Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The virtualization of free space and action : advancing a mondernized model of power inequity attenuation… Turner, Roger Anthony 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2014_september_turner_roger.pdf [ 2.93MB ]
Metadata
JSON: 24-1.0102482.json
JSON-LD: 24-1.0102482-ld.json
RDF/XML (Pretty): 24-1.0102482-rdf.xml
RDF/JSON: 24-1.0102482-rdf.json
Turtle: 24-1.0102482-turtle.txt
N-Triples: 24-1.0102482-rdf-ntriples.txt
Original Record: 24-1.0102482-source.json
Full Text
24-1.0102482-fulltext.txt
Citation
24-1.0102482.ris

Full Text

THE VIRTUALIZATION OF FREE SPACE AND ACTION: ADVANCING A MODERNIZED MODEL OF POWER INEQUITY ATTENUATION AND THE FREE SPACE INDEX (FSIx)  by  Roger Anthony Turner M.B.A., The University of Arkansas, 2003 B.B.A., The University of Central Arkansas, 2002  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Business Administration)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) June 2014 © Roger Anthony Turner, 2014 ii Abstract Although international electronic technologies have not provided a direct portal to a Utopian world of fairness and equality as some had dreamed, they have important implications for socio-organizational power that have thus far been under-considered.  Once a power structure is in place, it is generally self-reinforcing—powerful actors have motivation and wherewithal to subordinate others, and less powerful actors are constrained from resisting.  However, when a technology is introduced to a social system, it creates opportunities for interaction patterns governing power within that system to evolve.  International electronic technologies create new and rapidly changing, virtualized contexts for computer mediated communication, social media broadcasting, social networking, coordination, and action.  These contexts erode geographic, social, and psychological boundaries that have traditionally determined how, and if, power would be utilized, accepted, resisted, or challenged.  In this dissertation I present a modernized model of power that takes these changes into account and report six related empirical studies.  In advancing my model, I also draw from, refine, and extend free space theory. I argue that these technologies embed sheltered interaction contexts where the less powerful can express themselves and interact more freely.  These spaces can spawn social movements and other forms of collective resistance and ultimately result in social- and/or organizational change.  In Studies 1-3, I create the Free Space Index to identify such spaces both online and offline.  I collect data in Canada, the USA, and Denmark for cross-societal validation.  Studies 4-6 test two central propositions underling my model. The first is that electronic technologies discourage some power-reinforcing behaviours by raising perceived iii retribution risks; Study 4 examines this in an organizational decision-making context.  The second is that these technologies promote action challenging prevailing power structures.  Study 5 shows the effect of self-interest, a key offline action predictor, differs online.  Study 6 demonstrates that electronic technologies promote action by reducing participation costs—congruent with the slacktivism moniker often applied to Internet mediated social activism—but also by attenuating a number of socio-psychological constraints that discourage offline action.   I discuss the implications and limitations of my model and empirical work and suggest future research directions.      iv Preface Portions of this dissertation are based on collaborative work with Marc-David L. Seidel, who provided valuable guidance and insightful feedback to improve the theory advanced and the research that begins to test it.  An abstract of portions of Chapter 1 are reported in: Turner, R. A., & Seidel, M. D. L. (2013). Acting against authority, amended: Toward a modernized model of inequality attenuation. Administrative Sciences Association of Canada Organizational Theory Division Conference Proceedings, 34(1).  As first author, I was responsible for the initial theoretical framing, writing the first draft of the manuscript, and incorporating collaborator and conference reviewer feedback into the final manuscript.  Portions of Chapters 1 and 2 comprised: Turner, R. A. & Seidel, M. D. L. (2014). Resistance not resisted? Toward a validated measure of social spaces’ safety for resistance: The Free Space Index (FSIx).  Presented in I. Marti, D. K. Mumby, D. Seidl, and R. Thomas (Conveners) Organization Studies summer workshop, Corfu, Greece. As first author, I was responsible for the initial theoretical framing and research designs, collecting and analyzing the data, writing the first draft, and incorporating collaborator feedback into the final manuscript. An abstract of the first study in Chapter 4 is reported in: Turner, R. A. & Seidel, M. D. L. (2012). Predicting social action. Administrative Sciences Association of Canada Organizational Theory Division Conference Proceedings, 33(1). As first author, I was responsible for the initial theoretical framing and research designs, data collection and analysis, writing the first draft, and incorporating collaborator and conference reviewer feedback into the final manuscript. v The UBC Behavioural Research Ethics Board approved the research comprising this dissertation under certificate numbers H11-03415, H12-00917, H13-01731, and H13-02956. I gratefully acknowledge funding for portions of this research from the Administrative Sciences Association of Canada – Canadian Journal of Administrative Sciences PhD Research Grant, Copenhagen Business School – Sauder School of Business Partnership Fund Research Grant, Administrative Sciences Association of Canada PhD Student Travel Grant, Sauder School of Business Dean’s Doctoral Research Grant, and Sauder School of Business OBHR Division Conference Travel Grant. I also gratefully acknowledge additional support from grants awarded to Marc-David L. Seidel by the Social Sciences and Humanities Research Council of Canada and the Sauder School of Business and a grant awarded to Nancy Langton by the Sauder School of Business. vi Table of Contents Abstract ............................................................................................................... ii Preface ................................................................................................................. iv Table of Contents ................................................................................................ vi List of Tables ....................................................................................................... x List of Figures .................................................................................................... xii List of Equations............................................................................................... xiv Acknowledgements ............................................................................................ xv Dedication ........................................................................................................ xvii Chapter  1: Introduction ....................................................................................... 1 1.1 Illustrative Case .................................................................................................................. 4 1.2 Technology as Conduit for Change ................................................................................ 7 1.3 General Overview ............................................................................................................. 8 1.4 Characterizing “The Internet,” Circa 2014 ..................................................................... 9 1.5 Implications of the Internet for Power and Resistance ............................................. 12 1.5.1 The Internet and Resources for Resistance ............................................................ 14 1.5.1.1 Crowdsourcing as Mobilization ........................................................................ 20 1.5.2 The Internet and Fear of Punishment for Challenging the Status Quo ............. 22 1.5.3 The Internet and the Perceived Legitimacy of the Status Quo ............................ 24 1.6 The Internet and Social Action ..................................................................................... 27 1.7 The Internet and Risk of Abusing Power .................................................................... 39 1.8 Free Space – Construct Refinement and Extension into Virtual Contexts ............ 41 1.8.1 What Free Space Is and Is Not ................................................................................. 43 vii 1.8.2 Free Space in Online Contexts ................................................................................. 48 1.8.3 Free for X, Free for Y?: Topical Specificity of Free Space ................................... 50 1.8.4 The Multidimensionality of Free Space: Protection from Six Risk Categories .. 52 1.8.5 Free Space through the Eye of the Beholder.......................................................... 55 Chapter  2: Establishing a Measure of Free Space ........................................... 57 2.1 Conceptualization ............................................................................................................ 58 2.2 Characterizing the Measurement Model ...................................................................... 59 2.3 Generating and Refining Item Pool .............................................................................. 62 2.4 Study 1: Assessing Item-Dimension Fit ....................................................................... 62 2.4.1 Study 1 Participants .................................................................................................... 63 2.4.2 Study 1 Analyses and Results .................................................................................... 63 2.5 Study 2: Selecting Items and Testing Measurement Model ...................................... 68 2.5.1 Study 2 Participants and Procedures ........................................................................ 68 2.5.2 Study 2 Analyses and Results .................................................................................... 70 2.5.2.1 Item Selection ...................................................................................................... 71 2.5.2.2 Assessing Multicollinearity ................................................................................. 72 2.5.2.3 Assessing Convergent Validity .......................................................................... 72 2.5.2.4 Test of Model Specification ............................................................................... 73 2.5.2.5 FSIx Score Calculation ....................................................................................... 75 2.5.2.6 Hypothesis Testing .............................................................................................. 76 2.6 Study 3: FSIx Cross-Cultural Assessment ................................................................... 78 2.6.1 Study 3 Participants and Procedure ......................................................................... 81 2.6.2 Study 3 Measures ........................................................................................................ 82 2.6.3 Study 3 Analyses and Results .................................................................................... 84 viii 2.6.4 Study 3 Hypothesis Testing and Results ................................................................. 86 2.6.4.1 Free Space, by Society......................................................................................... 86 2.6.4.2 Moderating Effect of Self-perceived Power .................................................... 90 2.6.4.3 FSIx Dimensionality ........................................................................................... 92 2.7 FSIx Usage Guidelines ................................................................................................... 94 2.8 Chapter 2 Discussion ...................................................................................................... 98 Chapter  3: Testing Impact on Hierarchy-Sustaining Action ..........................102 3.1 Study 4 Hypotheses ....................................................................................................... 102 3.2 Study 4 Overview .......................................................................................................... 107 3.3 Study 4 Sample ............................................................................................................... 108 3.4 Study 4 Procedure and Measures ................................................................................ 109 3.5 Study 4 Results ............................................................................................................... 117 3.6 Chapter 3 Discussion .................................................................................................... 121 Chapter  4: Testing Impact on Social Action ...................................................123 4.1 Study 5 ............................................................................................................................. 124 4.1.1 Study 5 Hypotheses .................................................................................................. 125 4.1.2 Study 5 Procedures and Sample .............................................................................. 128 4.1.3 Study 5 Measures ...................................................................................................... 130 4.1.4 Study 5 Results .......................................................................................................... 131 4.1.5 Study 5 Discussion .................................................................................................... 138 4.2 Study 6 ............................................................................................................................. 139 4.2.1 Study 6 Hypotheses .................................................................................................. 140 4.2.2 Study 6 Overview ...................................................................................................... 146 4.2.3 Study 6 Procedures and Sample .............................................................................. 146 ix 4.2.4 Study 6 Measures ...................................................................................................... 149 4.2.5 Study 6 Results .......................................................................................................... 153 4.2.5.1 Main Effects ....................................................................................................... 160 4.2.5.2 Contextualized Effects ..................................................................................... 162 4.2.6 Study 6 Discussion and Post-Hoc Analyses ......................................................... 173 4.3 Chapter 4 Discussion .................................................................................................... 176 Chapter  5: Conclusion......................................................................................180 5.1 Limitations and Future Directions .............................................................................. 180 5.1.1 Theoretical Boundaries ............................................................................................ 180         5.1.2     Feeding off Grassroots? Formal Advocacy Organizations and (Apparently)                             Organic Online Social Movements ......................................................................... 184 5.1.3 Virtualized Free Space .............................................................................................. 185 5.1.3.1 Identifying and Characterizing Virtualized Free Space ................................ 190 5.2 Implications .................................................................................................................... 191 5.3 Summary ......................................................................................................................... 193 References ........................................................................................................ 200 Appendix A........................................................................................................251 Appendix B ...................................................................................................... 256 Appendix C ...................................................................................................... 260 Appendix D .......................................................................................................261 Appendix E ...................................................................................................... 266 Appendix F ....................................................................................................... 267  x List of Tables Table 1: Categories of Risk from which Free Space Need Provide Protection. ..................... 54 Table 2: Initial FSIx Item Pool and Strongest Fit Risk Categories ........................................ 65 Table 3: Discriminant Validity of FSIx Dimensions: Squared Interconstruct Correlations (Study 2) ........................................................................................................................... 75 Table 4 Summary Statistics and Correlations (Study 3) ........................................................ 87 Table 5 OLS Regression Predicting Free Space Comprised by Gay Pride Festival Events   in Three Societies (Study 3) ............................................................................................. 88 Table 6 Differences in Effect between FSIx Dimensions for Data from Gay Pride Festival Events in Three Societies for Advocating Preference in Hiring and University Admissions Decisions for LGBTQ Persons (Study 3) .................................................... 93 Table 7: Summary Statistics and Correlations (Study 4) ....................................................... 118 Table 8: Multiple Regressions Predicting Risk to Firm and Risk to Self (Study 4) .............. 118 Table 9: Multiple Regressions Predicting Final Offer and Offer Change (Study 4) ............ 120 Table 10: Summary Statistics and Correlations: Advocating One’s Position on a Social    Issue to a Group of Friends (Study 5) ........................................................................... 132 Table 11: Means, Standard Deviations, and Correlations: Advocating One’s Position on a Social Issue within the Context of a Community Discussion (Study 5) ....................... 133 Table 12: Ordinary Least Squares Regression Predicting Willingness to Advocate to a   Group of Friends (Study 5) ............................................................................................ 135 Table 13: Ordinary Least Squares Regression Predicting Willingness to Advocate within   the Context of a Community Discussion (Study 5) ....................................................... 137 Table 14: Outbound Social Tie Strength by Advocacy Act (Study 6) ................................... 156 xi Table 15: Summary Statistics and Correlations among Study Variables: Willingness to Advocate One’s Position on a Social Issue (Study 6) .................................................... 157 Table 16: Multilevel Maximum Likelihood Analysis, Main Effects Model Predicting Willingness to Advocate (Study 6) ................................................................................. 159 Table 17: Multilevel Maximum Likelihood Analysis of Context Effects, Predicting Willingness to Advocate (Study 6) ................................................................................. 163 Table 18: Factors Affecting Willingness to Advocate in Offline and Online Contexts    among Social Actors Who are Not Vested and Vested in the Focal Issue (Study 6) .... 172  xii List of Figures Figure 1. Percentage of US adults utilizing the Internet by age group, 2000-2012. ................ 10 Figure 2. Modernized model of social power and resistance. ................................................ 13 Figure 3. The global digital divide. ......................................................................................... 19 Figure 4. PLS path coefficients for theorized FSIx measurement model (Study 2). .............. 74 Figure 5. Mean FSIx score, by condition (Study 2). ............................................................... 77 Figure 6. FSIx PLS path coefficients for data collected in three international cities         (Study 3). .......................................................................................................................... 85 Figure 7. FSIx Score by Free Space Embedding Social Context (Study 3). ........................... 90 Figure 8. Conditional effect of Self-perceived Power (i.e., Internal Locus of Control)           on free space perceptions for Gay Pride Festival events in Vancouver, BC, Canada, Copenhagen, Denmark, and Memphis, TN, USA, for advocating preference in       hiring and university admissions decisions for LGBTQ persons, as measured by the FSIx (Study 3). ................................................................................................................. 91 Figure 9. Population weighted mean social stigma perceptions for focal groups affected (i.e., vested groups) in scenarios, with 95% confidence intervals (Study 6). ................ 154 Figure 10. Mean perceived societal attitude toward programs solely benefitting focal     vested groups in scenarios, with 95% confidence intervals (Study 6). .......................... 155 Figure 11. Nature of Online x Vested x Attitude Strength interaction predicting willingness to advocate (Study 6). .................................................................................................... 167 Figure 12. Nature of Online x Vested x Proactivity interaction predicting willingness to advocate (Study 6). ......................................................................................................... 168 Figure 13. Nature of Online x Vested x Anticipated Social Support interaction predicting willingness to advocate (Study 6). ................................................................................. 169 xiii Figure 14. Nature of Online x Vested x Anticipated Positive Affect interaction predicting willingness to advocate (Study 6). ................................................................................. 170 xiv List of Equations Equation 1: FSIx Score. ........................................................................................................... 76  xv Acknowledgements They say it takes a village, and for me that has certainly been true.  The village of people supporting me throughout my academic journey is a wonderful one.  Marie Mitchell and Mary Uhl-Bien helped me find my way.  Sandra Robinson had confidence in me from the start and has contributed greatly to my development.  Daniel Skarlicki showed me what it takes to see a manuscript all the way through to fruition.  Nancy Langton and Kate White gave generously of their time and energy to serve on my committee and have been tremendously supportive and developmental.  Thank you all!  And, thank you especially to Marc-David Seidel, who has been a wonderful mentor, collaborator, and committee chair.  I appreciate your guidance, encouragement, insight, constructive criticism, generosity, and witty banter more than I can say. I was very fortunate to have an awesome cohort to learn with and to learn from—Marjan, Leah, and Luke, it has been an honor to experience this with you.   Finally, my parents, Kay and Roger, have always encouraged me to follow my heart—even when it leads me far from home.  My sisters love me enough to go along with that.  My grandparents convinced me that I could be anything that I wanted to be and loved me with hearts wide open.  Jennifer has always been there, even when she’s not.  Roger Fisher and Terry Love each played an integral role in making me who I am today.  Matt encouraged me down the homestretch.  I am grateful for each of you and for the rest of my wonderful friends and family who are too many to mention but loved all the same!   xvi I also gratefully acknowledge funding for my doctoral studies from the Dean Earle D. MacPhee Memorial Fellowship in Business Administration, UBC Four Year Doctoral Fellowship, and UBC International Graduate Student Scholarship.    xvii Dedication  For the late Marlene LeDuc who insisted that I “not be ordinary”     1 Chapter  1: Introduction1 October 20:  Seriously they want me to wear purple because five queers committed suicide. The only way im wearin it for them is if they all commit suicide. (sic) – Clint McCance, school board vice-president, in a Facebook post responding to a movement to combat bullying in schools in the wake of a rash of gay youth suicides October 28:  I am going to resign from the school board. – Clint McCance, in an interview on CNN’s Anderson Cooper: 360; he tendered his official resignation three days later Hierarchy is the single most prevalent form of organization (Magee & Galinsky, 2008). One might think that because hierarchies are characterized by privilege for some and exploitation and subjugation for others, they would elicit a great deal of contentiousness and that challenges would abound from those who they disadvantage. However, historically that has most often not been the case, as the patterns of domination and legitimation inherent to hierarchies tend to perpetuate inequalities, discourage challenges, and, ultimately, deter reorganization (Martorana, Galinsky, & Rao, 2005).                                                     1 An earlier version of portions of this chapter comprised a conference paper that received honorable mention for best paper in organizational theory at the 2013 ASAC annual conference.  That paper was a collaborative work with Marc-David L. Seidel.  A full manuscript is in preparation for the peer review process.  2 As of October 19, 2010, Mr. McCance held an elected position of power within his community, vice president of the local school board (Mail Foreign Service, 2010b).  In late September and early October of that year, in memory of several gay teenagers who had recently committed suicide as a result of bullying, users of the social networking website Facebook promoted a national campaign to raise awareness of bullying in schools, asking people to wear purple on October 20 (Heussner, 2010a).  On that date, Mr. McCance engaged in a behavior that, left unchecked, would act to reinforce an established hierarchy (in this case, the social sexual hierarchy in which heterosexuality occupies the highest level: see, e.g., Herek, 1990). He posted the following response to the Spirit Day campaign on his Facebook profile page2 (Broverman, 2010):  Seriously they want me to wear purple because five queers committed suicide. The only way im wearin it for them is if they all commit suicide. […] being a fag doesn't give you the right to ruin the rest of our lives. If you get easily offended by being called a fag then dont tell anyone you are a fag. Keep that                                                   2 Screen captures of the postings were subsequently displayed in numerous online locations.  I archived two examples, now perpetually available at the following urls, on April 5, 2014:  https://web.archive.org/web/20140405171348/http://www.advocate.com/news/news-features/2010/10/26/arkansas-school-board-member-thinks-fags-should-die and https://web.archive.org/web/20140405171638/http://floridaagenda.com/wp-content/uploads/2010/11/ClintMccanceFacebook.jpg.  3 shit to yourself. […] It pisses me off though that we make a special purple fag day for them. I like that fags cant procreate. I also enjoy the fact that they often give each other aids and die. If you arent against it, you might as well be for it. […] I would disown my kids they were gay. (Sic)     Mr. McCance’s posting came to the attention of a member of the lower-power group (i.e., a non-heterosexual person) through bridging ties (Burt, 1992)—connections between two otherwise disconnected (Facebook) social networks (Cooper, 2010d; Roberts, 2010).  I was that person.  Shocked that a school official from my hometown had responded to the anti-bullying event with what he perceived to be more bullying (Cooper, 2010d), I wrote letters to McCance and the school board calling for his resignation or removal from his position of power in the school system (Watson, 2010). I also alerted the Human Rights Campaign of these events (Simon, 2010). The events that followed brought me to the realization that the processes influencing power structures may be changing as a result of the introduction and growing popularity of electronic technologies for social interaction and communication.  That realization served as the motivation for this dissertation, and as I have worked on this stream of research, I have realized that these changes apply equally well to all forms of power abuse.  My goal in this thesis and dissertation research is to leverage and generalize what started as a personal experience to make a strong theoretical contribution to the social power and social movement literatures.  4 1.1 Illustrative Case On October 26, 2010, The Advocate, a gay, lesbian, bisexual, and transgendered magazine published a story (i.e., Broverman, 2010) about McCance’s Facebook postings. That story, which included screen captures of McCance’s posting, generated a tremendous amount of web traffic to the site (Simon, 2010).  Thereafter, as David Koon, associate editor of The Arkansas Times (2011) put it, “the smelly breeze of Clint McCance's comments soon blew itself into a hurricane. Within a few days, a comment that might have been ignored only a few years before had focused the bright light of worldwide attention down on McCance.”   A frenzy of online activity fueled that hurricane.  A Facebook group entitled Fire Clint McCance was created (Facebook.com, 2010a), and within a couple of days, 65,000 people had joined it, initially at a rate of over 1,000 new members per hour (Simon, 2010).  An electronic petition was also established at change.org calling for McCance to be disassociated from the school district (Heussner, 2010b; Jones, 2010).  The school district received a steady flow of phone calls, emails, and letters asking for McCance to be removed from his position or resign.  The school district superintendent reported that he had personally received over 15,000 of these emails and that the district office telephones rang "nonstop" with calls about the issue (Koon, 2010).  Many of the emails the district received were generated automatically as a result of people adding their names electronically to the change.org petition, but people also disseminated the direct contact information for the school and its officials online (e.g., Facebook.com, 2010b). Amid this, the cable television news network CNN interviewed the “whistleblower”—me—in primetime on Anderson Cooper’s AC:360 (Cooper, 2010d); 5 MSNBC followed suit (Roberts, 2010).  Network news stations (e.g., ABC News, 2010; Hayes, 2010), as well as cable’s CNN and Fox News (e.g., Associated Press, 2010), reported the story.  McCance endured public condemnations of his actions from state and national officials, including the US Education Secretary (e.g., Cooper, 2010a; Mail Foreign Service, 2010b). Celebrities (e.g., Takei, 2010, a Youtube video that has been viewed over 1,600,000 times) and popular opinion leaders, such as psychologist and television host Dr. Phil McGrath and television host Ellen DeGeneres (Cooper, 2010c; Stewart, 2010) also called him to task.  Ostensibly angered individuals shared McCance’s name, home address and telephone number, email address, cellphone number, and business name, address, and telephone number online (in, e.g., comments section of SuchIsLifeVideos, 2010). He later reported having received a torrent of emails and phone calls, some of which conveyed death threats, prompting him to send his wife and two children out of state, install a security system in his home (CNN Wire Staff, 2010a, 2010b; Mail Foreign Service, 2010a), and reportedly carry a pistol for protection (Koon, 2010).  Protestors organized a public rally calling for his resignation through Facebook (Facebook.com, 2010a) and the event drew national media attention (CNN Wire Staff, 2010c).   Amid the sudden and constant barrage of pressure he faced, McCance made the decision to vacate his elected position of power (see, e.g., Simon, 2010).  On October 28, just eight days after posting the homophobic comments that triggered this “hurricane” (Koon, 2011) and just three days after The Advocate broke the story online, McCance also appeared live on AC360. There, he apologized for his actions and made known his intention 6 to resign (CNN Wire Staff, 2010b; Mail Foreign Service, 2010a). The school board unanimously accepted McCance’s letter of resignation (available: Arkansas Matters, 2010) on November 2 (CNN Wire Staff, 2010a).  It is theoretically possible that a movement to countervail Mr. McCance’s actions would have occurred and that he would have ultimately vacated his formal position of power in the absence of the Internet and related technologies—as Horney (1939) described, those who are disadvantaged by the prevailing organizational form may act against that system and those whose actions sustain it, and history is replete with stories of the rise and fall of standard bearers, social elites, dominant businesses and organizations, and even “superpower” nations.  However, a diverse body of theory and research suggests this is somewhat unlikely.  In fact, people’s general tendency has been not to challenge those with greater power or higher standing than themselves, actively combat behaviors and entrenched ideologies that at once reflect and reinforce status quo power distributions, or even engage in more benevolent action to promote the interests of the disadvantaged more generally (e.g., Keltner, Gruenfeld, & Anderson, 2003; Magee & Galinsky, 2008; Martorana, Galinsky, & Rao, 2005; Ratner & Miller, 2001).   A chain of reasoning delineated by Graham (2012) suggests that a purely local movement might have been especially unlikely in this case: Arkansas is in the Bible Belt (Barton, 2010), research has shown that a large portion of those living in the Bible Belt identify as fundamentalist Christians (Barton, 2010), and fundamentalist Christians tend to be high in social dominance orientation (Hathcoat & Barnes, 2010; Whitley & Lee, 2000), a belief that one’s personal group is dominant and superior to all other groups (Pratto, 7 Sidanius, Stallworth, & Malle, 1994), and to hold highly negative views toward homosexuality (Barton, 2010; Herek, 1987, 2004).  In such an environment, both LGBT individuals and their allies would be discouraged from acting by fear of negative social repercussions (Keltner et al., 2003; Magee & Galinsky, 2008).  Although outside actors might have intervened, in order for that to happen, information about the subjugating acts would need reach them.  As I will describe below, however, prior to the introduction of Internet technology, social networks that would allow for this information to spread by word of mouth largely have been geographically constrained, and elites controlled access to mass-communication channels that would allow for it to spread outside of personal social networks.  In this case, however, the information reached me nearly 4,000 miles away, and through the Internet-mediated response it elicited and the mainstream media attention that response garnered, was subsequently broadcasted far and wide. 1.2 Technology as Conduit for Change The question of when and how people are likely to resist versus conform to the status quo has important implications for the organization of status and power within societies and organizations, and as the case above serves to illustrate, its answer seems to be changing.  Social systems—organizations, social hierarchies, and the like—are social constructions whose perpetuities rely on their own repetition (Thomas, 1994). But, when those who are a part of a social system interact with one another and with other elements comprising that structure, governing patterns can evolve, meaning that the socially constructed order may be imperfectly reproduced (Giddens, 1979). Technologies are one of the prevalent elements of social systems with which people interact.  When new technologies are introduced, they are 8 “infused with objectives—that is, reflective of the interests or goals of particular groups within the social system,” but they are at the same time “objective—that is, reflective of a logic, a set of rules and conditions, independent of the social system” (Thomas, 1994, p. 19), meaning that members may put them to different uses than those installing them had anticipated.  As such, new technologies bring with them the potential for unforeseen variation to arise within the system, creating opportunities for change that previously did not exist or were unrecognized (Barley, 1986; Cohen, 1997).   1.3  General Overview Since the mid-1990’s, as the Internet has become increasingly woven into everyday life.  As I have begun to describe above and will subsequently expound upon, as Internet technology becomes an ever-more central beam in the structure of organizations and societies throughout the world, it brings with it new mediums for interaction, communication, organizing, surveillance, social support, coercion, and resistance that did not exist when most of the research on social inequality attenuation occurred.  Although the opening vignette comprises a perhaps powerful illustration of the capacity for this technology to mediate a new form of collective resistance, my focus here is broader than that.  In this thesis, I advance updates to a model of power and resistance that acknowledges the growing ubiquity of electronically-mediated social interaction in everyday life.  These technologies, I argue, mediate access to virtualized interaction spaces where challenges may be spawned (consistent with Rao & Dutta, 2012); allow for coordination and communication among the less powerful and their higher-power allies that can precipitate successful challenge; and, consequently, raise the risk of exposure and negative return for powerful 9 individuals, collectives, and/or organizations who abuse their power or engage in otherwise illegitimate acts that reinforce or redouble their relative advantages in the system. I argue that the wide-spread availability of the Internet provides means for non-elites and their allies to overcome obstacles that have previously dissuaded them from resisting the status quo and/or working proactively to improve the standing of those disadvantaged within existent hierarchies.  I argue that modern communication and social interaction technologies allow for a new form of collective punishment for abusing power, social media fuelled hurricanes such as that experienced by Mr. McCance (Koon, 2011).  I test two central propositions of my model—(1) that the virtualized context comprised by the Internet encourages social action participation and (2) that that the potential for Internet-mediated action against entities who abuse their power serves to raise the perceived risk associated with such action and, ultimately, to discourage it.  I also refine a key construct in the model and advance a validated measure of it to facilitate future research.  First, however, because the Internet is a very fluid concept, I briefly characterize it at present to provide context for this work.  1.4 Characterizing “The Internet,” Circa 2014 The Internet is an “electronic network of networks that links people and information through computers and other digital devices” (DiMaggio, Hargittai, Neuman, & Robinson, 2001, p. 307). Access to the Internet is readily available in most parts of the industrialized world.  As of the most recent world Internet survey, North America ranks among the leaders in Internet availability (Internet World Stats, 2012).  As of 2014, 87 percent of US adults were active online (Fox & Raine, 2014). Usage rates will continue to rise as older generations 10 among whom Internet use is much less pervasive (although still high—57 percent of Americans over 65 were Internet users in 2014) give way to younger generations for whom Internet use is almost ubiquitous (97 percent of 18-29 year old Americans were) (Fox & Raine, 2014). Figure 1, below, shows the upward trend lines for Internet use among all age groups in the US.    Figure 1. Percentage of US adults utilizing the Internet by age group, 2000-2012. Figure 1 from © Zickuhr, K., & Madden, M. (2012, June 6). Older Adults and Internet Use. PewResearch Internet Project. Washington, DC: Pew Research Center. Retrieved February 14, 2014, from www.pewinternet.org/2012/06/06/main-report-15/. By permission from publisher.   The Internet comprises “a meta-medium: a set of layered services that make it easy to construct new media with almost any properties one likes” (Agre, 1998, as reported in DiMaggio, et al., 2001, p. 309).  As it has become increasingly woven into everyday life, the Internet has come to represent not only wires and hardware but also all of the uses to which 11 those wires and hardware are put, and because it is such a flexible technology, those uses are innumerable and constantly evolving.  Examples of common uses at present include email, video-conferencing (e.g., Skype); social networking (e.g., Facebook, Twitter, Google+); professional networking (e.g., LinkedIn); information search and retrieval (i.e., Googleing); peer-to-peer file sharing (i.e., torrenting); shopping; gaming; entertainment; and creating, archiving, and sharing user-created content (e.g., YouTube, blogging, Instagram, Snapchat).   Current Internet technology also allows for the joining of individual efforts and resources toward common goals (Seidel & Stewart, 2011).  Examples of this include the collective curation of knowledge (e.g., Wikipedia.org: see, e.g., Burke, 2014), political action (e.g., change.org, whitehouse.gov: see, e.g., Lee, 2013), and open-source software (OSS) development (for a recent review of OSS communities, see Martinez-Torres & Diaz-Fernandez, 2013). They also include other forms of collective problem solving (i.e., crowdsourcing and expertsourcing: see e.g., Madsen, Woolley, & Sarangee, 2012) and the collective funding (i.e., crowdfunding: see, e.g., Mollick, 2014) of business ventures, scientific research, and creative projects (e.g., kickstarter.com, gofundme.com).   Increasingly, online activities take place on the go.  Applications (i.e., apps) to facilitate them on mobile hardware such as smartphones and tablet computers (see, e.g., Ghose & Han, 2014) proliferate as mobile hardware supplants the desktop computer as dominant nexus to the Internet. The tipping point when the number of global users accessing the Internet on mobile devices exceed those accessing it with desktops is expected to occur in 2014 (Boyle, 2014).  12 1.5 Implications of the Internet for Power and Resistance Figure 2, below, shows my model.  In part, that model reflects established relationships between social power, hierarchy sustaining acts, the perceived legitimacy of those acts, and response from others to those acts3.  Hierarchies are so pervasive and so resilient as to be considered inevitable (Sidanius, 1993).  They tend to self-perpetuate because, as Adam Smith pointed out in The Wealth of Nations in 1776 (English translation, 1976), it is the typical nature of things that resources beget social power and social power begets resources.  As such, those who have social power advantages—those higher in the hierarchy—are both motivated to engage in acts that buttress the structure that favors them and privy to resources that enable them to do so (Magee & Galinsky, 2008; Sidanius & Pratto, 1999; Weber, 1947).   Among individuals, this might manifest coercion, threats, harassment, hate speech, resource withdrawal, or even violence against members of lower-power groups (e.g., Berdahl, 2007; Bourhis, 1994; Costarelli, 2006), for example.  Such acts are sufficiently prevalent among social groups in North American that prejudiced action is perceived not as the exception but as the rule (Finchilescu, 2010; Vorauer, Main, & O'Connell, 1998).  For organizations, it might manifest such things as predatory pricing (McGee, 1958), market exploitation and other anti-competitive acts (Vickers, 2005), or exorbitant interest rates                                                   3 For reviews of the literatures establishing these relationships, see, e.g., Magee and Galinsky (2003), Martorana et al., (2005), and Snyder and Omoto (2007).   13     Figure 2. Modernized model of social power and resistance. charged for loans made to those with nowhere else to turn, both domestically (Karger, 2004) and abroad (Hudon & Sandberg, 2013).  Although these actions might seem very different on the surface, they share a common outcome; they maintain or extend the more powerful entity’s advantaged standing.   Despite the inherent inequity and pervasiveness of exploitive and subjugating acts in hierarchies, existent literature tells us that people generally have not acted against them for several reasons (Magee & Galinsky, 2008).  I focus here on three of the most central of these.  One: They cannot access and mobilize sufficient resources to successfully take on more powerful entities (e.g., McCarthy & Zald, 1973; McCarthy & Zald, 1977; A. Smith, 14 1976 [1776]). Two: They fear punishment and seek foremost to avoid it (e.g., Anderson & Berdahl, 2002; Keltner et al., 2003). And, three: They perceive that the system is legitimate, despite their disadvantaged position within it, and thus are not motivated to challenge it (e.g., Jost, Pelham, & Carvallo, 2002; Kay et al., 2009; Martorana et al., 2005; M. Weber, 1947; Wright, Taylor, & Moghaddam, 1990).  I will expand upon these and address implications of virtualization technology (i.e., “the Internet”) for each, in Sections 1.5.1 through 1.5.3, respectively.   My primary proposed enhancements in the model above include the moderating influence of virtualized context on (i.) resistance decisions and (ii.) the success of resistance efforts, along with (iii.) the subsequent risk associated with undertaking hierarchy sustaining action.  I also propose that (iv.) virtualized context for resistance bolsters the delegitimizing effect of resistive action on unjust hierarchy-sustaining acts going forward and that (v.) free spaces (and especially those embedded within virtualized contexts) may be especially conducive to individual- and collective- acts of resistance.  I underpin and elaborate these propositions in subsequent sections of this chapter. 1.5.1 The Internet and Resources for Resistance As the social movement literature highlights, those for whom the outcomes of reorganizing power would have the strongest positive valence are likely to have low expectancies of achieving those outcomes and, thus, have little motivation to act. Those at lower levels of hierarchies are at a power disadvantage and because power translates into resource access, those who are most disadvantaged by the status quo are also those with access to the fewest resources to challenge it (Zald & McCarthy, 1979).   15 A good communication network is the most basic resource that must be available to the less powerful if a challenge to the status quo is to succeed (Langton, 1987).  Prior to the introduction of Internet technology, access to communication channels was largely limited (cf. Gans, 2005; Gans, 2011).  However, the Internet encompasses the one-to-many broadcasting features of newspapers, radio, and television; the one-to-one communication features of letters, telephone calls, facsimiles, and even face to face interaction (as through video calls/conferencing, e.g., Skype); and the many-to-many communication features of community and group discussion (Stalder, 2006).  Moreover, the Internet allows for communication to occur almost instantaneously over long distances, and, through it, messages can be disseminated broadly at virtually no cost (Seidel & Stewart, 2011). This helps the less powerful to express and disseminate their grievances, build alliances and coalitions, marshal resources, and organize for resistance.  It also can result in reduced information asymmetry between the resister and the powerful during conflict, helping to further leverage the field of contest (Earl, McKee Hurwitz, Mejia Mesinas, Tolan, & Arlotti, 2013). The broad communication of information online is made possible, in part, by Internet-mediated social networking, a very popular online activity.  Of American Internet users, for example, more than one in two access social networking websites such as Facebook or Google+ on a given day, and 86 percent of users 18-29 do so (Zickuhr & Madden, 2012b).  There, they can post information visible to their ties in the network, visible to any user of the website (at the poster’s discretion), as well as communicate directly with individual ties or groups of ties through the website’s direct messaging functionality. 16 Information may also be disseminated through email—over 90 percent of Internet users utilize email (Zickuhr & Madden, 2012b)—or through Twitter, a social media broadcasting platform through which users send more than 500 million messages (Tweets) on an average day, and through it, as many as 143,199 messages have been broadcasted in a single second (Krikorian, 2013). Some other online communication options include video calling/conferencing (e.g., Skype), instant messaging (e.g., Windows Messenger; Yahoo! Messenger); and interacting within forums, blogs, and interest groups. Because multiplex communication channels are readily available wherever the Internet is introduced and come at little or, in most cases, no incremental cost, the historical communication advantage enjoyed by the more powerful should be reduced.    This is not without caveat, however, because access to the Internet is not universal.  Instead, people differ in their capacity to access the Internet in a variety of ways (van Dijk, 2005, 2012).  The most basic of these relates to material access—some people live in societies or pockets of societies in which Internet connectivity is not physically available.  Others live where Internet connectivity is widely available, but they cannot afford to purchase Internet-enabled hardware of their own, and there is little or no Internet-enabled hardware publically available.  For those for whom the Internet is not materially accessible, its capacity to mediate access to an inexpensive and reliable communication network clearly cannot be realized.   In North America, race has comprised a material access divide in Internet access.  However, mobile Internet access is eroding that division.  For example, a 2010 Pew Internet and American Life survey showed that a greater proportion of “Black, non-Hispanic” (87 17 percent) and “Hispanic” (87 percent) respondents owned cell phones than did “White, non-Hispanic” respondents (80 percent) (A. C. T. Smith, 2010).  A 2012 survey by that same organization found 64 percent of “Black, non-Hispanic” and 63 percent of “Hispanic” respondents used their cell phones to access the Internet, significantly outstripping the 52 percent of “White, non-Hispanic” respondents who did so (A. C. T. Smith, 2012). From a global perspective, the digital divide has been construed as one between rich and poor nations or developed and developing ones (see, D. S. White, Gunasekaran, Shea, & Ariguzo, 2011). But, White et al. (2011) conducted a study clustering countries according to per capita pervasiveness of home computers, Internet access, and Internet bandwidth. It found countries such as Antiqua, Aruba, Barbados, Chile, Estonia, Hungary, Jamaica, Lithuania, Slovak Republic, and Slovenia grouped in the top tier alongside (and in some cases above) countries such as Canada, the United States, and Japan.  Figure 3 reproduces those authors’ depiction of the “global digital divide.”   Moreover, people increasingly access the Internet through alternative hardware (i.e., not personal computers), even in the most highly economically disadvantaged societies.  Reports indicate, for example, that it is surprisingly common for people within countries in the lowest tiers of computer-mediated Internet access to utilize mobile hardware. The Pew Research Center (i.e., Wike, 2014) concluded that, “In a remarkably short period of time, Internet and mobile technology have become a part of everyday life for some in the emerging and developing world. Cell phones, in particular, are almost omnipresent in many nations” (p. 2).  Although that study showed that smartphones are less ubiquitous than standard cell phones, a significant number of people in these countries reported using the 18 Internet on a daily basis (Wike, 2014).  Tying this back, in Jordan, a Tier 3 country in White et al.’s (2011) classification, 95 percent of those polled in the Pew study owned cellphones, 37 percent owned a smart phone, and over half were Internet users.  Further, even tightly-controlled North Korea has begun to allow limited access to mobile Internet within its borders (BBC, 2013).  Thus, although there are certainly limitations on access that can impede the Internet’s capacity to mediate access to an inexpensive and reliable communication network in a global sense, access is much more readily available than ever before, and it will be ever more available as wireless Internet capabilities continue to spread and inexpensive Internet-enabled mobile devices continue to proliferate.  Access limitations may not stop at physical availability or financially feasibility, however.  The electronic flow of communication and information may itself be monitored and metered, as might be the case with an authoritarian government that censors content deemed threatening or reprobate.  At present, however, even strict Internet control protocols like those comprising the “Great Firewall of China” can be functionally circumvented with readily available, free technology (e.g., Arthur, 2010; Tor, 2013). In the words of Chinese artist and activist Ai Weiwei (2012), “they [the Chinese government] must understand it's not possible for them to control the Internet unless they shut it off," and that is just not something they are willing to do at present. Indeed, the Internet can facilitate challenge (Kuhn, 2009) and activist activity (e.g., Xie, 2008) in China (for a recent treatment of online activism in China, see Yang, 2013).     19               Figure 3. The global digital divide.  Reprinted from © White, D. S., Gunasekaran, A., Shea, T. P., & Ariguzo, G. C. (2011). Mapping the global digital divide. International Journal of Business Information Systems, 7(2), 207-219. Page 216. By permission from Inderscience Publishers.  Inderscience retains copyright of the figure and article from which the figure was taken. 20 That being said, battles over the flow of information between Internet censors and those acting to circumvent them are ongoing.  Especially at times and in locations in which people are explicitly or functionally blocked from interacting and communicating through the Internet, an online context for resistive action may create more opportunity for some than for others.  Groups and individuals with advanced “skills access,” which is to say those with the knowledge and ability to utilize the Internet in complex ways (van Dijk, 2012), should be better positioned to circumnavigate controls, mask their activities to avoid being detected or identified (Leberknight, Chiang, & Wong, 2012)4 or camouflage their true intentions, transfer and access resources, and the like, in order to undertake successful resistance. On the other hand, monitors may detect those who lack sufficient skills access but attempt to utilize the Internet to alter the status quo and thus subject those individuals to (further) punishment as a result.  This should be especially true in situations in which the government or another entity with the capacity to monitor and control the flow of electronic information is also the target of that challenge. 1.5.1.1 Crowdsourcing as Mobilization  Internet-mediated communication and social interaction should also have a second-order effect in helping the less powerful overcome resource deficits because it makes                                                   4 There are highly technically-capable organizations and resource providers (e.g., internetfreedom.org, www.torproject.org) who seek specifically to enable those who do not have a high degree of technical competence to get around barriers and protect themselves in their online resistance efforts. 21 crowdsourcing more feasible.  Crowdsourcing, a term coined by Howe (2006, 2008), is the act of outsourcing a task to a crowd rather than doing it yourself or outsourcing it to a designated agent.  Crowdsourcing capabilities should attenuate the influence of traditional resource differentials between those low and high in standing.   The Internet allows for crowdsourced efforts to combat power abuse and exploitation. This is to say that through the Internet, grievances can be publicized and calls for action made, and the consequential, seemingly trivial contributions of those with few resources can then accumulate to a level sufficient to redress the motivating grievances. Crowdsourcing allows each “little person” in the crowd to contribute a bit of resource toward a problem—signing e-petitions, joining a Facebook group, forwarding an email, posting or reposting information about a grievance, or even watching a YouTube video could all contribute (e.g., Bernoff & Schadler, 2010; CNN Wire Staff, 2010b; Tomassoni, 2012).  Crowdsourcing only works if the crowd is sufficiently large, with enough members having sufficient motivation and sufficient resources to solve the problem motivating the call to action (Afuah & Tucci, 2012).  Extremely large social networks like those the Internet facilitates, along with the diverse flow of information within those large networks should yield an increased likelihood that a sufficient number of actors will be reached.  Moreover, by allowing people to participate electronically, the Internet may also expand the potential actor set to include demographics such as the elderly or those with disabilities who otherwise might be excluded from resistance efforts (cf. Mukherjee, 2010). A crowdsourced approach to mobilization is not without potential drawbacks, however.  For example, social change is often lubricated by supportive media coverage, and 22 many people get their news from local outlets (e.g., 71 percent of US adults tune in for local television news broadcasts: Olmstead, 2013).  Here lies the potential rub: Research by Andrews and Caren (2010) suggests that social movements gain greater traction in local media when they are large, local, conventional, and highly organized.  Although crowdsourced mobilizations are likely to be large, local news organizations may be less likely to perceive them as conventional, local, and highly organized, especially as compared to a traditional social movement visibly led by a local organization.  On the other hand, however, crowdsourcing may facilitate mobilization that would otherwise not occur, and a crowdsourced mobilization, especially against a target in a news organization’s local area, has infinitely more potential to elicit local media coverage than one that never occurs at all. 1.5.2 The Internet and Fear of Punishment for Challenging the Status Quo Fear of punishment is another constraint on action to resist established power structures or challenge the acts through which they are sustained.  Actors lower in a hierarchy rely on those above them for resources. The resource advantage enjoyed by higher level members predisposes them toward opportunity seeking, whereas the resource disadvantage suffered by lower level members predisposes them toward conflict avoidance (Anderson & Berdahl, 2002; Anderson & Spataro, 2005).  Because those low in power are motivated primarily to avoid punishment and retain what resources they have (Anderson & Berdahl, 2002; Fiske, 1993; Keltner et al., 2003), they are little inclined to challenge the status quo, even though doing so could ultimately advance their standing.  However, hierarchies and the exploitation they can enable do not always go unchecked.  Those low in power are not limited only to avoiding more powerful others; they 23 also can act against them and against the system that perpetuates the power imbalance they face in their everyday lives (Horney, 1939).  Scholars have argued that this is more likely to occur when those lower in power feel powerful—even when they are in an objective power disadvantage (Anderson & Spataro, 2005; Martorana et al., 2005).  Internet-mediated interaction should act to bolster felt power among those at lower levels of hierarchies that exceeds the power that they experience in traditional interactions.  Perhaps the most central reason this should be the case is that the Internet allows for interactions to occur outside the watchful eyes of typical social interaction spaces; the Internet houses virtualized free space. Free space, which will be discussed in much greater detail in Section 1.8, is a context for social interaction in which ordinary citizens can voice and enact their volitions without deference to more powerful others (Evans & Boyte, 1986; Fantasia & Hirsch, 1995).  Free spaces are liminal, situated in social spaces between private lives and large-scale institutions (Evans & Boyte, 1986; V. W. Turner, 1986).  They are expected to insulate the less powerful, at least to some extent, from the risks and controls present in other social contexts and thereby facilitate interactions that otherwise would be too risky to undertake (Gamson, 1996).   The Internet allows for people to control the relative visibility of their actions online, up to and including acting virtually completely anonymously.  Internet-savvy subversives can conceal their identities behind false names, firewalls, virtual proxy networks, encrypted data transmissions, network relay routing, and even more sophisticated techniques (see, e.g., www.torproject.org, www.wefightcensorship.org/online-survival-kithtml.html, Leberknight et al., 2012).  The influence of fear of punishment should be less meaningful when actors 24 expect that their identities will not be discovered, and to the extent that actors feel efficacious in acting anonymously online, they should perceive little risk from acting. Thus, any constraining influence that their objective standing in the social hierarchy would normally offer should be reduced online.  In fact, it may make them especially likely to act—“those in low power may feel the disposition to act against others when they have a sense of power and especially when that sense of power is in sharp contrast to actual levels of hierarchical power and control” (Martorana et al., 2005, p. 299; see also Anderson & Spataro, 2005).   1.5.3 The Internet and the Perceived Legitimacy of the Status Quo Another constraint on challenges of an ensconced hierarchy is that if actors perceive that system to be legitimate, they are unlikely to act against it (e.g., Ellemers, Wilke, & Vanknippenberg, 1993; Hornsey, Spears, Cremers, & Hogg, 2003; Ng & Cram, 1988; S. C. Wright et al., 1990).  One reason that people rebel against the status quo so infrequently is that prevailing social structures are pervaded by hierarchy-sustaining myths that actors within those structures internalize. These myths ultimately leading those at lower levels of the hierarchy to accept their lots (Kay et al., 2009; Kay & Jost, 2003; Sidanius & Pratto, 1999).  Because those low in power are motivated to “see the way things are as the way things should be” (Kay et al., 2009, p. 421), they tend to justify the status quo and rationalize acts of subjugation perpetrated against them (e.g., Jost & Banaji, 1994; Jost et al., 2002; Pratto et al., 1994; Sidanius & Pratto, 1999).  As a result, in many cases hierarchical abuses might not be viewed as overly problematic and thus offer little motivation for countervailing action.   25 Introducing contradictions, or competing models of how things should be, is means of undermining legitimacy (cf. Clemens & Cook, 1999; Mann, 1986).  However, when those atop the hierarchy control access to the media, information that challenges the status quo is relatively less likely to come to light, whereas hierarchy sustaining myths are likely to be commonplace.  For example, as Herbert Gans (2005, 2011) argued, traditional top-down news formats privilege particular views with the consequence that what gets covered is a sliver of the actual happenings.  It also allows news-givers to frame the information and thus influence the meaning that viewers give it.   As I discussed above, the Internet provides easy access to virtual free spaces in which minority perspectives can be cultivated.  The Internet also allows for those perspectives to be encoded in any form of media and communicated quickly and broadly through many-to-many-communication channels.  The inflow of alternative perspectives introduces new ideas—new possibilities—about the legitimacy of the status quo hierarchy and acts that sustain it.  If there are alternatives, then the status quo is not inherently inevitable. “Revolution becomes possible once institutions, however fragile or robust, are no longer perceived as inevitable” (Clemens & Cook, 1999, p. 449). Moreover, because of challenge acts that occur online, illegitimating perspectives may also enter traditional media streams, and it is likely that much of the potency of online action against the status quo lies in this bottom-up directing of “news.”  Whether or not actors explicitly endorse the perspectives embedded in a message, through the very act of clicking a link or passing along a viral message, they implicitly endorse it and enhance its credibility (Harvey, Stewart, & Ewing, 2011).  When a message “gets enough clicks” —click credibility—26 it can become a “legitimate” news story.  Prior to the Internet, news about hierarchal abuses disseminated in a mainstream media outlet would have been framed by that outlet and, thus, most likely be stilted toward maintaining the status quo (Entman, 2007).  Although activists might have reframed the information to drum up support, the initial (media-given) frame would have been advantaged in relation to the activist’s (secondary) frame (Druckman, Fein, & Leeper, 2012; C. Ryan & Gamson, 2006).  Through the Internet, the message may be first framed by lower-power actors and given legitimacy through perceived popular sentiment (i.e., “click power”) before entering the mainstream media, perhaps even with the Internet-mediated frame intact (e.g., Bernoff & Schadler, 2010).  Implications of this are likely significant:  Druckman et al. (2012) found that early frames on political issues are strongly favored, with people tending to dogmatically adhere to the opinions they form in response to the first frame to which they are exposed. Finally, the Internet allows communication among many “friends” (i.e., online social network ties) and such interactions may result in people who otherwise would have viewed a hierarchy-sustaining act through system justifying glasses as an illegitimate act—as a grievance—instead.  Barley (1991) found that third parties can sometimes convince victims that outcomes they had first perceived as bad luck were actually human rights violations, for example.  Goldman (2001) similarly showed that friends, family members, and colleagues significantly influenced layoff victims’ decisions to file legal claims against the companies who had dismissed them.  Concordantly, motivated Facebook friends, for example, might be able to engage in strategic communications to sway others to share their perspective that a power laden act was illegitimate.    27 1.6 The Internet and Social Action One way in which people can act in such a way as to resist or alter established socio-organizational powers structures is to engage in social action.  Social action is activity undertaken by individuals for a collective purpose, such as the advancement of, or resistance to, a particular cause, ideology, or idea that affects a society or one or more groups within a society (cf. Brunsting & Postmes, 2002).  This topic has long been a central focus in social science research (e.g., Klandermans, 1984a; McAdam, 1986; Merton, 1936; Olson, 1965; Snow, Zurcher, & Sheldon, 1980), and as a result, we know much about social action in its various forms—advocacy, activism, volunteerism, social movement participation, etc. However, society has changed dramatically in the time since much of the seminal social action research occurred - particularly in regards to advancements in inexpensive communication technologies that can enable coordination of efforts.  As the Internet has become increasingly woven into everyday life, much has been made of its potential social implications (DiMaggio et al., 2001), with some heralding the Internet as the great social equalizer (e.g., Shirky, 2008), others holding a more ambivalent view (e.g., Gamson, 1996; Morozov, 2011), and still others bemoaning it as the invisible shears shredding knowledge, belief, culture, community, and other fabrics of society (e.g., Arnold, 2013; Keen, 2007). This much is clear: Context can play a pivotal role in behavioral decision-making (Johns, 2006), and in the modern era, people’s everyday existences are increasingly embedded in the context of cyberspace (Internet World Stats, 2012; Raine, Purcell, & Smith, 2011; Zickuhr & Madden, 2012b).  28 Because of the widespread use of Internet-connected devices and increasing focus on low cost, high-bandwidth wired and wireless Internet availability (Davidson & Lieberman, 2010), people are engaging with one another as never before. For example, there are 1.06 billion monthly active users of the social networking website Facebook, with nearly 60 percent of these being active daily users (Tam, 2013), and Twitter users Tweet 3.5 billion messages in an average week (Holt, 2013). The trend in Internet-mediated social connectivity is likely to continue into the foreseeable future; the technology firm Cisco predicts that 50 billion devices (e.g., cellular phones, laptops, tablet computers, etc.) will in use by the year 2020 and financial firm Morgan Stanley estimates that number could actually be as high as 75 billion (Danova, 2013; Warman, 2012).  One of the ways connective technology may affect the world is through its impacts on social- and political action participation. For example, an analysis of a US national data set by Gil de Zúñiga, Jung, and Valenzuela (2012) showed that actively seeking information via social network sites was positively predictive of social capital and both civic and political action participation (both offline and on-), after controlling for factors such as political knowledge and efficacy, size and frequency of interaction with political discussion networks, and media consumption.  In somewhat the same vein, Shah and co-authors (Shah, Cho, Eveland, & Kwak, 2005) analyzed data from a two-wave US national panel survey to predict (offline) civic engagement, or the frequency with which respondents “did volunteer work, went to a club meeting, worked on a community project, went to a community or neighborhood meeting, and worked on behalf of a social group or cause” (p. 540).  They found that respondents’ online political information seeking (which they characterize as use 29 of the Internet as a resource) and interactive civic messaging (use of the Internet as a forum) both strongly influenced their subsequent offline civic engagement.  Further work (i.e., Shah et al., 2007) showed that campaign ads promoted online political information seeking and interactive political messaging, and whereas ads emphasizing “attack” messages discouraged offline political information seeking (i.e., via print and broadcast media), this was not the case for online political information seeking.  The Internet also has broadened what Tilly (1995) and Tarrow (1998) termed the contentious repertoire, creating new mechanisms through which the less powerful assert their claims and/or pursue shared interests. Some websites are established with the sole purpose of serving as a forum for grievance claims against corporations or brands (see, Hollenbeck & Zinkhan, 2006). Others routinely host or directly link to online petitions, or e-petitions, boycotts, and email- letter writing campaigns (Earl, 2006). Tens- and even hundreds- of thousands of people join Facebook groups centered on specific social issues (e.g., Facebook.com, 2013; Facebook.com, 2014) and the Tweets of everyday individuals contribute to social change, even in relatively tightly controlled China (Kuhn, 2009). Moreover, it is not unusual- for a single e-petition to garner 1,000,000 or more signatures (e.g., AAVAZ.org, 2013; Hawkins, 2011; T. Martin & Fulton, 2012).5                                                    5 For further examples of the Internet’s additions to the contentious repertoire, see Van Laer and Van Aelst’s (2010) online social action typology and/or Earl, Kimport, Prieto, Rush, and Reynoso’s (2010) internet activism typology. 30 The efficacy of connective technology to promote change has been a topic of debate (e.g., Gladwell, 2010; Hindman, 2009; K. Lewis, Gray, & Meierhenrich, 2014), and some authors have applied the moniker slacktivism to much of the online contentious repertoire. This scarlet S reflects a judgment about online action being too easy and not demanding long-term commitment. It also acts to set actions like joining social cause pages on Facebook, Tweeting about an issue, or signing an e-petitions apart from the ostensibly more desirable, “real” social actions in which one might engage (see, e.g. Christensen, 2011; Jutras, 2009; Leonard, 2009; Morozov, 2009).  However, 83 percent of the respondents to the Pew Internet and American life poll (n = 2303) believed that the Internet affects the impact of collective action on society, and 59 percent believed that the Internet has a “major impact” on that outcome (Raine et al., 2011).   Indeed, Twitter allows collective action participants and observers to be on-the-ground reporters who relay updated information outside the locus of action in real-time (e.g., Veenstra, Iyer, Hossain, & Park, 2014), for example, and has given a “voice to imprisoned journalists in Egypt and fueled a rallying cry for users to donate money for relief efforts in Haiti” (Kang, 2010, p. 1). Facebook has been credited as an essential resource for Barack Obama and his supporters in his successful bid to become the first African American elected president of the United States (Fraser & Dutta, 2008), and the Internet was a “key tool” (Wingrove, 2010) facilitating the election of Canada’s first Muslim mayor, Naheed Nenshi.  Activism through social media and e-petitions also has exerted important influence in political struggles and directly led to policy changes by the US government (Lee, 2013; Vogt, 2013).  And, few would disagree that Edward Snowden’s hacking and subsequent 31 dissemination of information about the spying efforts of the US National Security Administration (Butler, 2013; A. West, 2013) was a significant resistance act with far reaching effects.  Although “empirical studies of online activism are surprisingly scarce” (K. Lewis et al., 2014, p. 1) overall, the potential for the Internet to facilitate social action and precipitate change is also underscored in academic literature.  For example, Diani (2000) argued that connectivity technologies may allow for greater leveraging of existing bonds and solidarities and thus permit more effective mobilization than would otherwise be possible.  Further, based on their review of the literature, Van Laer and Van Aelst (2010) concluded that, although not without limitations, the Internet has been a boon for social movements in that it has provided “new and improved opportunities to engage in social and political action” (p. 1146).  Harlow and Harp (2012) surveyed participants in the US and Latin America to investigate the role of social networking sites (e.g., Facebook) in social action mobilization.  They found not only that the people of both regions used social networking sites to mobilize supporters for both online and offline forms of social action but also that regardless of whether respondents’ social action took place mostly online or offline or in both contexts equally, their offline social action participation was no different (read as: “slacktivism” does not necessarily supplant “activism”).  Yang (2013) analyzed over a decade’s worth of data on social movements, voluntary organizing, and politics in China relating to both struggles for recognition and struggles against exploitation and oppression and concluded that through online activism, the Chinese people have “transformed personhood, society, and politics” 32 within that country (p. 1).  Moreover, Kristofferson, White, and Peloza (2014) showed that in some circumstances, so-called slactivistic online social action served as an initial investment that lead to subsequent “meaningful support” for a cause (p. 1149).   Some research also has focused on whether and how the Internet affects social action decision-making more explicitly.  Van Zomeren (2013) overviewed “four core social‐psychological motivations to undertake collective action” (p. 378). They were collective efficacy, which comprises reasoned or expectancy-based considerations; collective identity; emotion, which encompasses such feelings as injustice and anger; and morality.  Alberici and Milesi, (2013) administered surveys to a sample of participants at two (offline) demonstrations in Italy, both of which had been organized primarily through online communication, to assess if and how the effect of these four motivations differed, depending on the extent to which activists engaged in online political discussions.  In that data, efficacy and morality predicted greater intention to attend a future (offline) demonstration among higher (vs. lower) frequency online discussers.  Politicized (i.e., shared) identity had a significant effect only among higher frequency discussers, whereas anger had no effect on demonstration attendance intention among them.   Brunsting and Postmes conducted a study, reported in two articles (i.e., Brunsting & Postmes, 2002; Postmes & Brunsting, 2002), in which they surveyed activists and sympathizers associated with a large environmental organization in the Netherlands, as well as nonactivists, to characterize and compare motives for online and offline social movement 33 participation.6  They predicted that the relative weight of factors predicting collective action participation would vary for online versus offline collective action.  More specifically, they predicted that identification with the cause (which they include as part of a broader class of “affective factors”) would play a more central role in offline social action, with cognitive factors (e.g., cost/benefit analyses) playing a lesser role.  Conversely, they predicted that pattern would be reversed for online social action.  In advancing those predictions, the authors drew from a number of paradigms, including relative deprivation theory (Runciman, 1966), social identity theory (Tajfel & Turner, 1986; Van Zomeren, Postmes, & Spears, 2008), the theory of reasoned action (Ajzen, 1985, 1991), and, most predominantly (and in the case of Postmes & Brunsting, 2002, exclusively), the expectancy value model of collective action (Klandermans, 1984a, 1984b).   The social action intentions investigated in that research related to participation in demonstrations and blockades (“hard” or conflictive action); letter writing and petition signing (“soft” or persuasive action); and their online analogues.  In general, for these acts, cognitive predictors were found to matter proportionally more and movement identification proportionally less in predicting online social action intentions, as had been predicted. Moreover, peripheral members of the movement were found to be more easily recruited for online action participation.                                                   6 Postmes (2007) later delineated three core (socio-)psychological motivations for collective action—a sense of injustice, a sense of efficacy, and a sense of collective identity—and discussed some ways in which each might be affected by the Internet.   34 I further unpack when and how the Internet should affect social action decision-making, below.  Specifically, successful challenges often rely not only on action by those directly affected by a cause but also on action by those who are not (e.g., Downs, 1957; Mueller, 1979; Olson, 1965).  However, little is yet known about why decisions to engage in social action may differ between online versus offline contexts for those who are not directly affected by the cause but who might prove to be sympathetic allies.  This is especially true for social action that carries significant social risks, such as advocating a socially stigmatized issue to those with whom one has social ties (cf. writing a letter to one’s government representative advocating for an environmental issue, as in Postmes and Brinsting’s 2002 articles).  In addressing this gap, I consider how decisions by the same individual to advocate an issue may differ online versus offline, not only because of technology’s potential direct effects but also because of its potential higher order effects (i.e., as a social context).  I explicitly take into account whether or not that individual is directly affected by the focal issue, as well as factors related to the issue itself.  As research in traditional contexts has shown, although people are often unwilling to engage in social action in general, that aversion can be especially strong for those who are not at risk of being directly affected by the issue at hand.  Ratner and Miller (2001) suggested that this disparity in willingness to act between individuals who are members of the group that is affected by an issue (i.e., those with vested interest in that issue) and those who are not (i.e., those without vested interest) arises from a prevalent North American social norm for people’s actions to be, first and foremost, self-interested. In keeping with that hypothesis, they found that those who were not members of a social group affected by a social issue (i.e., 35 lacking vested interest) were significantly less likely to take action on the issue, even among those who were sympathetic to the cause. Further, they found that this aversion to action arose from a fear of social punishment for engaging in non-self-benefitting action. Given this, coupled with the general tendency for people to form social connections with others who are much like themselves (i.e., form homophilous ties) (McPherson, Smith-Lovin, & Cook, 2001), crossing the vestedness barrier to forge alliances with members of non-vested groups who are willing to take action has been difficult, historically (Ibarra, 1992).  However, because of the aforementioned changes in the way that people network and communicate with one another, there are undoubtedly more social connections between more people in more places who are communicating with each other in more ways today than at any other point in history.  As described previously, differential access results in the more powerful being disproportionately present in online contexts, but those on the lower rungs of social hierarchies are increasingly present. Within North American, for example, from 2000 to 2011 the prevalence of Internet access among low-income Americans more than doubled, such that by 2011 nearly two-thirds were going online (Zickuhr & Smith, 2012), while the Internet access rate for Canadians in the lowest quartile of household income rose from 58.7 percent in 2005 to 76.2 percent in 2009 (Statistics Canada, 2010).     Those acting on social causes in online contexts do not always have vested interest in those causes, and likely are only connected by weak network ties to the effort. As an example, the more than 170,000 people who signed a petition calling on the South African government to stop the occurrences of “corrective rape” of its lesbian citizens lived in over 160 different countries (Change.org, 2011; Reno, 2013). Signatories not living in South 36 Africa acted on the cause despite not having vested interest in it, as did any non-lesbians signatories living in South Africa. Although people lacking vested interest in a cause undoubtedly sometimes act for the benefit of that cause in offline contexts as well, the sheer volume and ubiquity of prosocial action participation in online contexts suggests that important new insights may be gained by examining social action in the context of the newly evolving mediums for social (inter)action that comprise online environments.  There are several reasons that third-party action decisions may differ in online contexts as opposed to traditional contexts.  People are attentive to social perceptions of their actions when they act publicly on social issues (White & Peloza, 2009), and one of the primary reasons that those not affected by an issue do not act is that they anticipate social punishment if they do (Ratner & Miller, 2001).  However, the Internet likely provides free space where higher power sympathizers can act in ways that are likely to go unnoticed by other high power people, just as it provides means for lower power actors to escape detection.  Further, as will be discussed in more detail in Section 1.7, the Internet positions free spaces and nexuses of action at geographical distance from actors, which should reduce potential social detriments within home and local community.  The net effect of this should be to reduce the importance of vestedness—direct interest—in a social issue on decisions to engage in action against the status quo. Another reason that third parties may not act, even when they believe that they should is that their motivation and intention to act is weak, relative to that of a vested party. Because of this comparatively weak motivation, intervening forces within would-be actors’ daily lives would be prone to derail them from ultimately acting on their beliefs (Ajzen, 1991; 37 Weber & Gillespie, 1998).  However, through Internet technology, people can act immediately, signing an electronic petition, sharing a story on their Facebook walls or through a Tweet, emailing their lawmakers, making a donation electronically, and so forth.  As such, the Internet should increase the tightness of coupling between beliefs and ultimate action by shortening the process cycle between motivation and action and thereby reduce opportunities for exogenous forces to interfere.     Finally, third parties should be more likely to act in online contexts because the Internet allows for the easy and inexpensive sharing of photos, videos, and screen captures.  Power holders, power holding third parties in this case, tend to experience action-oriented emotions in response to injustice, which spurs them to action (Keltner, Ellsworth, & Edwards, 1993; Tiedens, Ellsworth, & Mesquita, 2000).  The Internet makes the sharing of hard evidence of injustice—photos, screen captures, videos—easier and cheaper than ever before, and because these communications are disseminated virally, their reach is broad. Moreover, emotional arousal is expected to encourage the emergence of challenges from free space interactions (Rao & Dutta, 2012), and these types of media, and video in particular, likely have the greatest capacity to elicit the strongest emotional reactions and thus encourage third parties to act, perhaps even despite social repercussions that may come from doing so.  As Internet-mediated social connectedness and communication grow ever more ubiquitous, online social action is likely to become increasingly prevalent.  Thus, it is important to understand the capacity for online context to affect willingness to act, both for those who are vested in issues and those who are not. Like Brunsting and Postmes 38 (Brunsting & Postmes, 2002; Postmes & Brunsting, 2002) in their online versus offline social action decision making research, I focus on rational determinants of social action participation decisions (see also, Feather & Newton, 1982; Klandermans, 1984a).  In essence, as rational beings, people weigh the likely benefits against the known costs and anticipated negative consequences of the activities in which they might engage (Ajzen, 1991; Beach & Mitchell, 1978; Cabantous, Gond, & Johnson-Cramer, 2010), including activities with social implications (Klandermans, 1984a).  Specifically, when the predicted costs and anticipated negative consequences of an activity outweigh predicted benefits and positive consequences, people are less likely to act than when that balance is inverted.  It might be for this reason that people sometimes do not act as they believe they should (cf. J. Weber & Gillespie, 1998).  The benefits or positive consequences one might reap from engaging in social action are diverse.  They may relate to the actor’s direct tangible/financial, social, affective and/or psychological gains from undertaking social action, to the success of the broader social movement to which the actor’s action is a part, or both (Feather & Newton, 1982; Klandermans, 1984a, 1984b).  Conversely, costs or negative consequences are those things the actor expends, surrenders, loses, forgoes, or consumes—money, energy, time, happiness, social esteem, etc.—and anything experienced negatively—pain, fatigue, etc.—in order to participate (Feather & Newton, 1982; McAdam, 1986).  Factors likely to raise the perceived benefits of social action or reduce its costs should increase willingness to act; factors likely to reduce perceived benefits or raise the costs of participation should decrease it.   39 Virtualized context should affect willingness to engage in social action in part by reducing the direct costs of participation.  An online social act likely requires the expenditure of less time, effort, and money than a comparable offline social act, which is to say that, all else equal, online action should be less prohibitive than functionally similar offline action. Compare, for example, the time, energy, and expense involved in driving even across town to attend a community meeting to discuss a social issue to that involved in participating in an online community discussion of that same issue.  People should thus be more willing to engage in social actions in online contexts than in offline contexts, all else equal.  However, virtualized context is also likely to have higher-order implications for social action decision-making.  It is widely accepted that contexts affect decisions, behavior, and outcomes in social contexts (Johns, 2006; Payne, 1982), and calculations of the likely costs and benefits of advocacy are likely to differ between online and offline contexts.  For example, people have been shown to evaluate risks less severely in online contexts (Debatin, Lovejoy, Horn, & Hughes, 2009; Drennan & Previte, 2006).  Congruent with this, online versus offline context should alter the influence of the other willingness-to-act predictors.  1.7 The Internet and Risk of Abusing Power Because the Internet has the capacity to mediate challenge acts by those low in power and by their higher power supporters, it also has the capacity to deter those higher in power from engaging in abuses of power that perpetuate the hierarchy that favors them, especially when that hierarchy-sustaining act would occur in public (i.e., not under conditions of anonymity or free space protection). As I delineated above, the Internet should make action to countervail hierarchy-sustaining more likely than before and thus make engaging in such 40 behaviors more risky in general. Beyond that, though, the Internet creates an additional source of risk for engaging in hierarchy-sustaining acts, especially those that are particularly grievous. Specifically, when blatant subordinating acts occur, they may be made public online, and information about both the acts and the entities or persons perpetrating them can spread virally through C-forms (Seidel & Stewart, 2011) such as Facebook, Twitter, or YouTube. That information can, in turn, motivate an aggressive collective response by the victim, other members of the victim’s social group, and/or their allies.  As the opening case illustrated (Section 1.1), Internet-mediated communication and social connectivity can facilitate the flow of information about abusive hierarchical acts in a given locale to the outside world.  Once members of the lower powered group involved and their higher-powered allies become aware of such an act, the Internet also provides an efficient means through which they can coordinate their response.  Through coordination, actors can amplify their combined power beyond a simple aggregate (Arendt, 1970), and in cases in which hierarchical acts are especially grievous, countervailing responses—crowd-sourced punishment—can be so frenzied as to leave the more powerful perpetrator-cum-target with “no place to hide” (Chee-Sing, 2011, p. 4).  That might have been the feeling of Mr. McCance, who described being inundated with threats and harsh words at home and at work and fearing for the safety of his family (CNN Wire Staff, 2010b). A rich body of literature demonstrates that risk of punishment acts to deter negative behavior, both in organizational contexts (e.g., Hollinger & Clark, 1982) and in society at large (e.g., deterring crime: Shover, 1996).  Greater risks act as greater deterrents (e.g., Taylor, Walton, & Young, 1973).  Thus, increased frequency and veracity of countervailing 41 action as mediated by the Internet should raise perceived risks of engaging in (illegitimate) hierarchy-sustaining acts.   Moreover, to the extent that Internet-mediated countervailing action involves large numbers of actors and/or is communicated broadly through social media, it should act to reduce the perceived legitimacy of future hierarchy-sustaining acts than would offline, largely localized countervailing action.  1.8 Free Space – Construct Refinement and Extension into Virtual Contexts As I described above, on the surface, it might seem that challenges to establish power structures that favor some over others and attempts to countervail the actions that reinforce those structures would abound, with or without electronic communication and coordination technologies.  That is not the case, however.  There exists a generally pervasive aversion among human beings to challenge the more powerful among us, and challenges to established power structures are rare, even (and especially) among those who would stand to gain the most.  For change to one of these discriminatory power structures to be tenable through organized resistance, non-elites must have access to spaces where they can act outside the normal confines pervading the social system—spaces that provide insulation, at least temporarily, from the threats and cultural programming that otherwise entangle them (Gamson, 1996).  There, subordinated voices can be heard, stigmatizing identities expressed, and/or dissent vented away from the “on-stage power of the dominant” (Scott, 1990, p. 115) but also outside the confines of hearth and home (Evans & Boyte, 1986). Such spaces are believed to facilitate ongoing interactions among those who share a subordinated voice, and from them, common cause, emotional empowerment, and collective action frames may arise 42 to fuel collective resistance (Gamson, 1996; Rao & Dutta, 2012). Although protected spaces show promise as important “weapons of the weak” for spurring change through resistance efforts (Rao & Dutta, 2012, p. 625), heretofore the existence and nature of these spaces have been assumed, described, or approximated, because no measure of them exists. As such, conceptualizations of the construct vary widely. In this paper, I refine the construct, clarify its boundaries, and validate a universal measure of it. The construct I am describing has been called by several names (see Polletta, 1999, for a review). Sara Evans and Harry C. Boyte used the term free space in seminal works in this domain (Evans, 1979; Evans & Boyte, 1986): “Put simply, free spaces are settings between private lives and large scale institutions where ordinary citizens can act with dignity, independence, and vision” (Evans & Boyte, 1986, p. 17). This term has been used often by scholars (Polletta, 1999), and I also use it here.  To date, free space has been discussed primarily in physical contexts such as religious festivals, communal gatherings, block clubs, cooperatives, taverns, or movement half-way houses (Rao & Dutta, 2012), but today free spaces may also exist within virtual, online environments, with interactions taking place through computer mediated communication (CMC) (for one recent overview of CMC technologies, see Herring, Stein, & Virtanen, 2013). Gamson pointed out early on in the Internet era (1996) that people interact in a variety of “cyberspaces” online, and because each of those spaces operates with its own set of rules for access and participation, some are assuredly more “free” than are others. Rao and Dutta (2012) suggested that social media platforms such as Facebook and Twitter, specifically, may serve as large-scale free spaces. Other virtualized interaction contexts such 43 as online communities, issue-specific websites, discussion groups, blogs, posting boards, online gaming platforms, and social virtual worlds such as Second Life (cf. Saunders, Rutkowski, van Genuchten, Vogel, & Orrego, 2011) may as well. As I will discuss below, online free spaces should serve a similar function as offline free spaces but may also be less constrained because they lower geographic boundaries and allow for mass participation. The growing ubiquity of Internet-mediated social interaction brings with it a tremendous volume and diversity of spaces in which people can come together. Thus, clearly understanding free space and being able to identify and characterize it a priori should be increasingly important in predicting change. In pursuit of those ends, I seek to theoretically clarify this important construct and delineate its defining characteristics. In specifying my conceptualization of free space, I first establish stronger conceptual boundaries of what free space is and is not. I then discuss free space in virtualized, online contexts. Next, I consider the issue specificity and temporal variability of free spaces. Thereafter, I describe the range of risks (i.e., threats from control mechanisms) from which free spaces must provide protection. Finally, I consider the role of the individual as perceiver of the freedom for resistance offered by a given social interaction context.  Throughout, I advance specific hypotheses that will be tested in three studies that will be reported in Chapter 2 of this thesis. 1.8.1 What Free Space Is and Is Not I define a free space as a distinct social setting that provides protection from formal controls, cultural norms and practices, and other mechanisms of subordination implicit to socio-organizational power hierarchies that otherwise dissuade non-elites from expressing 44 their identities, ideologies, interests, or concerns. Although this definition is informed by previous work (e.g., Evans & Boyte, 1986; Fantasia & Hirsch, 1995; Rao & Dutta, 2012), my conceptualization is more tightly bounded than that suggested in some other works. Specifically, descriptions of the free space construct have often been explicitly or implicitly conflated with some of its potential outcomes and intermediary processes that might occur within it.  To date, discussions of free spaces and the limited empirical works incorporating the construct have largely identified free spaces retrospectively, based on their output (e.g., revolts, social movements, activist action), and perhaps as a result, outcomes sometimes have been incorporated into the construct itself (see Polletta, 1999). However, as the organizational ecology literature (for one review, see Carroll & Khessina, 2005) holds, conditions within a given resource space can be such that it becomes “at risk” of birthing attempts by actors in that space to come together in an attempt to leverage available resources to achieve common goals, but the conditions that make a space “at risk” do not guarantee that founding attempts will be made or that those which are made will be successful (Hannan & Freeman, 1987; Lomi, 1995; Wade, Swaminathan, & Saxon, 1998). Specifically, the conditions within a resource space can be such that it is “at risk” of birthing organizations, whereas those organizations otherwise would be less likely to arise (e.g., Carroll, 1985; Hannan & Freeman, 1987). Or, said differently, conditions within that space can be such that it is especially conducive to the founding of one or more organizations formed with the intent of leveraging available resources to achieve goals, as when high ethnic identity among a space’s members spurs the founding of ethnic organizations, whereas high 45 social polarization makes such foundings comparatively undesirable or even dangerous (E. West, 1995).  From this perspective, social spaces may be, to greater or lesser extents, “at risk” of birthing conglomerated efforts to leverage available resources to achieve resistance goals. These goals might include such things as advancing the standing of those low in social power, challenging the subjugating acts of those high in social power, and/or otherwise promoting legitimacy and fairness in hierarchies. Under my conceptualization of the construct, free spaces should be more conducive to birthing action to promote such goals than are more contested spaces, but this action need neither succeed nor even occur in order for free space to exist. Some free spaces are likely to be potential incubators of such action, wombs from which fairness-directed change efforts might emerge. However, they must not be identified based upon the existence or success of resistance, neither by individuals nor through collective action.  Similarly, the free space construct should not be contaminated with potential intermediating processes and structures that might link them to resistance efforts. A few examples serve to illustrate: In their seminal work on free space, Evans and Boyte (1986) discussed the importance of shared perceptions of community within free space, Polletta (1999) highlights the importance of the nature of social ties within free space, and Rao and Dutta (2012) emphasized the role of emotional contagion. I do not disagree that free spaces populated by actors with shared grievances should be conducive to the establishment of communal feelings and that structures of association are meaningful predictors of mobilization, nor do I expect that emotional contagion will not often occur within free 46 space. In some cases, in fact, these potential intermediaries almost assuredly influence the birth of resistance movements. But to including these in the construct conflates the ‘what’ with the ‘how’ and the ‘why.’  To put this in terms of physical science, one might think of the relationship between free spaces and collective challenges as a chemical reaction. Chemical reactions are not possible without reactants—the materials from which the reactions may arise (Atkins & De Paula, 2006), and the speed at which they occur is dependent on reactant concentrations.  When reactants are more concentrated, they have increased interactions per unit of time, and reactions are accelerated (Atkins & De Paula, 2006; Avery, 1974).  Subordinated ‘reactants’ may be pushed towards lower pressure areas (i.e., free spaces) when pressure is high elsewhere in their atmosphere (as through pressure gradient force) (cf. Jacob, 2011), increasing their concentration in the lower pressure environment.  The presence of ‘reactants’ within a free space should make the production of collective change agents possible and such occurrences should be hastened by greater concentrations of ‘reactants’ within that environment.  Further, in chemical reactions, reactants combine with one another to yield products that have properties different from the reactants themselves (Atkins & De Paula, 2006). Individual subordinated ‘reactants’ may well lack the resources, confidence, or acumen to engage in challenge on their own, but through interactions with other ‘reactants’ within the free space beaker, a synthesized product with those characteristics may arise.  Moreover, a chemical reaction may occur in steps, yielding intermediary products that facilitate the production of the ultimate product (Atkins and De Paula, 2006). In this same way, things 47 like communal feelings and emotional contagion likely represent intermediary products in the overall process yielding collective challenge agents.  Extending this analogy even more, the atmosphere outside of the free space beaker is also likely to affect the reactions that occur within it. When heat is applied to a chemical reaction environment, for example, reactants move faster and interact with one another more frequently; this encourages the reaction (Atkins & De Paula, 2012). Moreover, if too much heat is applied for too long, a runaway reaction may occur, and this runaway reaction may have explosive consequences, especially if there is no way for the pressure exerted on the reactants to escape into the outside environment (cf. Stoessel, 2008). Similarly, when ‘heat’ is exerted on the subordinated ‘reactants’ populating a free space, resistance may more swiftly arise.  In cases of extreme ‘heat’ (e.g., persecution, prosecution) and/or when benign tactics fail to relieve pressure and incremental change is untenable, this may trigger more aggressive and potentially explosive outcomes.   Finally, the reactants necessary for a reaction to occur may be present within a chemist’s beaker, but if the threshold for activation energy (i.e., energy required for a reaction to start and carry on spontaneously) is not achieved, a reaction will not occur until the proper catalyst is added to trigger it (Spencer, Bodner, & Rickard, 2010).  Likewise, a free space may be populated with subordinated ‘reactants’ but not yield collective resistance absent a catalyst. However, just because a catalyst has not yet been added does not mean that the appropriate ‘beaker’ does not exist or that the necessary ‘reactants’ are not present within it. By identifying the beakers—free spaces—organizational and social scientists will be able to observe which catalyst(s) ultimately precipitate reactions as well as the conditions inside and 48 outside the space that tend to encourage those outcomes, hinder them, or make them more or less severe.  1.8.2 Free Space in Online Contexts Rao and Dutta (2012) reported a study in which they demonstrate the capacity for free spaces to promote resistance that leads to meaningful change. The context for their study was mutinies within nineteenth-century Bengal armies and the free spaces were communal gatherings and processions. However, today’s electronic meeting and interaction spaces should have the capacity to serve theoretically similar, yet less constrained, roles. The defining features of free space outlined above have been discussed almost exclusively in the context of physical places in past research (cf.  Gamson, 1996), but they should be equally applicable within virtualized social interaction spaces. Beyond that, though, access- and resource- barriers that might otherwise impede communication and coordination can be dramatically lower in online contexts, and there are virtually no geographic barriers to online interaction, whereas interaction within physical space is subject to geographic constraint (Seidel & Stewart, 2011). Thus, characteristics that define free space may exert differential influence, depending on whether it is embedded in a physical place or is virtualized, online.  Virtualized free spaces should have the capacity to serve as meeting arenas where social movements can be spawned and activated (cf. Haug, 2013). Through multiplex CMC technologies, actors can both share grievances (i.e., Antony & Thomas, 2010) and have rich and meaningful interactions, without the requirement that they be physically present (see, e.g., Bargh & McKenna, 2004; Preece & Maloney-Krichmar, 2003). By removing geographic constraints on participation, the Internet allows for large numbers of participants to join in 49 combined efforts (Howe, 2008; Seidel & Stewart, 2011). Leadership for these combined efforts by geographically dispersed actors can then emerge organically based on merit and ability (O'Mahony, 2007; O'Mahony & Bechky, 2008). Moreover, virtualized free spaces should provide protections for the less powerful beyond those available within physically-embedded ones. Failed resistance brings the risk of punishment to those who have taken part, and fear of this can dissuade participation (Mason, 1996). Large numbers of participants can reduce perceived risk of acting by raising expectations that a sufficient number of actors can be roused for challenge acts to be successful, rather than failing and triggering punishment (Oliver, 1989). Additionally, large numbers and geographic actor dispersion should also reduce the risk of punishment to any individual actor if resistance efforts fail by raising transaction costs for targets who might wish to identify, locate, and punish these far-flung actors, as compared to if the free space was less populated and in a bounded, physical location adjacent to the locus of resistance.  To the extent that these characteristics of virtualized free space decrease action constraints that non-elites would otherwise face, non-elites should experience levels of empowerment that are in contrast with their everyday experiences (cf. Anderson & Berdahl, 2002; Keltner et al., 2003). Such experiences have been shown to decrease the salience of risk and other goal impediments in decision making (Whitson et al., 2013); increase optimism and promote risk-taking in pursuit of goals (Anderson & Galinsky, 2006); and decrease behavioral inhibition (Anderson & Berdahl, 2002; Hirsh, Galinsky, & Zhong, 2011; Keltner et al., 2003). In short, feeling powerful should encourage resistance (Martorana et al., 50 2005), and virtualized free spaces should make people feel even more powerful than traditional free spaces, all else equal. 1.8.3 Free for X, Free for Y?: Topical Specificity of Free Space As delineated above, non-elites tend to be highly sensitive to the risk of punishment and this can discourage them from acts that challenge the status quo, even when change would be to their benefit (Anderson & Berdahl, 2002; Fiske, 1993; Keltner et al., 2003; Magee & Galinsky, 2008). People who feel motivated to undertake resistance related to a given social issue should be less apprehensive about doing so within a context that they view as offering protection from harms that might otherwise arise from their actions. Moreover, the greater the extent to which a space is seen as free in relation to the general society in which it is imbedded, the greater should be their willingness to act there, all else equal—“Those in low power may feel the disposition to act against others when they have a sense of power and especially when that sense of power is in sharp contrast to actual levels of hierarchical power and control” (Martorana et al., 2005, p. 299; see also Anderson & Spataro, 2005).  However, social context that serves as a free space in relation to one social issue or cause may not for another. For example, Polletta (1999) described one oft-mentioned example of free space: The Southern black church, removed from white control and central to life of black communities, figures in most surveys of free spaces. For the emerging civil rights movement it provided meeting spaces to develop strategy and commitment, a 51 network of charismatic movement leaders, and an idiom that persuasively joined Constitutional ideals with Christian ones. (p. 4) Some have likened the lesbian, gay, bisexual, and transgendered (LGBT) rights movement to the civil rights movement, but would the free space described by Polletta also have served as a free space for LGBT voices? Evidence suggests not. For example, the Coalition of African American Pastors strongly objects to the comparison between the civil rights movement and the gay rights movement even being made, describing LGBT activists as having “hijacked” the civil rights movement (see, e.g., Lee-St. John, 2005). Additionally, based on analysis of data from 31 studies from 1973 to 2002, Lewis (2003, p. 75) concluded, “Blacks disapprove of homosexuality more strongly than whites. Even in the most recent survey years, nearly three-quarters of blacks say that homosexual relations are always wrong, and over one-third say that AIDS might be God’s punishment for immoral sexual behavior.” Moreover, Lewis also found that the average perception of homosexuality held by black churchgoers attending services weekly was significantly more negative than that held by black participants in general.  As I have just illustrated, it seems unlikely that the free space for one subordinated group described by Polletta in the passage above would have been equally free for another subordinated group. Perhaps today, or in the future, it would, which underscores another important point: Not only are free spaces likely to differ according to topic at a given point in time, but a given free space is likely to appear, disappear, and evolve as new members join, existing members leave, and/or the normative views of its membership evolve.  52 To the extent that a space is seen as offering protection for resistance related to a given social cause or issue at the point when one considers engaging in resistance, it should attenuate fears and encourage action, ceteris paribus. Hypothesis 1: The apprehension that a potential resister feels about acting in a given social setting is a negative function of the extent to which he/she perceives that setting as free space for the focal issue. Hypothesis 2: The likelihood that a potential resister will act in a given social setting is a positive function of the extent to which he/she perceives that setting as free space for the focal issue. 1.8.4 The Multidimensionality of Free Space: Protection from Six Risk Categories Although the general nature of free space might be readily discernable at this point, to more fully define the construct, I must expand upon the risks that make it necessary. Some examples serve to illustrate: A Bengal soldier might shy from advocating mutiny against the oppressive regime that commands him (as in Rao & Dutta, 2012) for fear of government-sanctioned punishments such as execution or imprisonment. One employee might not resist speak out against her employer for fear of being retaliated against by the organization (cf. Parmerlee, Near, & Jensen, 1982); another might not do so for fear of being labeled “whistleblower” and ostracized by his colleagues (cf. McDonald & Ahem, 2000). A woman living in the time leading to the women’s suffrage movement might not have spoken out for the rights she deserved for fear of discord with her misogynistic husband (cf. Stark, 2007); a progressive-minded man living in that same time might not voice his opposition to women being treated unequally for fear of being harangued by his friends at the lodge or 53 kicked out of the organization altogether (cf. Ratner & Miller, 2001). The lady living with HIV, knowing the strength of the social stigma against her illness, might not speak out for fear of making her status known and losing her standing as a revered member of her local community or workplace (e.g., Clair, Beatty, & MacLean, 2005); another might not do so for fear that others would inflict physical harm on her or treat her as if she were tainted and unclean (Sandelowski, Lambe, & Barroso, 2004). And, one might not join a community program to combat gang activity for fear that members of his family will be harmed as a result (cf. Healey, 1995).  As the above examples illustrate, a person who says or does things, either individually or as part of a collective that might be seen as challenging established institutions; threatening powerful actors, collectives, or organizations; threatening social order; or otherwise challenging established socio-organizational structures (i.e., engages in resistance acts) could face risks from a range of potential sources and those risks could affect them directly or indirectly, by threatening those whom they love. Consistent with this, my literature review suggests resistance risks are of six distinct types, as shown in Table 1.  The relative mix of these risks for resistance for a given social issue is also likely to vary, even for similar social interaction spaces, in accordance with the laws, norms, and cultural conventions of the broader societies embedding them. For example, an office dinner party in the United Sates might not comprise free space to voice support for equal marriage rights being extended to gay men and women throughout the world because to do so would likely elicit ostracism from coworkers (cf. Herek, 2004). If that office dinner party were in Russia, however, the additional risk of being imprisoned for violating the country’s legal 54  Table 1: Categories of Risk from which Free Space Need Provide Protection. Institutional Related to negative outcomes a person might experience as a result of the reactions of governments or their agents who observe or learn of resistance acts (e.g., law enforcement, military/paramilitary, courts or legal systems, etc.) (e.g., Machain, Clifton, & Regan, 2011; Rao & Dutta, 2012; Taylor et al., 1973).  Societal, active Related to threats of harm or acts of harm directed at a person arising from individuals or groups of individuals within that person’s community or broader society who observe or learn of resistance acts (e.g., Cooley, 1922; Ferree, 2004; Herek, 1990; Matsuda, Lawrence, & Crenshaw, 1993).   Societal, passive Related to changes in a person’s general social standing or reputation within the person’s community or broader society and/or changes in the person’s social capital arising from individuals or groups of individuals within that person’s community or broader society who observe or learn of resistance acts (e.g., Borgatti & Foster, 2003; Goffman, 1963; Herek, 2007; Scheff, 1988).   Strong ties Related to negative outcomes a person might experience as a result of the reactions of those with whom the person values close, meaningful, and valued interpersonal connections (e.g., family members, close friends, romantic partners) who observe or learn of resistance acts (e.g., Brown, Clasen, & Eicher, 1986; Ratner & Miller, 2001; Richman & Leary, 2009; Rusbult, 1980; B. R. E. Wright, Caspi, Moffitt, & Paternoster, 2004).   Professional Related to negative outcomes a person might experience as a result of the reactions of the person’s employer, business community, or business relations (e.g., supervisors, coworkers, customers, vendors/suppliers, etc.), who observe or learn of resistance acts (e.g., Bernhein, 1994; A. Smith, 1976 [1776]; B. R. E. Wright et al., 2004).   Collateral Related to a person's valued social ties being subjected to risk as a result of their association with that person and that person engaging in resistance—collateral damage (i.e., Branigan, 2012; Hunjan & Towson, 2007; cf. Goffman, 1963).    55 prohibition against “homosexual propaganda” (see, e.g., Free Press, 2013) would come into play. On the other hand, in Denmark, where same-sex marriage is legal and attitudes toward homosexuality are generally positive, there would be much less risk of any kind from voicing such sentiments, minimizing the potential for the dinner party—or any given interaction context—to be especially free. Thus, I argue, social interaction contexts comprise free space to the extent they provide protection from a range of risks that are generally present in the embedding institutional context, that these risks emanate from a variety of sources, and that the magnitude, distribution, and covariance of these various risks differ depending in part on that broader context. As a result, the capacity for a space to be free differs depending on the embedding environment. Hypothesis 3: A social setting perceived as comprising greater free space in a broader institutional context that is generally less receptive to the focal resistance issue (as reflected in, e.g., laws, regulations, social norms, social attitudes, cultural conventions) will be perceived as comprising lesser free space in a broader institutional context that is generally more receptive to it. 1.8.5 Free Space through the Eye of the Beholder I have argued that free space offers protection for non-elites—those who are subjugated, disadvantaged, or generally low in power—to express their views and grievances or otherwise engage in acts of resistance whereas they otherwise would be less likely to do so. As a final step in conceptualizing free space, I consider the role of the would-be actor who must make sense of a space and the safety it may or may not offer for resistance within 56 the broader societal context embedding it. Key here is that objective truth and perceived truth do not always equate. Regardless of one’s objective power in a socio-organizational system, the power he or she feels influences risk perceptions in potentially combative situations (Kipnis, 1976). Those who perceive of themselves as weaklings should view a societal context that is generally averse to a given social issue, for example, as harboring relatively greater risk than those who perceived of themselves as powerful. Those who perceive themselves as powerful should be less weary of the general societal context, in general.  By comparison, for the same potential free space in the same embedding societal context, those low in perceived power should view the free space as especially protective.  Hypothesis 4: Self-perceived power attenuates free space perceptions such that those high in self-perceived power will view free space embedded within a hostile institutional context as less free, as compared to those low in self-perceived power.   My hope is that my efforts in clarifying this construct, advancing theory around it, and creating a measure for it that can be customized for use in any setting, online or offline (i.e., Chapter 2) will encourage future research on free space and its role in organizational- and social- change and that such research will proceed in a more systematic and interconnected manner.  The measure itself will also facilitate further testing of the overall theory I advance through this thesis.   57 Chapter  2: Establishing a Measure of Free Space7  As I described in Chapter 1, one might think that because hierarchical structures are characterized by privilege for some and subjugation for others, they would elicit a great deal of contentiousness and that efforts to increase the fairness of those systems would abound from those whom they disadvantage. However, historically that has most often not been the case, in large part because of a generally pervasive aversion among human beings toward drawing the attention, and potential ire, of those more powerful (Martorana et al., 2005). As Rao and Dutta (2012) recently described, however, one “weapon for the weak” that enables them to combat their subordinated realities may be comprised by free space.  These spaces are expected to provide the protections necessary for subordinated voices to be heard, for stigmatizing identities to be expressed, for dissent to be vented. They may also facilitate ongoing interactions among those who share a subordinated voice and act as nurturing environments from which collective resistance may be spawned. Although free space shows promise as an important resource for the less powerful, a measure to identify it proactively is necessary for systematic investigation of its role in promoting action by those subordinated within a social system to attenuate unfairness and inequity in that system.  Thus, I conducted three studies, reported below, to develop and                                                   7 This chapter is based on collaborative work with Marc-David L. Seidel, and an abstract of it was presented at the 2014 Organization Studies Summer Conference.  A full manuscript also including portions of Chapter 1 is in preparation for the the peer review process.  58 validate such a measure.  The first of these studies comprises the item generation and refinement phase. The second comprises the specification and testing of theorized measurement model and the testing of Hypotheses 1 and 2, which were advanced in Section 1.8 of this thesis. The third entails further validation of the measure through a nomological network approach (Cronbach & Meehl, 1955), with data collected at expected free spaces in Canada, Denmark, and the United States to also facilitate cross-cultural comparisons, and the testing of Hypotheses 3 and 4 (also advanced in Section 1.8).  Because of the growing ubiquity of Internet-mediated social interaction and the tremendous volume and diversity of Internet-mediated social interaction contexts, I expect that being able to identify and characterize free spaces will become even more important going forward; as such, the measure is designed for application in both physical and virtualized social contexts. I begin by characterizing the measurement model and reporting the process through which I generated and refined the pool of items comprising it. 2.1 Conceptualization The first, and most essential, step in creating a measure for any construct is to develop a sound conceptual definition of that construct and its conceptual boundaries (Mackenzie, Podsakoff, & Podsakoff, 2011; Nunnally & Bernstein, 1994).  In Chapter 1,  I defined free space as a social setting that provides protection from formal controls, cultural norms and practices, and other mechanisms of subordination implicit to socio-organizational power hierarchies that otherwise dissuade non-elites from expressing their identities, ideologies, interests, or concerns.  In Sections 1.8.1 – 1.8.5 of that chapter, I detailed my conceptualization of the construct and its boundaries more explicitly.  In doing so, I 59 identified six categories, or dimensions, of risk for which free space needs provide protection (see Section 1.8.4, Table 1).  The name I have given the measure arising from this studies reported below is the Free Space Index.  Its acronym, FSIx, is perhaps especially apropos given that the measure is intended to assess protection from six categories of risk.  2.2 Characterizing the Measurement Model Many measurement models in social and organizational research are reflective, which is to say that they are comprised of a group of items seen as manifestations, or effects, of an underlying latent construct (Diamantopoulos & Siguaw, 2006). In reflective measures, the latent variable represents a common cause shared by all of the items, and each of the items is a reflection of that common cause (i.e., the latent construct) and are thus, to a lesser or greater degree, interchangeable (Bollen & Lennox, 1991). A scale comprised of reflective items intended to represent the free space construct—perhaps “I feel safe here to...", "Nothing bad will happen to me if I ... here", "I would feel safe to express myself here...", etc.—could be used to characterize the level of safety or protectiveness for resistance of a given social space, but that approach would sacrifice potentially important information. Specifically, as I described in Section 1.8.5, the risks from which a social space need provide protections in order to comprise free space arise from different sources and may affect the resister directly or, in the case of collateral risk, indirectly.  In some cases (e.g., an act of terrorism), resistance might trigger all of these risks, whereas in other cases (e.g., peacefully protesting workplace mistreatment), resistance might trigger only specific risk types. A reflective free space scale would tell us only the extent to which a social space is seen as safe for resistance efforts; it would do little to identify which types of risk are in place 60 within a space to prevent it from being safe for resistance. It is my intention to specify a measurement model that will do both. An alternative to reflective measurement specification, formative measurement specification has a long standing history (Blalock, 1968; Curtis & Jackson, 1962; Land, 1970) and recently has received much attention from organizational research methodologists (Bollen & Davis, 2009; Bollen & Ting, 2000; Diamantopoulos, Riefler, & Roth, 2008; Diamantopoulos & Siguaw, 2006; Diamantopoulos & Winklhofer, 2001; J. R. Edwards, 2011; Howell, Breivik, & Wilcox, 2007; Mackenzie et al., 2011; Petter, Straub, & Rai, 2007). In formative measurement models, the focal latent construct is not reflected in the items comprising its measure but, instead, caused by them. For example, overall job satisfaction (latent construct, η) could be seen as arising from satisfaction with specific facets of one’s job: the work itself (x1), pay (x2), coworkers (x3), supervisors (x4), and advancement opportunity (x5) (MacKenzie, Podsakoff, & Jarvis, 2005). In such a model, although the indicator variables may sometimes be correlated with one another, each comprises a distinct dimension of the overall construct, and together those dimensions give the latent variable its meaning (Diamantopoulos et al., 2008; J. R. Edwards, 2011). Typical examples of organizational constructs treated in this way are socio-economic status (Hauser, 1973; Hauser & Goldberger, 1971), career success (Judge & Bretz, 1994), business relationship value (Ulaga & Eggert, 2006), and quality of life (Bollen & Ting, 2000; Fayers, Hand, Bjordal, & Groenvold, 1997) (additional examples are provided in Diamantopoulos et al., 2008, Table 1).  61 A formative measurement approach corresponds well with my conceptualization of free space and its measure. Unfortunately, formative measurement is not without its drawbacks or its detractors. Edwards (2011), for example, presents criticisms arranged in six themes: dimensionality, internal consistency, identification, measurement error, construct validity, and causality; I will not duplicate his efforts here. His conclusion, however, is succinct and damning: “The shortcomings of formative measurement lead to the inexorable conclusion that formative measurement models should be abandoned” (Edwards, 2011, p. 382). Although others would assuredly disagree with that assessment, Edwards goes on to say, “Fortunately, the objectives of formative measurement models can be served by alternative models that incorporate reflective measurement and, by doing so, avoid the shortcomings of formative measurement.”  I will take this best-of-both-worlds approach with the FSIx, following the recommendations of both Edwards (2011) and Mackenzie et al. (2011) to construct a measure that has both reflective and formative components. Specifically, I will treat protection from each risk category as a dimension that combines with the others to comprise the freeness of a social space for resistance (i.e., risk categories will be formative components of the latent free space construct) and treat each risk category dimension as a latent construct that is reflected in greater than one indicator (i.e., risk categories will themselves be measured with multiple indicators, reflectively) (J. R. Edwards, 2011; Iacobucci, 2010; MacKenzie et al., 2005; Mackenzie et al., 2011). 62 2.3 Generating and Refining Item Pool Following from my planned specification for the FSIx measurement model, I treated protection from each of the six risk categories identified above as a dimension of the higher order free space construct, with each to be measured with greater than one reflective item. Based on my conceptualization of the construct and its theoretical boundaries, I generated a pool of items larger than the number I anticipated would comprise the final measure (Nunnally & Bernstein, 1994). Specifically, I generated sixty items of which no fewer than seven items were intended to correspond with each of the risk types identified above. I solicited feedback on those items from a group of professors and doctoral students at a large Canadian university who were trained in organization theory in an effort to maximize content validity—the degree to which items reflect their intended content universe (Straub, Boudreau, & Gefen, 2004)—before proceeding further.  I made minor modifications to several items as a result of their feedback.  2.4 Study 1: Assessing Item-Dimension Fit  To more formally assess content validity and inform the selection of valid items for the eventual measure, I followed the method advanced by Hinkin and Tracey (1999) and recommended by Mackenzie et al. (2011). Specifically, I constructed a matrix in which definitions of each of the six risk dimension/category were listed along the top and the 60 potential items (plus 4 repeated items to facilitate response consistency) were listed in the rows. This matrix was then presented to participants, with the order of items randomized for each participant, and they rated the extent to which each item was representative of each risk category.  63 2.4.1 Study 1 Participants A sample of 61 members of the Mechanical Turk data panel living in the United States initially served as participants in this study.  Mechanical Turk samples have been shown conducive for social science experiments (e.g., Alter, Oppenheimer, & Zemla, 2010; Ruedy, Moore, Gino, & Schweitzer, 2013) and compare favorably with samples drawn from more traditional pools (Buhrmester, Kwang, & Gosling, 2011; Sprouse, 2011).  A correct CAPTCHA verification response at the end of the study was required for responses to be included in the data set, and I verified that no more than one response was associated with any given Mechanical Turk ID.  Moreover, following the recommendations of Huang, Curran, Keeney, Poposki, and DeShon (2012), I embedded a series of four questions within the questionnaire that were not related to the subject matter but to which participants responding attentively should provide the same answers. Data from five participants were later excluded on this basis.  The 28 males and 28 females (n = 56) remaining in the sample ranged from 19 to 60 years old (M = 32.46, SD = 9.55). Participants’ educational attainment was similar to that of the general population in that age range (see, U. S. Census Bureau, 2013). Of the sample, 29 percent (vs. 30 percent of the population) reported high school/equivalent as ultimate attainment, whereas 68 percent (vs. 60 percent of the population) reported attaining post-secondary learning. They spent an average of 32.5 minutes completing the questionnaire. 2.4.2 Study 1 Analyses and Results Following recommendations of Mackenzie, Podsakoff, and Podsakoff (2011), I first computed a two-way mixed model consistency interclass correlation on participant’s 64 responses to the four items repeated in the questionnaire matrix (Shrout & Fleiss, 1979). An interclass correlation coefficient (ICC) of 1.00 signifies perfect reliability; the ICC for this data, 0.915, indicated that participants were highly consistent in their responses. Then, as recommended by Hinkin and Tracey (1999), I restructured the data into long form (each item's N = 56 participants x 6 category ratings = 336, less any missing values) and performed a one-way repeated measures ANOVA for each item (see, Yao, Wu, & Yang, 2008). This approach is essential (Mackenzie et al., 2011) to account for error misspecification that might otherwise arise from each participant making multiple ratings (Winer, 1971). I set participant id as the panel variable. There was significant variation in fit ratings among categories for each item (p < .001). After each ANOVA, I engaged the Stata margins post-estimation command to generate 95 percent confidence intervals for the marginal mean for category fit, by category, for each item. Items for which the 95 percent confidence interval for the strongest fit category did not overlap that for the next strongest fit category were assigned to that strongest fit category and deemed eligible for further consideration. When the confidence intervals for an item's category fit overlapped between the strongest fit category and the next strongest fit category, I did not retain that item. All items presented to participants and the strongest fit categories to which they were assigned (if any) are shown in Table 2.   65 Table 2: Initial FSIx Item Pool and Strongest Fit Risk Categories  Item #  Item Strongest Fit Risk Category ST1 Decreased emotional support from close friends  Strong Ties ST2 Loss of respect and good will from family Strong Ties ST3 Having conflict with one’s relationship partner—spouse, partner, boyfriend, girlfriend, etc. Strong Ties ST4 Being made to feel unwelcome or unwanted by family members Strong Ties ST5 Being held in greater esteem by cherished relatives [Reverse coded] Strong Ties ST6 Gaining greater respect and good will from one’s close friends [Reverse coded] Strong Ties ST7 Loss of respect and good will from one’s mentors or childhood heroes Strong Ties ST8 Decreased ability to attract or keep an ideal mate Strong Ties CO1 Family, friends, or other valued social relations being physically or verbally assaulted because of their association with oneself Collateral CO2 Family, friends, or other valued social relations having decreased opportunities for success and prosperity because of their association with oneself Collateral CO3 Causing the authorities to scrutinize people known to associate with oneself Collateral CO4 Family, friends, or other valued social relations losing standing in their communities because of their association with oneself  Collateral CO5 People one cares about being the target of the harmful words or actions of others because of their association with oneself Collateral CO6 Family, friends, or other highly valued social relations being watched by the authorities because of their association with oneself  Collateral CO7 Shaming one’s family Collateral IN1 Being put on a government "watch list" by public authorities Institutional IN2 Being physically punished by public authorities Institutional IN3 Being put to death by public authorities Institutional IN4 Having to pay fines or financial penalties to public authorities Institutional IN5 Being condemned by public authorities Institutional IN6 Being jailed or imprisoned by public authorities  Institutional IN7 Being exiled or deported from one’s homeland Institutional     (table continues)  66  Item #  Item Strongest Fit Risk Category SA1 Being the target of hateful words / verbal assaults Societal, Active SA2 Receiving death threats  Societal, Active SA3 Being ridiculed and harassed by others within society Societal, Active SA4 Being physically assaulted by others within society Societal, Active SA5 Being threatened with physical harm by others within society Societal, Active SA6 Being stabbed or shot Societal, Active SA7 Being held against one’s will by those other than public authorities Societal, Active SA8 Angering the others present in the context Societal, Active SP1 Being treated as if one is immoral, tainted, or unwholesome Societal, Passive SP2 Achieving higher social standing in the community where one lives [Reverse coded] Societal, Passive SP3 Having less influence in society Societal, Passive SP4 Being seen as unfit for leadership positions within society Societal, Passive SP5 Loss of respect and good will from one’s neighbors Societal, Passive SP6 Having decreased access to sources of money / financial resources outside the workplace Societal, Passive SP7 Loss of respect and good will from other members of a valued religious or spiritual organization Societal, Passive PR1 Getting fired / losing employment Professional PR2 Getting in trouble at work Professional PR3 Being viewed less favorably by one’s employer Professional PR4 Losing the support of valued business relations Professional PR5 Having better future job prospects [Reverse coded] Professional PR6 Having fewer professional opportunities Professional PR7 Achieving a more favorable professional reputation [Reverse coded] Professional PR8 Harming the reputation of one’s employing organization Professional  Items not cleanly classified into a risk category XX1 Being publically shamed  XX2 Being kicked out of a valued club or social group  XX3 Having everyone present within the context convey agreement with or sympathy toward the views one has expressed [Reverse coded]      (table continues)  67  Item #  Item Strongest Fit Risk Category XX4 Receiving support from others present within the context, should one later express those same things elsewhere [Reverse coded] XX5 Being made to feel unwelcome within a valued religious or spiritual organization  XX6 Hearing disagreement from others present within the context  XX7 Being less likely to secure or retain an elected or appointed position within the government  XX8 Receiving support from others present within the context [Reverse coded]  XX9 Having less influence in government or politics  XX10 Finding allies among those present within the context [Reverse coded]  XX11 Having ‘common cause’ with the others present within the context [Reverse coded]  XX12 Damaging the reputation of one’s family   XX13 Losing access to valued non-monetary physical resources (e.g., lands, properties, natural resources)  XX14 Achieving greater wealth [Reverse coded]  XX15 Eliciting discord and disagreement from the others present in the context   Note. Italicized items comprise those subsequently selected (see Study 2) to form the FSIx.     68 2.5 Study 2: Selecting Items and Testing Measurement Model Study 1 allowed me to identify items that were representative of each of the six categories of risk for which free space needs provide protection. I undertook Study 2 to select best items from those pools, construct the FSIx measurement model, and tentatively test it. I also tested Hypotheses 1 and 2 in this study to provide initial information about this measure’s utility. 2.5.1 Study 2 Participants and Procedures  Study 2 participants were 43 females and 35 males (n=78) living in the United States who ranged from 19 to 67 years old (M=31.47, SD=9.34).  This study comprised a scenario-based experiment with participants recruited from Mechanical Turk, an acceptable data panel for such studies in organizational research (i.e., Skarlicki & Turner, 2014). Education attainment among participants was again similar to that among the US population (see, U. S. Census Bureau, 2013)—high school was the ultimate attainment for 25 percent of participants (vs. 29 percent of the population), whereas 67 percent had attained post-secondary education (vs. 60 percent of the population).  They spent an average of 19 minutes completing the study, and I ensured the veracity of their submissions in the same manner as in Study 1.  Conditions. Participants read one of five prompts, presented to them at random. Each prompt included two scenarios. In one, the person is described undertaking resistance in a general public setting; in the other, that same person is described undertaking the same or essentially the same action in a potential free space. The five conditions varied in terms of resistance issue and act (i.e., verbally expressing a desire for the totalitarian government in 69 his/her country to be forcefully overthrown; disclosing managers’ corrupt and abusive workplace behaviors; enacting an alternate gender identity; advocating affirmative action for the benefit of native peoples; disseminating a photo of youth setting himself on fire in protest of a military regime). Because the FSIx is intended for application in both offline and online spaces, some resistance contexts were Internet-mediated and others were not. Appendix A reports the texts of the five conditions. Measures. After reading their scenarios, participants indicated the likelihood (1 = very unlikely, 6 = very likely) of each potential harmful outcome that might arise from resistance action (i.e., item) from the actor’s perspective, (A) assuming that action took place in the unprotected/ public social space and (B) assuming that action took place instead in the potential free space. All items were presented in fully randomized order to each participant. Differences in each participant’s rating for each item [(A)-(B)] comprise that item’s safety score, a measure of the relative safety of the potential free space versus the unprotected space. Subsequently, participants responded to three additional items (1 = strongly disagree, 7 = strongly agree). I intended one as a global item, or an item that gets at the essence of the construct for which the measure is being developed, to facilitate item testing (Diamantopoulos & Winklhofer, 2001): “For people who wish to express the views held by the person in the scenario above or engage in those actions, the second context (as in the second part of the scenario) offers protection from bad things that might happen in the first context (as in the second part of the scenario)." The other two were measures of the free space outcomes predicted in Hypotheses 1 and 2: “If I were the person in the scenario 70 above, I would be more anxious and apprehensive about expressing those views or undertaking those actions within the first context (as in the first part of the scenario) than within the second context (as in the second part of the scenario)” and “If I were the person in the scenario above, I would definitely be more likely to express those views or undertake those actions within the second context (as in the second part of the scenario), than within the first context (as in the first part of the scenario).” Last, participants responded to demographics questions. 2.5.2 Study 2 Analyses and Results First I used the Study 2 data to select the best items from those retained in Study 1 to construct a valid index with a manageable number of items. In the initial step of this process, I verified that all items in each risk category’s pool were reflections of a common dimension as expected or, in other words, that the safety scores for each risk category’s items loaded to a single factor in principle components analyses (John & Benet-Martinez, 2000). For each risk category, with the exceptions of two, a single factor solution emerged. For the ‘strong ties’ and ‘professional’ risk item pools, however, two-factor solutions emerged. Further investigation revealed that the two items in each of those two pools (i.e., items ST5, ST6, PR5, and PR7 in Table 2) that loaded apart from the others were those that were positively worded and reverse coded, which suggested the two-factor solutions embedded methods artifacts. Internal consistency of the safety scores within each risk category (with those four items retained in their respective pools) was high, with values ranging from α=.80, for the ‘societal, passive’ pool, to α=.95, for the ‘institutional’ pool.  71 2.5.2.1 Item Selection  Together, the results reported immediately above provided reassurance that Study 1 item sorting was effective and that all items in each risk category pool were viable options. Of these options, however, a smaller subset was needed for a more manageable, practical measure. My goal in selecting a final set of items maximizing the extent to which those selected were characterized by both representativeness of their given risk category and distinctiveness, or lack of overlap with other risk categories. I deemed the first of these important because in my theorized measurement model, the six risk categories comprise reflective first-order constructs.  I deemed the latter important because in that model, these first-order constructs then are to serve as distinct, formative sub-dimensions of the higher-order free space construct, and high levels of multicollinearity in a formative measure can obscure the distinct influence of each indicator on the higher order (latent) construct (Diamantopoulos & Siguaw, 2006).  In order to facilitate the selection of 'best items' following these criteria, I calculated item-test correlations (i.e., correlations between the safety score value for each item and the aggregate of those for the other items comprising a category) to determine which items were strongly representative of the others in their respective categories. Items with an item-test correlation (i.e., representativeness) of at least .70 were retained as candidates. I then calculated the correlation of each item safety score with each of the other categories’ composite. Items with an average external-categories overlap (i.e., distinctiveness) of .70 or lower were retained as candidates. Finally, for the items remaining in the pool, I calculated representativeness:overlap ratios, and the items within each category with the two highest 72 ratios were considered best items. These analyses are reported by risk category (i.e., planned FSIx sub-dimension) in Appendix B. The two items emerging from these analyses to represent each of the six risk dimensions in the overall FSIx measure are shown in italics in Table 2, above. 2.5.2.2 Assessing Multicollinearity I then tested to ensure that my method for selecting the 12 best items had successfully yielded items that could be combined into a measure that would not be subject to problems arising from multicollinearity. A common rule of thumb for preventing such problems is that no item in a measure should have a variance inflation factor, or VIF, in excess of 10 (see, e.g., Hair, Anderson, Tatham, & Black, 1998; Kennedy, 2003), although a cutoff of 4 has also been advanced as a more conservative guide (Petter et al., 2007). I calculated the VIF for each of the 12 items and found the maximum among them (VIF = 3.31) was beneath even the more conservative threshold. Thus, multicollinearity was not a meaningful threat among the selected items.  2.5.2.3 Assessing Convergent Validity  Diamantopoulos and Winklhofer (2001) argued that each valid item in an index must be significantly correlated with an item that is not part of that index but summarizes the essence of the construct it purports to measure. Those authors described that criterion as an indication of external validity, but it might be more accurately described as a metric for convergent validity, or "the degree to which [an] operationalization is similar to (converges on) other operationalizations to which it theoretically should be similar” (Trochim & Donnelly, 2008, p. 61). Following Diamantopoulos and Winklhofer’s methodological 73 recommendation, I correlated each of the 12 items with the study’s global item, verifying that each was positively and significantly related to that item (at p < .05 or better).  2.5.2.4 Test of Model Specification As previously explained, I conceptualized the FSIx as a formative measure, an index, with six underlying dimensions to be captured reflectively through multiple items. To test this model specification (Figure 4), I utilized partial least squares (PLS) structural equation modeling (Hair, Hult, Ringle, & Sarstedt, 2014). I selected this approach because it accommodates both reflective and formative measures (Fornell & Bookstein, 1982), and both were present. I used SmartPLS for this analysis (Ringle, Wende, & Will, 2005) and followed the recommendations of Hair and colleagues (Hair, 2013; Hair et al., 2014) in doing so.  Figure 4 displays parameter estimates. As that figure shows, (1) each item safety score was predictive of its given first-order category construct (i.e., FSIx dimension), and (2) each of the dimensions was significantly related with the higher-order Free Space construct. I utilized SmartPLS's bootstrapping procedure with 5,000 resamples to facilitate significance testing. Each parameter estimate shown is significant at the p < .01 level.  I then undertook a number of additional steps to ensure the measurement model was appropriately specified and included no bad items. I began by examining the outer loadings for items in the model. Outer loading values ranged from .82 to .94, well above the .701 outer loading guideline (Hair et al., 2014). Next, I assessed composite reliabilities for each risk dimension. All were sufficiently high, ranging from .85 to .93. Third, I tested convergent validity. Convergent validity is demonstrated when the average variance extracted (AVE) for 74  Figure 4. PLS path coefficients for theorized FSIx measurement model (Study 2).  All path coefficients and composite reliability statistics (ρ) significant at the p < .01 level. See Appendix A for full condition details.75 Table 3: Discriminant Validity of FSIx Dimensions: Squared Interconstruct Correlations (Study 2)  Institutional Strong Ties Collateral Societal, Active Societal, Passive Professional Institutional (0.877)      Strong Ties 0.208 (0.816)     Collateral 0.487 0.455 (0.868)    Societal, Active 0.358 0.529 0.488 (0.837)   Societal, Passive 0.175 0.491 0.313 0.483 (0.736)  Professional 0.164 0.313 0.298 0.348 0.455 (0.871)  Note. AVE levels are presented (in parentheses) on the diagonal. The Fornell-Larcker criterion for discriminant validity is met when a construct (i.e., dimension in this measurement model specification) AVE level exceeds each of the squared interconstruct correlations of which that construct is a part. This criterion was met for all six dimensions.   a composite exceeds .50 (Fornell & Larcker, 1981; Hair et al., 2014), and as Table 3 shows, AVE levels ranged from a low of .74 (Social, passive) to .88 (Institutional).  Finally, I assessed discriminant validity in two ways. First, I ensured that the AVE for each FSIx dimension exceeded its squared interconstruct correlation with each of the other dimensions (the Fornell-Larcker criterion); as shown in Table 3, this guideline was not violated. Last, I ensured that each dimension’s items loaded more strongly to that dimension than to any other (Hair, 2013); all did.  2.5.2.5 FSIx Score Calculation After successfully establishing appropriate outer loadings of all items and composite reliabilities, convergent validity, and discriminant validity for all six dimensions, I calculated individual-level FSIx scores from the twelve item safety scores (SS) using Equation 1, below (cf. Ulaga & Eggert, 2006). 76 Equation 1: FSIx Score. ݔܫܵܨ = 	∑ ܵܵ݅ଵଶ௜ୀଵ6  I set the Equation 1 denominator to six to adjust the measure’s range to -10 to 10.8  In the current sample, the range was -3.50 to 9.83. Mean FSIx scores by condition, which I report in Figure 5 below, each exceeded zero, p < .05.  2.5.2.6 Hypothesis Testing I next tested Hypotheses 1 and 2 and, in doing so, also tested the measure’s preliminary utility. Hypothesis 1 predicted that the extent to which a potential resister would be apprehensive about acting in a given social space would be a negative function of that person’s perception of that space being free for resistance on the focal issue. This would be born out if FSIx score was found to mediate the relationship between context for action (i.e., Condition) and apprehensiveness about acting in that context versus in the broader societal context embedding it. It was.                                                    8 The highest value in the FSIx instrument’s scale is six (6 = Very likely), and the lowest is one (1 = Very unlikely), meaning that the range for an item’s safety score is from negative five (-5 = 1-6) to five (5 = 6-1). Correspondingly, the raw total for the 12 items safety score could range from  -60 to 60. 77  Figure 5. Mean FSIx score, by condition (Study 2). Condition texts are reported in Appendix A. Whiskers denote 95% confidence intervals. Mean for each condition is significantly greater than 0 (p < .05). Condition factor variables (see, e.g., StataCorp, 2013) (hereafter Conditions) predicted FSIx score, F(4, 73) = 15.44, p < .001, and the criterion, F(4, 73) = 2.55, p < .05. However, when I predicted the criterion from Conditions and FSIx score in tandem, F(5, 72) = 4.76, p < .001, the effect of FSIx score was significant (t = 3.47, p < .001), but the effect of no Condition was (t-vales ranged from .01 to 1.48, p > .10). These findings demonstrate mediation (Baron & Kenny, 1986) and are consistent with Hypothesis 1.  Hypothesis 2 predicted that extent to which a potential resister is likely to act in a given social setting is a positive function of her/his perception of that setting’s relative protection for resistance on the focal issue. This prediction would be supported if FSIx score mediated the relationship between Contexts and likelihood of engaging in resistance acts. That was my finding. 78 Conditions alone predicted the criterion, F(4, 73) = 5.56, p < .001, but when I predicted the criterion from Conditions and FSIx score in tandem, F(5, 72) = 7.32, p < .001, the effect of FSIx score was significant (t = 3.36, p < .001) but the effect of no Condition was (t-values ranged from .20 to 1.34, p > .10). These findings demonstrate mediation.  Study 2 yielded a strong potential universal refined measure of free space, the Free Space Index (FSIx). It also provided support for Hypotheses 1 and 2. These findings provided an initial assessment of the validity of the FSIx, as it functioned as expected. 2.6 Study 3: FSIx Cross-Cultural Assessment In Study 3, I gather data from a different sample pool. With this data, I repeat the measurement model testing described above, but I also extend those analyses. I test Hypotheses 3 and 4 and, in so doing, leverage a nomological network approach (Cronbach & Meehl, 1955) to further assess the validity of the FSIx (J. R. Edwards, 2001). Under that approach, when a measure relates to other constructs in a manner consistent with broader theory, one can be more confident in the validity of that measure (Mackenzie et al., 2011). Hypothesis 3 predicts how FSIx scores should vary in different cultural contexts. Hypothesis 4 predicts when and how FSIx scores should be affected by the rater’s self-perceived power.  Rao and Dutta (2012) showed that carnivals and festivals can comprise free space, with sedition and insubordination displayed publically there in ways that are masked or that can be claimed as perfectly innocent (see also, Scott, 1990). The festivals examined in their study were religious festivals that served as free spaces for constricted soldiers in advance of their rebellion in the 1857 Bengal Native Army. In the current study, I investigate another type of festival that I also expect to comprise free space—Gay Pride Festivals.  79 Because homosexuality is a subordinated identity and stigmatizing trait but carries no clear physical markers, gay and bisexual men and women often conceal their sexuality within generalized social contexts (Herek, 2000; Ragins, 2008). Gay Pride Festivals likely provide protection for the expression of these identities and the grievances of those who are not heterosexual, compared to society at large—i.e., they should comprise free space—and Study 4 participants considered the protection offered by Gay Pride Festivals for advocating hiring and university admissions preference for non-heterosexual individuals. I have argued that the extent to which a given interaction space offers protection from resistance risks for a given social issue is dependent in part on the receptiveness of the broader social context in which that space is embedded (Hypothesis 3), and this study tests that argument. Societies differ in overall negativity of attitude toward homosexuality, and, as such, FSIx scores should vary in accordance with positivity of societal attitudes toward non-heterosexual individuals in the city in which the festival is held. With that in mind, I recruited 315 participants at Gay Pride Festivals to take part in this study. Host cities were Copenhagen, Denmark; Vancouver, British Columbia, Canada; and Memphis, Tennessee, USA, and I drew participants in roughly equal numbers from each.  According to a 2008 multinational social attitudes survey conducted by the International Social Survey Program (www.issp.org), less than 25 percent of people living in Denmark held generally negative views toward homosexuality, whereas nearly 60 percent of people in the United States held such attitudes. In 1989, Denmark became the first country in the world to legally recognize same-sex unions. More recently, in 2005, Canada legalized same-sex marriage. However, as of 2013 when I conducted this study, nearly 75 percent of 80 US states (including Tennessee) expressly forbade these unions. Corrales (2010) ranked major cities throughout the world on LGBT friendliness, and Copenhagen ranked 10th whereas Vancouver ranked 30th. Although Memphis was not included in that list, the highest ranked US cities were Chicago (54th) and Los Angeles (55th). By way of comparison, both Chicago and Los Angeles were also among the top 15 of 137 major US cities ranked in the 2012 Municipality LGBT Equality Index report (Human Rights Campaign Foundation, 2012), while Memphis was among the bottom 20. Thus, in assessing the free space offered by Pride Festivals in their cities for advocating preference in hiring and university admissions decisions for non-heterosexual individuals, scores from participants in Vancouver should be higher than those from participants in Copenhagen but lower than those from participants in Memphis.  Finally, I have argued that free space offers protection for those low in power to express their subordinated views whereas they otherwise would be less likely to do so, but people with objectively similar levels of power do not always perceive their own power equally. Although gay and lesbian persons generally hold socially subordinated, stigmatizing identities and thus have impaired social power relative to their heterosexual counterparts (e.g., Herek, 2000; Ragins & Cornwell, 2001), their perceptions of their own power within society undoubtedly vary. Regardless of one’s objective power, perceived power influences the risk one perceives in potentially combative situations (Kipnis, 1976). Those high in self-perceived power should view a society that does not support homosexuality as relatively less risky than those low in self-perceived power. Conversely, those low in self-perceived power who live in areas in which homosexuality is not socially accepted should view Pride Festivals 81 as especially protective. This study tests that prediction (Hypothesis 4). If responses to the FSIx differ as expected, that also will provide additional evidence that my conceptualization of the construct and its measure is valid. 2.6.1 Study 3 Participants and Procedure I administered a questionnaire to a sample of attendees of Gay Pride 2013 Festival events in Vancouver (n = 123), Copenhagen (n = 100), and Memphis (n = 92) (total N = 315). Most participants (86 percent) were Caucasian, 3 percent were Asian, 4 percent were black, 4 percent were Hispanic, and the remainder were of other race or did not report racial identity. Regarding sexuality, 51 percent reported being attracted only to members of the same sex, 20 percent reported being attracted mostly to members of the same sex, 9 percent reported being attracted to both sexes equally, 5 percent reported being attracted to mostly members of the opposite sex, and 11 percent reported being attracted only to members of the opposite sex. There was no significant difference between cities in the proportion of sexualities reported χ2(25)=31.54, p = .17. Average participant age, M=31.5 years, SD=10.98, did not vary across cities, F(2, 312)=0.25, p = .79.  Participants had the option of completing the questionnaire on paper or online, with their personal electronic devices. In Vancouver, 79 percent of participants completed the paper form; in Memphis, 96 percent did so. Participants in Copenhagen also had the option 82 of completing the questionnaire in either medium in English or Danish.9 Of these participants, 26 percent completed the online form in English, 26 percent completed the online form in Danish, four percent completed the paper form in English, and 44 percent completed the paper form in Danish. The amount of time it took participants in all locations to complete the paper form was not measured, but participants who completed the (identical) questionnaire online did so in approximately 12 min, on average. There was no difference in the amount of time taken depending on language, F(1,73)=0.35, p = .56, or host city, F(2,72)=1.08, p = .35.  2.6.2 Study 3 Measures Free space. I provide a generic version of the FSIx instrument in Appendix C. In the current study, I customized it such that participants made two sets of ratings for the likelihood they would experience each of the 12 FSIx items if they were to openly advocate for LGBTQ persons to receive preference in hiring and university admissions decisions in order to make up for past and present discrimination.10  The first set of ratings, (A),                                                   9 The Danish version was translated from English by an organizations faculty living in Copenhagen and back-translated by another colleague there. I worked with a third colleague to resolve minor translation discrepancies. 10 I focused on a specific contentious issue that would allow me to begin assessing the dimensionality of the construct, as I discuss below. The instrument is customizable for use with any contentious or political issue, subordinated group or identity, or general social cause. It can be focused to a specific resistance act (as in this study) or more broadly (e.g., expressing dissatisfaction 83 corresponded to advocacy that would take place publically in a prototypic societal context in their geographic area (i.e., British Columbia, Denmark, Tennessee).  The second set of ratings, (B), corresponded to advocacy that would take place INSTEAD at a Gay Pride Festival event in the host city (1 = Very unlikely to 6 = Very likely). I later used these ratings to calculate the safety score for each item [(A) – (B)] and then calculated FSIx values using the Formula 1 (reported in Study 2).  Self-perceived power. I used six items from Rotter’s (1966) Locus of Control (LOC) scale to measure self-perceived power (Azzam, Beaulieu, & Bugental, 2007). There are two primary reasons for this. First, LOC refers to the extent to which individuals believe that they exert influence on their own outcomes, rather than those outcomes being determined by such things as powerful others, society or luck (P. B. Smith, Trompenaars, & Dugan, 1995) and, as such, is indicative of perceived power. Second, this measure has been shown to yield very similar results when administered in Denmark and the United States (P. B. Smith et al., 1995), as well as in the United States and Canada (Parsons & Schneider, 1974). Participants made forced-choice decisions between statements reflecting internal (1 corresponds with I control) or external (0 corresponds with Others control) perceptions. An example pair is: “What happens to me is my own doing” and “Sometimes I feel that I don’t have enough control over the direction of my life.” Values on this summated scale could                                                                                                                                                               with the government; discussing unethical things going on at work; etc.), depending on the research question being addressed. 84 range from 0 (high external LOC/ low internal LOC) to 6 (low external LOC/ high internal LOC). Higher scores indicate greater self-perceived power. Control variables. Age, Sexuality, Race, and Gender identity comprised individual-level control variables. Controls related to the data collection instrument were Language (0 = English and 1 = Danish) and Medium (0 = electronic and 1 = paper). 2.6.3 Study 3 Analyses and Results   As in Study 2, I utilized partial least squares (PLS) structural equation modeling to test the measurement model (Hair et al., 2014). Parameter estimates from each city’s data, along with parameter estimates for the combined data set are displayed in Figure 6. Echoing Study 2 results, both within each city’s data and in the combined data set, (1) each safety score item was predictive of its given FSIx dimension, and (2) each dimension was significantly related with the higher-order Free Space construct. I again utilized SmartPLS's bootstrapping procedure with 5,000 resamples for significance testing. All parameter estimates were significant at the p < .01 level.  I repeated the measurement model testing steps from study 2. Outer loading values for the full dataset ranged from .71 to .90, with each exceeding the .701 unconditional retention guideline (Hair, et al., 2014).11 In Figure 6, I report each dimension’s composite reliability (ρ). The items demonstrated convergent validity, with each dimension’s AVE value exceeding the .50 guideline (Fornell & Larcker, 1981; Hair et al., 2014). Each dimension’s AVE also exceed its squared interconstruct correlation with each of the other dimensions,                                                   11 Item SA5 had an outer loading less than .701 in the Memphis data subset. 85 Figure 6. FSIx PLS path coefficients for data collected in three international cities (Study 3).  All path coefficients and composite reliabilities statistics (ρ) significant at   p < .01 level. 86 meaning the Farnell-Larcker criterion was met, and each dimension’s items loaded more strongly to that dimension than to any other. This demonstrates discriminant validity (Hair, 2013).  2.6.4 Study 3 Hypothesis Testing and Results As a first step in hypothesis testing, I verified that selecting Copenhagen, Memphis, and Vancouver as data collection sites had achieved variance in prevailing attitudes toward homosexuality in the broader societal contexts in which each potential free space (i.e., Pride Festival) was embedded. As a proxy of local attitudes toward homosexuality, near the end of the questionnaire, participants had reported their level of agreement with the statement, “LGBTQ people in this city and in nearby areas experience quite a lot of discrimination” (1 = disagree strongly, 7 = agree strongly). There were significant differences between host cities in responses to this item, controlling for individual differences (i.e., Gender Identity, Sexuality, Age, Race, and Self-perceived power) and instrument differences (i.e., Language and Medium), F(19, 295) = 2.90, p < .001. The marginal mean for Memphis (4.67, SE = .20) exceeded that for both Vancouver (3.85, SE = .16; Mdiff = .82, p < .001) and Copenhagen (3.34, SE = .24; Mdiff = 1.33, p < .001); the mean for Copenhagen exceeded that for Vancouver mathematically (Mdiff = .51), but this difference was not statistically significant (p = .13). 2.6.4.1 Free Space, by Society  The relative protectiveness of a focal interaction space for resistance for a given social issue should depend in part on general attitudes toward the issue within the broader embedding context. The order of host cities’ general favorability toward homosexuality, as 87 reported above, was Copenhagen – Vancouver – Memphis. Given this, Hypothesis 3 would predict that mean FSIx score in our data is highest in Memphis and lowest in Copenhagen. I used OLS regression to test this, generating factor variables for all categorical variables (StataCorp, 2013).   Table 4. Summary Statistics and Correlations (Study 3) Variable  Mean    SD    1      2 1. Free space   1.62   1.67       2. Self-perceived power   2.95   1.29 -0.07  3. Age 31.48 10.98 -0.06 0.17**  Note. N = 315. ** p < .01. I report summary statistics in Table 4 and OLS regression results in Table 5. In Model 1, I regressed FSIx on host city. That model was significant, with each host city factor variable predicting perceived protectiveness of gay pride festivals for resistance on the LGBTQ-centric issue. To better characterize FSIx score differences between cites, I then ran post-estimation commands to estimate (margins) and compare (lincom) marginal means (UCLA Institute for Digital Research and Education, 2013b). The emerging pattern of results, depicted in Figure 7, below, was as predicted: Memphis Pride comprised greater free space than both Vancouver Pride and Copenhagen Pride, p-values < .01, whereas Vancouver Pride and Copenhagen Pride did not differ meaningfully from one another. Through Model 2, I verified that host city effects were robust to the influence of demographic- and data instrument- control variables. In Model 3, I added the remaining individual difference, self-perceived power; host city effects remained intact.   88 Table 5. OLS Regression Predicting Free Space Comprised by Gay Pride Festival Events in Three Societies (Study 3)  N = 315 Variable Model 1 Model 2 Model 3 Model 4 City   Memphis (reference)   Vancouver   -0.60** (.21)   -0.53* (.23)   -0.52* (.23)   -2.17*** (.60)   Copenhagen -1.56*** (.22) -1.08** (.37) -1.06** (.37) -2.17** (.65) Self-perceived power†   -0.11 (.07) -0.44** (.14)  Self-perceived power† x City:   Memphis (reference)   Vancouver       0.54** (.18)   Copenhagen    0.37* (.18) Gender   Female (reference)   Male    -0.14 (.20)   -0.14 (.20)   -0.20 (.19)   Transgender  0.10 (.57) 0.04 (.56) 0.07 (.33)   Gender queer  -0.41 (.75) -0.41 (.75) -0.45 (.75)   Other  -0.11 (1.63) -0.37 (1.64) 0.05 (1.63) Sexuality—attraction to:   Same sex only (reference)   Same sex mostly    0.20 (.24)   0.18 (.24)   0.23 (.24)   Both sexes equally  0.02 (.33) -0.02 (.33) 0.06 (.33)   Opposite sex mostly  -0.45 (.42) -0.46 (.42) -0.43 (.41)   Opposite sex only  0.16 (.31) 0.14 (.31) 0.07 (.31)   Other  0.38 (.49) 0.35 (.49) 0.61 (.49) Age†  -0.00 (.01) -0.00 (.01) -0.00 (.01)                        (table continues)   89 Variable Model 1 Model 2 Model 3 Model 4 Race   Caucasian (reference)             Asian  0.73 (.51) 0.75 (.51) 0.68 (.51)   Black  0.16 (.48) 0.18 (.48) 0.27 (.47)   Hispanic  -0.24 (.45) -0.23 (.45) -0.23 (.44)   Other  -0.57 (.52) -0.60 (.52) -0.48* (.51) Instrument medium   Online (reference)   Paper    0.52* (.24)   0.52* (.24)   0.52* (.24) Instrument language   English (reference)   Danish    -0.28 (.34)   -0.30 (.34)   -0.28 (.34) Intercept 2.36*** (.16) 1.99*** (.41) 2.24*** (.45) 3.16*** (.55) R2 .14*** .18*** .19*** .22***  Note. Unstandardized coefficients, with standard errors in parentheses. † Mean centered. * p < .05. ** p < .01. *** p < .001.    90   Figure 7. FSIx Score by Free Space Embedding Social Context (Study 3). Protectiveness of Gay Pride Festival events in Vancouver, BC, Canada, Copenhagen, Denmark, and Memphis, TN, USA, for advocating preference in hiring and university admissions decisions for LGBTQ persons, as measured by the FSIx. Predicted marginal means with continuous and factor variables related to individual differences and factor variables for the language and medium of the FSIx instrument each held constant at its mean. a Mdiff = 1.06, SE = .37. b Mdiff = 0.52, SE = .23. c Mdiff =.54, SE = .34.  * p<.05.  ** p<.01. 2.6.4.2 Moderating Effect of Self-perceived Power  In Model 4, I added the interaction between self-perceived power and city factor variables. Although self-perceived power did not directly effect FSIx score in Model 3, that does not rule it out as a meaningful moderator (Baron & Kenny, 1986). Rather, in fact, 91 Model 4 showed the City x Self-perceived power interaction significant. Figure 8 depicts the nature of the interaction. I estimated the simple slopes (b), difference in simple slopes (∆b), and 95 percent confidence intervals reported there with Stata’s margins command.   Figure 8. Conditional effect of Self-perceived Power (i.e., Internal Locus of Control) on free space perceptions for Gay Pride Festival events in Vancouver, BC, Canada, Copenhagen, Denmark, and Memphis, TN, USA, for advocating preference in hiring and university admissions decisions for LGBTQ persons, as measured by the FSIx (Study 3).  (A) Vancouver simple slope:      0.11, 95% CI [-0.12,  0.34].  (B) Copenhagen simple slope: -0.06, 95% CI [-0.29,  0.16].  (C) Memphis simple slope:      -0.44, 95% CI [-0.71, -0.17].  (A) – (B):  0.17, 95% CI [-0.15,  0.49].  (A) – (C):  0.54, 95% CI [ 0.19,  0.90].  (B) – (C): -0.37, 95% CI [-0.73, -0.02].   92 These results suggest that self-perceived power’s apparently positive interaction with Vancouver and Copenhagen in Model 4 is better characterized as lack of a negative interaction effect that was present for Memphis, the comparison city, but not for Vancouver or Copenhagen. In Memphis where attitudes toward homosexuality were generally deemed unfavorable, the greater were participants’ self-perceived power levels, the less they saw their pride festival as being an especially protective advocacy environment. This negative relationship was absent in Copenhagen and Vancouver where attitudes toward homosexuality were generally deemed favorable. This result pattern was as predicted in Hypothesis 4.  2.6.4.3 FSIx Dimensionality Finally, I have characterized free space as a multi-dimensional construct, and as a last step in my analysis, I began to test that. Specifically, neither being homosexual nor advocating for gay rights was illegal or punishable by law in any study 3 host city. Although people could conceivably be punished on trumped-up charges for other crimes because of their sexuality in these locations, it seems this risk should have been rather small. Conversely, given the negative workplace implications of non-heterosexuality in general (DeJordy, 2008; Ragins, Singh, & Cornwell, 2007) and especially given the nature of the resistance act participants considered in the study, protection from professional risk should have been comparatively higher. To determine if this was the case, I restructured the data into long form with dimension scores populating a single variable (Dimension Score). Following the method outlined by Rabe-Hesketh and Skrondal (2012), I determined that within-subject dependence in the restructured data could bias multiple regression results, χ2 = 696.56, p < 93 .001. Thus, in order to obtain unbiased estimates, I utilized Stata’s xtreg multilevel regression command with the maximum likelihood estimation (mle) option engaged and paneled the data by participant id.  The predictive model including risk dimension, city, and controls for each individual difference and questionnaire language and medium explained a significant amount of variance in Dimension Score, χ2 = 279.17, p < .001. I used the margins command subsequently to estimate the unique effect of each FSIx dimension, with all other predictors held constant at their means, and then linear combinations of estimators (lincom) analyses to test the significance of difference between each possible pair of FSIx dimensions (UCLA Statistical Consulting Group, 2013). I report these results in Table 6.   Table 6. Differences in Effect between FSIx Dimensions for Data from Gay Pride Festival Events in Three Societies for Advocating Preference in Hiring and University Admissions Decisions for LGBTQ Persons (Study 3) |Marginx - Marginy| Dimension   Margin z 1. 2. 3. 4. 5. 1. Institutional 0.773 6.90*** 2. Societal, Active 1.602 14.29*** 0.829*** 3. Societal, Passive 2.414 21.54*** 1.641*** 0.813*** 4. Strong Ties 1.443 12.87*** 0.670*** 0.159 0.971*** 5. Professional 1.967 17.54*** 1.194*** 0.365*** 0.448** 0.524*** 6. Collateral 1.484 13.24*** 0.711*** 0.117 0.930*** 0.041 0.483***  Note.  Estimations of effect for each FSIx dimension on safety score following multilevel regression in which dimension safety score was regressed on dimension, controlling for host city, self-perceived power, host city x self-perceived power, gender factor variables, sexuality factor variables, race factor variables, age, and version and language of the questionnaire. Data paneled on participant id. Standard error of margins = 0.112. *** p <.001, ** p<.01, * p<.05.     94 Findings were as expected. Protection provided from institutional risk (MarginI = 0.78) was less than that provided from professional risk (MarginP = 1.97; MarginI – MarginP = -1.19, p < .001).  In fact, relative protection from institutional risk was lowest of all in this dataset, whereas relative protection from professional risk outstripped all other dimensions except passive societal risk.   Thus, this study provides evidence that this construct is multidimensional in that there were significant differences in dimension scores, and the nature of those differences was as expected. Moreover, this study’s results also demonstrated that individuals’ perceptions of a social interaction space’s protectiveness for resistance on a given issue are determined in part by the larger societal context in which that space is embedded and by the extent to which they perceive of themselves as having power in their everyday lives. That FSIx scores varied in accordance with these two constructs, as predicted, provides additional evidence of the measure’s validity (Cronbach & Meehl, 1955; Mackenzie et al., 2011). It also suggests that what makes a social interaction space free inviting for resistance acts is not only its structural elements (in this study, e.g., high concentration of resisters in the space, ability to be anonymous through masks and costumes, resistance acts smuggled into public under a carnival guise, etc.) but also the mentalities of the actors within it and the contrast those actors feel between the space and the broader institutional context in which they live their ‘unprotected’ lives.    2.7 FSIx Usage Guidelines The FSIx is a measure of individual perceptions that can be combined to characterize an interaction space more broadly, and the treatment of these scores in future research 95 should be guided by the outcome(s) one wishes to predict. Individual perceptions—simple FSIx scores—are likely to guide individual expressions of subordinated identities, grievances, etc., and other micro-resistance acts against the established order and/or the norms that otherwise reinforce it. It is at the individual level of analysis that I operationalized FSIx scores in the studies reported above.  In order to characterize free space at a higher, collective level of analysis, individual perceptions (i.e., FSIx scores) must be aggregated. There are alternatives for doing this (Chan, 1998; Schneider, Salvaggio, & Subirats, 2002), and the choice of method should be concordant with the research question at hand. One option is the direct consensus approach in which individual scores would be averaged and within-group agreement would serve as a prerequisite for the group-level variable to be meaningful (Chan, 1998). Under this approach, the group-level value for free space would be the average, or most typical, way in which its members describe it (cf. Schneider, et al., 2002).  A truly free space is likely characterized not only by high perceived protectiveness for resistance related to a given issue within that space, on average, by those populating the space but also by high consensus among them, which is in line with the direct consensus approach. Indeed, to borrow from organizational culture researchers who have argued that an organization in which there is lack of consensus about culture has no culture at all (J. Martin, 1992), it is quite possible that a social interaction context in which there is lack of consensus about protectiveness is not free. Within a truly free space, individual expressions should elicit little harm, thus providing reassurance both to the actor and likeminded others in the space that it truly provides protections, and it is within such a space that the potential 96 synergism of free space in building momentum for collective change efforts will more likely be realized.   Another option for aggregating individual scores to the collective level of analysis is the dispersion approach (Chan, 1998). Under this approach, the degree of shared perception (i.e., within-group variability) would itself treated as the focal construct (cf. Schneider, et al., 2002; see also, Shanock, Baran, Gentry, Pattison, & Heggestad, 2010). Using such an approach, Chatman, Polzer, Barsade, and Neale (1998) showed, for example, that a person’s workplace attitudes and behaviors are affected by the extent to which that person differs from his or her workgroup, whereas Gibson, Cooper, and Conger (2009) showed that difference in ratings between team members and leaders affected constructive conflict and, ultimately, goal achievement (see also, e.g., Ostroff, Atwater, & Feinberg, 2004). Within the current context, individual undervaluation of a space’s protection is likely to disguise opportunities for self-expression, support-building, and promotion of collective action, whereas overvaluation should increase the likelihood of individual actions and expressions that yield negative outcomes or punishment. Within-group variability has also been shown to affect collective outcomes (e.g., team productivity: Guzzo & Shea, 1992, see Schneider et al., 2002), and it is likely that variability in free space perceptions will meaningfully influence the intermediary processes such as community-building (Evans and Boyte, 1986) and emotional contagion (Rao and Dutta, 2012) believed to link free space to successful collective change efforts.  Edwards (1995) discussed how to use multivariate procedures in analyses in which agreement is treated as the outcome variable. This would be useful in future research 97 predicting free space perceptions at a higher level of analysis. Factors to be considered in such research might include societal and psychological differences such as those in the current research, characteristics of space actors’ social networks, and characteristics of the space itself (cf. Rao and Dutta, 2012).   Shanoch, et al. (2010) provided direction and tools for conducting polynomial regression with response surface analysis. This approach provides a means for examining the relationship between (1) degree of agreement, (2) degree of discrepancy, and/ or (3) direction of discrepancy and a focal outcome. Through it, future research could examine, for example, the threshold for a space to serve collective purposes. If, for instance, 60 percent of the people with access to the space perceive of it as free, and the remaining 40 percent do not, will it elicit the disinhibition necessary to spur what would otherwise be risky action (cf. Hirsh et al., 2011)? Is that space conducive to open communication, community building (cf. Evans and Boyte, 1986), emotional contagion (cf. Rao and Dutta, 2012) and, ultimately, the birth of collective change efforts? Or need virtually all those with access to the space perceived it as free for these outcomes to occur? What is the critical mass? Similarly, if the distribution of free-space perceptions at the individual level is bimodal, with some perceiving the space as extremely free and others perceiving it as not at all free, is the space’s utility in change efforts the same as if the aggregate perception was the same but the distribution of individual perceptions normal? Leveraging FSIx data to answer such questions will expand our understanding of these spaces as change incubators. No matter how, or if, FSIx data is aggregated, the multi-dimensionality of the construct and the six sets of scores reflecting that multi-dimensionality should not be 98 forgotten. The risks from which a social space need provide protections to be free arise from different sources, and their effects may be direct or, in the case of collateral risk, indirect. The measurement model specified for the FSIx allows researchers not only to characterize the extent to which a social interaction context comprises free space but also to identify which risks might be in place to prevent a social interaction context from being free. Such understanding can usefully inform change efforts, alerting agents to perhaps minor structural alterations necessary to enable change.  2.8 Chapter 2 Discussion Access to spaces where non-elites can act outside the legitimized and socially-mediated control mechanisms that pervade socio-organizational hierarchies can play an important role in facilitating change. Although researchers have discussed these ‘free spaces’ as incubators for change, and recent research has demonstrated its promise as an important resource for those who seek to alter prevailing power structures (i.e., Rao & Dutta, 2012), no universal measure of free space has been advanced. Instead, in work to date, free space been assumed, described, or approximated, and, perhaps as a result, the construct itself has often been conflated with its potential outcomes and the processes through which it might be linked to those outcomes. My research refines the free space construct and advances a validated measure of it to promote and facilitate unified research on the construct going forward. In Chapter 1, I discussed in detail what free space is and, equally importantly, what it is not. I characterized free spaces as distinct social settings that provide protection from formal controls, cultural norms and practices, and other mechanisms of subordination 99 implicit to socio-organizational power hierarchies that otherwise dissuade non-elites from expressing their identities, ideologies, interests, or concerns. This characterization takes into account that resistance does not always take the form of coordinated efforts. The individual is the basic driving force of (re)organization (Barnard, 1968; Felin & Foss, 2005), and individual acts of dissent may serve as the micro-foundations for broader change. That said, I also emphasized that free spaces should be more conducive incubators for collective efforts to promote change than are more contested spaces, but such efforts need not arise, much less succeed, in order for the free spaces themselves to exist.  I derived from the literature six categories of risk from which free space need provide protection—legitimate risks, such as government-laden imprisonment or monetary fine; societal risks of an active nature, such as death threats or physical harm; societal risks of a passive nature, such as forfeiting social capital or bearing stigma; strong ties risks, such as disharmony and strife within familial relations; professional risks, such as job loss or impaired business relations; and collateral risks, or harms befalling those one cares about because of their relationship with oneself. I also highlighted that protection from these risks is likely to vary, within the same location, depending on the issue at hand. In other words, free space is likely topically specific, such that a free space for issues related to one group or social issue would not serve that same role for issues related to another.  Moreover, although free space has been discussed primarily in physical contexts in the past, free space is also likely embodied in virtualized (i.e., online) contexts (Gamson, 1996; Rao and Dutta, 2012). I argued that free space in online contexts may differ from free space in traditional, physical social interaction contexts. More specifically, by removing 100 geographic constraints on participation, the Internet allows for large numbers of geographically dispersed participants to join in combined efforts (Howe, 2008; Seidel & Stewart, 2011). As a result, relative to that emerging from physical free spaces, change efforts arising from online free spaces may be more likely to muster sufficient participants to succeed. In the case of failure, the cost and effort required to identify, locate, and punish large numbers of far-flung participants also should reduce the risk borne by any given individual. I suggested that because of the growing ubiquity of computer mediated communication and the tremendous volume and diversity of online social interaction contexts, being able to identify and characterize free spaces will become even more important in predicting the emergence of change efforts going forward, and concordantly, I designed the FSIx for application in both physical and virtualized social contexts. In this chapter, I reported three studies through which I specified and validated this measure. In the first, I generated a pool of items to assess the comparative safety of a given social space from the six categories of risk identified, test the content validity of each, and eliminate those invalid. In the second, I further refined the item pool, selecting the two best items for each dimension, as determined by their representativeness of their dimension overall and distinctiveness from the other dimensions. I also specified and tentatively tested the measurement model. In the last, I cross-culturally validated the measurement model with data collected in Canada, Denmark, and the USA.  Through the latter two of these studies, I also tested a number of hypotheses. Data from Study 2 demonstrated that perceptions of free space, as measured by the FSIx, predicted apprehensiveness about engaging in resistance in a given social space and 101 willingness to do so. Data from Study 3 showed that the perceived protectiveness of similar social contexts varies in accordance with receptivity of the broader societal context to the issue at hand and, moreover, that the influence of broader societal context differs, depending on actors’ self-perceptions of their own power in everyday life. Beyond their own merit, these findings were also valuable in that they provided additional evidence for the validity of my conceptualization of the construct as reflected in the FSIx measure; according to the nomological network perspective, when a measure covaries with related constructs in accord with theoretically-derived predictions, one can be more confident in the validity of that measure (Mackenzie et al., 2011).  Finally, I provided guidelines for other researchers to use this measure going forward. It is my goal to encourage systematic and interconnected research on free space and its role in organizational and social change going forward. By clarifying this construct, advancing theory around it, and creating a universal measure for it that can be customized for use in any setting, online or offline, I hope to promote and facilitate future work in the area will build from common ground and thus result in a more integrated and cumulative body of knowledge about free space and its function in individual and collective acts of resistance. Such knowledge has wide-ranging implications for understanding individuals’ behaviors such as stigmatized identity disclosure and whistleblowing, as well as for the collective resistance and organizational change literatures more broadly.  102 Chapter  3: Testing Impact on Hierarchy-Sustaining Action  3.1 Study 4 Hypotheses As I discussed in Chapter 1, it has been the typical nature of things that those with greater power and resources are able to act as they choose toward those with lower power and resources and do so with relative impunity (for a review, see Magee & Galinsky, 2008).  Recent changes in the way that people network and communicate with one another may be changing that to some extent, however.  Tens of thousands of people join Facebook groups formed with the sole purpose of responding to acts seen as power abuses (e.g., Facebook.com, 2010a) and people are "Tweeting" their dissonance with government corruption, even in relatively tightly controlled China (Kuhn, 2009).  It is not uncommon for a single online petition to garner hundreds of thousands (BBC, 2011) or even millions of individuals (Hawkins, 2011).  Social-media broadcasted grievances against power-holders perceived as having behaved badly toward the less powerful can generate negative publicity (e.g., Dunne, 2010), threats of boycott (Oliveira, 2013), and direct verbal harassment and even the potential for bodily harm (Cooper, 2010b).   The illustrative case of the school board vice president (Section 1.1) that opened this thesis serves as one example of this.  A case commonly referred to as “United Breaks Guitars” is another.  One of United’s customers, Dave Carroll, provides context for it (as recounted by Dunne, 2010) : “In the spring of 2008, Sons of Maxwell were traveling to Nebraska for a one-week tour and my Taylor guitar was witnessed being thrown by United 103 Airlines baggage handlers in Chicago. I discovered later that the $3500 guitar was severely damaged. They didn’t deny the experience occurred but for nine months the various people I communicated with put the responsibility for dealing with the damage on everyone other than themselves and finally said they would do nothing to compensate me for my loss.  So I promised the last person to finally say ‘no’ to compensation that I would write and produce three songs about my experience with United Airlines and make videos for each to be viewed online by anyone in the world.” (p. 1)12 As Dunne (2010) described in his case study of this incident and its aftermath, Carroll posted the first of these videos13 on July 6, 2009.  By day’s end, it had been viewed more than 150,000 times.  Within three days, that number reached 500,000.  By August 21, it had been viewed 5,000,000 times.14  A flood of negative comments about United Airlines were posted in the video’s comments section, and discussions of the video soon dominated the blogosphere.  Within two weeks of being posted, the video was the third result returned for the Google query, “United Airlines” (Hammond, 2009).   As a result of these events, United apologized to Carroll, offered him flight vouchers (which he refused), and made a $3,000 donation to charity as a “gesture of goodwill.”  A                                                   12 This case was the basis for the Internet risk manipulation in Study 4. 13 http://www.youtube.com/watch?feature=player_embedded&v=5YGc4zOqozo 14 As of April 6, 2014, that number had risen to 13,855,023. 104 Daily Mail article claimed that the incident cost the company 10 percent of its market value (Wrenn, 2009), and although that claim is rather dubious, what is not is that despite his tremendous power- and resource disadvantages, Carroll leveraged technology to challenge the power abuse he had suffered at the hands of the far more powerful corporation and, ultimately, to impose his will.  As this case suggests, the greater risk for illegitimate actions mediated by virtualization technologies should be meaningful to organizations and their agents (e.g., Ms. Irlweg, who Carroll called out by name in the video as she who informed him that United’s final word on his claim was “no”).  Although it seems clear that online responses to acts of power exploitation comprise a source of risk both to organizations and their agents, these likelihoods have received little research attention thus far.  Thus, I advance the following hypotheses.    Hypothesis 5: Organizational decision makers will perceive greater risk to the organization from exploiting power when there is potential for a negative Internet-mediated response to the exploitation than when there is not.  Hypothesis 6: Organizational decision makers will perceive greater risk to themselves from the organization exploiting power when there is potential for a negative Internet-mediated response to the exploitation than when there is not. Further, a rich body of literature demonstrates that risk of negative outcomes acts to deter behavior likely to trigger those outcomes (e.g., Taylor et al., 1973), both in organizations (e.g., Hollinger & Clark, 1982) and broader society (e.g., Shover, 1996).  An organization unjustly gaining at the expense of the less powerful, for example, is likely to elicit feelings of relative depravation—judgments that the current situation is not fair or as it 105 should be (Folger & Martin, 1986)—that can elicit collective protest (see, Van Stekelenburg & Klandermans, 2013, for a recent overview of the protest motive literature).  Organizations typically experience being targeted in social action negatively—it can disrupt operations, harm reputation, and ultimately elicit impaired financial performance, for example (e.g., King, 2008, 2011; Vasi & King, 2012). Thus, the risk of Internet-mediated social action should to reduce organizations’ propensities toward exploitive behaviors. Hypothesis 7: The potential for Internet-mediated action against organizations who exploit power decreases power exploitation. Although risk arising from online responses to negatively perceived power-laden acts should influence organizations’ propensities toward engaging in such acts in general, that influence will likely vary in accordance with the organization’s relative power advantage with its transaction partner (hereafter, target).  “Whether one is referring to individual relations or relations between commercial organizations, power is widely considered to be a function of dependence” wrote Handley and Benton (2012, p. 254), with the power of A over B equal to, and based upon, the dependence of B upon A (Emerson, 1962). This dependence is rarely unilateral, however, but can be counterbalanced to greater or lesser extent by the dependence of A upon B (Emerson, 1962; Wilkinson & Kipnis, 1978).  As such, power can exist at a dyadic level, with its balance determined by how much each entity in exchange or relationship depends on the other for valued tangible, informational, or psycho-sociological resources (French & Raven, 1959; Handley & Benton Jr, 2012) (or, perhaps more precisely, by each entity’s perception of this: see, R. A. Turner & Schabram, 2012). Because of their comparatively greater dependence on a given target, organizations with smaller power 106 advantages should be less inclined toward exercising their power to accrue gains at the target’s expense, ex ante.  In such case, existence of a risk outside the dependence relationship—e.g., Internet-mediated backlash—thus should have comparatively less capacity to influence the organization’s decision. Hypothesis 8: Organizations’ power relative to targets moderates the negative relationship between Internet-mediated response risk and pursuit of organizational goals to targets’ detriment.  The relationship will be less negative when the organization’s power advantage is smaller (vs. larger).  Finally, as Anderson, John, and Keltner (2012) emphasized, although in an objective sense power arises from resources and the capacity to make decisions regarding them, power is also a psychological state (see also, e.g., Bugental, Blue, & Cruzcosa, 1989; Galinsky et al., 2003).  Research indicates that in some cases, the power felt by an organization’s agent coincides well with external indicators of that power, but this is often not the case (Anderson & Spataro, 2005; Anderson, Srivastava, Beer, Spataro, & Chatman, 2006; Fast & Chen, 2009). This is important because feeling powerful tends to orient people toward pursuing gains, whereas feeling less powerful tends to orient people toward avoiding conflict (Anderson & Galinsky, 2006; Keltner et al., 2003; Kipnis, 1976). The implication of this in the present context is that the greater the extent to which organizational decision-makers see themselves as powerful entities, the greater should be their willingness to exploit power in pursuit of organizational goals, absent substantial risks to constrain them.  Risk of Internet-mediated response should have greater potency in an absolute sense among them than 107 among those who feel little power and, concordantly, should be less predisposed to exploit power.   Hypothesis 9: Personal sense of power moderates the influence of Internet-mediated response risk in agents’ decisions to exploit power in pursuit of organizational goals.  Among agents with higher (vs. lower) personal sense of power, its influence will be stronger. 3.2 Study 4 Overview To test these hypotheses, I conducted a scenario-based experiment based largely on a real-life case, United Breaks Guitars (Bernoff & Schadler, 2010; Dunne, 2010; Wrenn, 2009), which I describe in some detail below. There were four conditions in the experiment, comprising a 2 x 2 factorial design. Participants were placed in the role of organizational decision maker for a commercial airline that was either high in market power or lower in market power (power: high vs. low). They were then presented with a grievance in which a customer requested reimbursement for an expensive music instrument damaged by the airline.  Subsequently, each viewed the United Breaks Guitars Youtube video, which highlights the potential for negative Internet-mediated response by a customer who feels an airline has bullied them in handling a damage claim. Half of the participants were then told that after talking with the customer, they are certain that no Internet-mediated response will arise, no matter how they handle the claim (threat of Internet-mediated response: low), whereas the other half were told that after talking with the customer, they had no new information (threat: high). They then reported how much money they would offer the 108 customer to close the claim and also the risk to the organization and themselves they perceived if the customer should not be happy with their offer.  3.3 Study 4 Sample Study 4 participants were 171 North Americans (97 percent living in the USA; 3 percent in Canada) who were members of Amazon’s Mechanical Turk data panel and self-identified as either (A) having attended business school at a college or university (70 percent), (B) having at least one full year of full-time managerial experience in a corporate environment at a mid-level management position or above (85 percent), or (C) both (56 percent).  Following Huang et al. (2012), I embedded a series of items throughout the study questionnaire that were not theoretically relevant but to which generally attentive participants should provide the same answers.15  Data from 11 participants were excluded on this basis.  The 109 males and 51 females (n=160) remaining in the sample ranged from 19 to 70 years old (M=31.77, SD=9.78).  They devoted an average of 34 minutes completing the study, and each received a $3.01 honorarium as thanks.  No Mechanical Turk ID was associated with greater than one response.                                                     15 For example, one item, which was presented as the last in the generalized sense of power series but not included in its measure, asked level of (dis/)agreement with the statement, “I am favored in my dealings with others because I was born on the thirtieth day of the second month of the year.”  That, of course, is impossible, and low frequency responses were those indicating some degree of agreement with that statement. 109 3.4 Study 4 Procedure and Measures After indicating consent to participate in the study, participants were told that they would read a scenario and imagine themselves in the role of Operations Manager for Global Airlines, reporting directly to the Chief Operating Officer. They were instructed to take into account all of the information they would be provided in the scenario, assume it all to be true, and answer questions as though they were the Operations Manager.  Their goals and responsibilities as Operations Manager, they were told, included promoting high stock price (both because this is best for stockholders and because a large portion of their compensation depends on it) and minimizing the company’s costs and expenses wherever possible, so long as doing so did not reduce profitability, compromise safety, or harm the company overall.  They were also told that stock price is influenced by such things as company reputation, managerial effectiveness, and financial performance.  This information was constant across participants and interwoven with the initial Power manipulation. Power manipulation, part 1.  In this manipulation, participants were randomly assigned to read one of two descriptions of the airline industry and Global’s position in it.  In the low power condition (Power = 0), participants were told that (A) Global is one of the world’s 20 commercial airlines and that because (B) customers have a large number of choices of airline for flights to most destinations and (C) several individual airlines have the capacity to handle all of the customer demand, (D) competition for customers in the industry is generally fierce.  In the high power condition (Power = 1), participants were told that (A) Global is the world’s most predominant commercial airline and that because (B) customers have very few alternatives in selecting an airline for flights to most destinations 110 and (C) only Global has enough capacity to handle most of the customer demand, (D) competition for customers in the industry is extremely low overall.   Call for remuneration (stock information).  Next, participants were told that when they arrived at their office this morning, an email from a customer awaited them.  The text of that email (reproduced in Appendix D) incorporated information adapted from a real-life event (see Dunne, 2010, Appendix 1).  It conveyed that Pat, the customer, sought compensation for the destruction of an expensive, historic violin by Global baggage handlers. Pat, a professional musician, had intended to carry the violin on to the plane but an announcement had been made at the gate that larger carry-ons would not fit in the plane’s bins and would have to be checked for pickup at the final destination. Pat protested but ultimately capitulated and checked the violin.  The gate agent told Pat not to worry and that he would affix the violin case with bright “Fragile” and “High Priority” tags.  While waiting to deplane at the final destination, one of Pat’s bandmates and others on the plane observed Global baggage handlers throwing the violin and the band’s other instruments around recklessly.  Upon picking up the violin at the carrousel, Pat found that the promised tags were not affixed and that the instrument’s soundboard was cracked.  Pat brought the situation to the attention of a variety of Global employees but was dismissed by most. The agent at the global luggage desk told Pat that he could only authorize repairs up to a maximum of $200 and that any amounts over that had to be approved by the Global Luggage Line.   Upon calling the Luggage Line, Pat learned that repair receipts must be submitted before an appeal of the $200 limit would be considered.  Pat then submitted to the Luggage 111 Line representative a receipt for repairs totaling $3,000, along with a notarized statement from the World Violin Museum attesting that the violin had been made 300 years ago by a master craftsman and had been worth $18,000 prior to damages but, post-repair, was worth only $3,000. Only then had Luggage Line representatives notified Pat that the maximum amount that they could authorize was $250 and that only the Operations Manager or a member of executive management could approve a higher amount.  Pat found the Operations Manager’s email address on the Global website.  Power manipulation, part 2.  Subsequently, participants experience the second part of the Power manipulation.  Those in the low power condition were told that (E) Pat’s home airport was a hub for one of Global’s top competitors and because of the heavy flow of that airline’s flights in and out of that location, Global sometimes struggles to fill its flights for that airport.  Those in the high power condition were told that (E) Pat’s home airport was one of Global’s hubs and because of the heavy flow of Global flights in and out of that location, there were few flights offered by the minor regional airlines also serving it. Participants in both conditions were told that there was no high-speed railway servicing the city and that the nearest other airport was three hours from the city by car.   Power manipulation check (measure).  Power in business relations arises largely from the balance of dependency among the parties involved, and dependency arises from control of a resource that is important and scarce and for which there is no readily-available substitute (see, e.g., Hickson, Hinings, Lee, Schneck, & Pennings, 1971; Mechanic, 1962; Mintzberg, 1983). In both power conditions, the ability to travel across long distances expediently is important for Pat’s professional livelihood and the ability to do so via high 112 speed rail or by flying from another airport other than that nearest Pat’s home is constrained.  Also in both power conditions, income from customers is important for Global to sustain its operations.   In the high power condition, the dependency balance is more skewed toward Global because Global controls a greater portion of the supply for air travel, general availability of air travel from alternative suppliers is low, and the environment in which Global operates is not a highly competitive one where the income from any given customer is likely to be materially important.  In the low power condition, the dependency balance is less skewed toward Global because the supply of air travel is more dispersed across providers, several airlines have excess capacity, Global struggles especially to fill flights to/from Pat’s home airport (and most of the cost of operating a flight is fixed, meaning that every empty seat is important), and Global operates in a generally competitive environment where income from any given customer is likely of greater importance.      In order to allow for the effectiveness of these manipulations to be assessed, after reading the power manipulation, part 2, participants were asked to select a point on a line to reflect the balance of power between Pat and Global (1 = Pat has all the power, 10 = Global has all the power). Organizational performance (measure). To allow control for the possibility that the manipulation impacted perceptions of the company’s performance, an extraneous factor that could foreseeably influence subsequent responses, participants then responded to three items (1 = strongly disagree, 7 = strongly agree; α= .93) comprising an organizational performance 113 measure. (A) Global is a well-managed company. (B) Global operates efficiently. (C) Global is an effective company (cf. Katz & Kahn, 1978; Ostroff & Schmitt, 1993). Global baggage claims policy (stock information).  Next, participants were told that as they considered their response to Pat’s email, they consulted the Global Airlines Policies and Procedures Manual. They were presented with the relevant section (reproduced in Appendix D). In short, it confirmed that the airport luggage desk could authorize up to $200 for repairs and that the Luggage Line could authorize up to $250 for repairs. It also stated, “Repair or replacement costs in excess of $250 per customer, per flight may have a strong and negative impact on the profitability of that flight and must be approved by the Operations Manager or a member of executive management (i.e., COO, CTO, CFO, CEO).” Initial offer (measure).  Participants were then reminded of the goals and responsibilities of the Operations Manager and asked to spend a few moments thinking about themselves as Operations Manager in the situation, to consider their responsibilities and goals in that role, and to enter an amount in the box provided to offer Pat to close the claim, based on what they knew so far. Introduction of Internet-mediated response threat (stock information).  Next, participants were told that while further pondering their response to Pat, they recall having seen a YouTube video about a similar situation and that after a quick Internet search, they locate it and watch it again.  They are asked if they would like to watch the version of the 114 video with- or without- subtitles16 and subsequently viewed a music video for the song “United Breaks Guitars” written by Dave Carroll and performed by Sons of Maxwell (Appendix D includes song lyrics and links to the video).  The song recounts Carroll’s experience in which United Airlines baggage handlers damaged his Taylor guitar and subsequently refused to compensate him for it in a scenario much like that experienced by Pat in the study.  After watching the video, participants were also told, “You remember that the video was featured on national news programs (e.g., CBS, NBC, Fox) and cable news networks (e.g., CNN), as well as in Time Magazine and the international press.  You also recall that within 4 days of the video being posted online, United Airlines' stock price fell 10 percent (equating to about $180 million in value). Some coverage of the incident attributed the stock price decrease to this video. Others doubted whether this price drop could be directly linked to the video, however; they noted that some other airlines had drops in their stock price on that date, as well, and that United's stock price had been on an up and down roller coaster all that quarter.”  (These statements are true; see, e.g., Bernoff & Schadler, 2010; Dunne, 2010; Wrenn, 2009). Internet response threat manipulation.  In this manipulation, participants were given one of two additional bits of information, at random.  In the low risk condition                                                   16 Subtitled version was made available primarily to accommodate potential participants who were hearing impaired; both versions were otherwise identical.  Roughly half of the participants elected to view the subtitled version. Version selected was not related with dependent variable value (r = .006, p = .943). 115 (Internet risk = 0), they were told: “You decide to give Pat a quick call to discuss the situation. Afterward, you are fully and completely certain that no matter what you decide, Pat will not post a music video on YouTube or make any other sort of attempt to utilize the Internet or social media to spread the word, even if dissatisfied with the outcome.  You tell Pat that your written decision will be forthcoming.”  In the high risk condition (Internet risk = 1), they were told: “You decide to give Pat a quick call to discuss the situation. Afterward, you have really learned nothing beyond what you already knew. You tell Pat that your written decision will be forthcoming.” Internet response threat manipulation check (measure).  After the Internet risk manipulation, participants responded to a single item that comprised a manipulation check.  It asked them to mark a point on a line to indicate the likelihood that Pat will attempt to use the Internet (YouTube, social media, etc.) to spread the word if unhappy with their response to Pat’s email (1 = Pat definitely will not – 0% chance, 10 = Pat definitely will – 100% chance).  Risk perceptions (measures).  Participants next responded to a series of individual items adapted from Soprano, Crielaard, and Piacenza (2010) through which they characterized the likely implication (1= extremely positive, 9 = extremely negative) if Pat were to be dissatisfied with their response and responded to that dissatisfaction in exactly the way that they had predicted for: Risk to firm (α= .88) • Global’s overall financial performance over the coming months directly resulting from changes in the amount of business it receives from Pat • Global’s overall financial performance over the coming months resulting from changes in the amount of business it receives from customers other than Pat • Global’s reputation among investors and potential investors throughout the country 116 • Global employees’ confidence in the company’s strategy • Global employees’ confidence in the company’s trustworthiness • Government regulators’ trust of the company  Risk to self (α= .89) • Your career at Global • Your professional reputation as a manager • Your personal reputation in your community  Final offer and offer change (measures).  Participants were then again reminded of the goals and responsibilities of the Operations Manager and asked to spend a few moments thinking about themselves as Operations Manager in the situation, to consider their responsibilities and goals in that role, and to enter an amount in the box provided to offer Pat to close the claim. This comprises final offer.  Differences between initial offer and final offer comprise offer change. ‘Happy Pat effect’ (measure).  Penultimately, participants indicated (A) the likelihood that Pat would use the Internet to spread the word if happy with the response (0 = definitely would not, 7 = definitely would) and (B) the effect on Global if Pat did do so (-3 = very negative effect, 3 = very positive effect).  These values were subsequently multiplied to comprise this variable.     Individual differences (measures). Finally, participants responded to individual difference variables.  They first responded to 8 items from Anderson, John, and Keltner (2012) to assess their generalized sense of power (1= strongly disagree, 7 = strongly agree; α= .90). A sample item is: “In my dealings with others, I can get them to listen to what I say.”  Next, they replied to three items from Lennox and Wolfe (1984) (1= certainly always true, 6 = certainly always false; α= .68).  A sample item is: “I have the ability to control the way I come across to 117 people, depending on the impression I wish to give them.”  Last, they reported education and demographic data. 3.5 Study 4 Results I first tested to ensure the power manipulation and Internet risk manipulation functioned as intended. The mean response to the power manipulation item were lower among those in the low power condition, M = 5.60, SD = 2.77, than among those in the high power condition, M = 8.47, SD = 1.50, F(1, 158) = 65.9, p < .001, indicating the manipulation succeeded.  There was no significant difference in perception of organizational performance, which had been measured as a safeguard against introducing an unintended manipulation, between those in the two power conditions F(1, 158) = 2.75, p = .099.  The mean response to the Internet risk manipulation were lower among those in the low Internet risk condition, M = 3.60, SD = 3.18, than among those in the high Internet risk condition, M = 7.69, SD = 2.30, F(1, 1.58) = 86.67, p < .001, indicating the manipulation succeeded.  Descriptives and correlations for the focal variables in Study 4 are reported in Table 7.   118 Table 7: Summary Statistics and Correlations (Study 4) Variable        M    SD      1    2   3   4   5   6   7 1. Initial offer 5238.45 5977.67     ---       2. Power 0.49 0.50   -0.12     ---      3. Final offer 7900.00 7519.35    0.73*** -0.14    ---     4. Offer change 2585.51 5058.46   -0.09 -0.08   0.61***   ---    5. Internet threat 0.50 0.50    0.02  0.05   0.07  0.06   ---   6. Risk to firm 5.92 1.14    0.17* -0.07   0.20*  0.09  0.09  (.88)  7. Risk to self 6.10 1.31    0.19* -0.14   0.23**  0.12  0.14 -0.72***   (.89) 8. Happy Pat effect 0.94 1.72    0.17* -0.18*   0.18*  0.05  0.05  0.15  -0.31*** 9. GSP 5.17 0.90    0.05  0.00  -0.04 -0.11 -0.10 -0.09  -0.12 10. Self-monitoring 2.63 0.85   -0.05 -0.01  -0.07 -0.07 -0.06 -0.00  -0.00 11. Age 31.77 9.78    0.28**  0.07   0.20**  0.04 -0.03  0.03   0.05 12. Gender 1.32 0.47   -0.05 -0.08  -0.13 -0.12 -0.01  0.28   0.15            Variable        8       9     10     11      12     8. Happy Pat effect       ---         9. GSP    -0.07    (.90)        10. Self-monitoring    -0.03   -0.38***     (.68)       11. Age     0.03    0.05    -0.01      ---      12. Gender     0.02    0.07     0.02     0.07    ---      Note. n = 160. Scale reliabilities are reported in parentheses on the diagonal where applicable. Power is coded as low power (0) and high power (1). Internet threat is coded as low (0) and high (1). Gender is coded as male (1) and female (2).  GSP: Generalized sense of power. * p < 0.05. ** p < 0.01. *** p < 0.001. Table 8: Multiple Regressions Predicting Risk to Firm and Risk to Self (Study 4)   Risk to firm    Risk to self Predictor variable  β   t   β   t Internet risk Power  GSP Self-monitoring  .20 -.16 -.13 -.05 1.07 0.88 1.22 0.46   .33†  -.35  -.21†  -.08 1.66 1.73 1.67 0.60 Age Gender Constant .00 .32 6.20 0.39 1.63 7.43***   .01   .42† 6.57*** 0.77 1.90 7.06 Model R2 .04    .08  F 1.07  2.20   Note. n = 160.  Coefficients reported are unstandardized.  Internet risk is coded as low risk (0) and high risk (1). Power is coded as low power (0) and high power (1).  Gender is coded as male (1) and female (2).  GSP: Generalized sense of power.  † p<.10. *** p < 0.001.  119 I tested my hypotheses through multiple regression and report the main effects model results in Table 8.17,18  Support for Hypothesis 5 would be demonstrated if participants in the high Internet risk condition perceived greater risk to the firm for exploiting its power by  not compensating the customer fully for losses suffered than those in the low Internet risk condition.  This was not the case in this data, β = .20, p = .285.  Support for Hypothesis 6 would be demonstrated if those in the high Internet risk condition perceived greater risk to themselves as an agent of the firm for exploiting the firm’s power over the customer by not compensating the customer fully than those in the low Internet risk condition.  However, Internet risk and perceived risk to self were only marginally related, β = .33, p = .099. Finding that those in the high Internet risk condition offered Pat, the claimant customer, an amount greater than those in the low Internet risk condition would support Hypothesis 7.  Although offers from the former group, M = $8410.00, SD = $7536.50, were                                                   17 One dependent variable value corresponded with a Cook’s d value > 4/n and was thus treated as an overly influential case and replaced as missing (Bollen & Jackman, 1990). Results for analyses on the dataset including the initial value were materially equivalent to those reported.  18 The dependent variable data in this study were bimodal.  In post-hoc analyses, I engaged a variety of data transformation techniques and alternative regression specifications (e.g., probit, binomial, negative binomial), as well as PLS path modeling, which is not dependent on normality of distribution and other assumptions underlying OLS regression, and multilevel modeling with initial and final offer nested within participant and verified that this did not explain my findings. 120 over $1,000 higher than those in the latter group, M = $7383.55, SD = $7514.47, this difference was not statistically significant19—Internet risk was not a meaningful predictor of amount offered. Nor was it a meaningful predictor of change in amount offered before the Internet risk manipulation to after it.  Table 9 displays these results.  Table 9: Multiple Regressions Predicting Final Offer and Offer Change (Study 4)      Final offer ($)   Offer change ($) Predictor variable        β   t       β   t Internet risk Power  Risk to firm Risk to self Happy Pat effect Generalized sense of power Self-monitoring Age Gender Constant    536.38 -1939.50    646.11    675.53    430.50   -254.29   -649.54    169.27**       -2953.94*  1791.15 0.47 1.67 0.91 1.02 1.23 0.37 0.90 2.88 2.39 0.29   291.17  -845.68     68.14   393.68    -27.13  -700.11  -642.23     29.21 1435.41 6346.19 0.35 1.01 0.13 0.83 0.11 1.43 1.25 0.69 1.63 1.45 Model R2         .16**          .06  F       3.12         1.05   Note. n = 160.  Coefficients reported are unstandardized.  Internet risk is coded as low risk (0) and high risk (1). Power is coded as low power (0) and high power (1).  Gender is coded as male (1) and female (2).  ** p < 0.01. * S < 0.05.  Hypotheses 8 and 9 predicted moderation of an expected effect (Hypothesis 7) that I did not find support for.  I formally tested these hypotheses, nonetheless, using the approach outlined by Aiken and West (1991).  For Hypothesis 8, I added the Internet Risk x Power interaction term to each of the models shown in Table 9.  The interaction term was a significant predictor of neither final offer, β = 3238.95, t = 1.38, p = .171, nor offer change,                                                   19 This difference also was insignificant when I analyzed this data through ANOVA, F(1, 157) = .74, p = .391. 121 β = 777.69, t = 0.46, p = .647.  For Hypothesis 9, I added the Internet Risk x General Sense of Power (centered) interaction term to each model shown in Table 9. This interaction term also did not predict final offer, β = 704.60, t = 0.52, p = .601, and although it was a significant predictor of offer change, β = 2446.26, t = 2.61, p = .010, the overall model did not predict a significant portion of that criterion, F(10, 146) = 1.67, R2 = 0.10, p = .093.  Thus, I found support for neither Hypothesis 8 nor Hypothesis 9 in this data.  3.6 Chapter 3 Discussion The data from Study 4 provided support for no hypothesis advanced in Chapter 3. Although it is possible that each of those hypotheses was simply a prediction that did not align with reality, the direction of effects found were consistent with my predictions.  One explanation, then, for failure to detect these effects at a statistically significant level is that the study was underpowered.  For example, results from a post-hoc Monte-Carlo simulation power analysis (following UCLA Institute for Digital Research and Education, 2014) with 1000 repetitions and alpha = .05 suggested that the likelihood of falsely confirming the null hypothesis that the Internet Risk x Power interaction exerts no influence on final offer (RE: Hypothesis 8) with a sample of the current size is 88.7 percent (simulated power = .113).  Further simulations suggested that a sample of 500 participants per cell (n = 2000) would lessen the risk of such an error to a much more reasonable level (simulated power = .73).  Although “post-hoc power analyses are not universally admired” (UCLA Institute for Digital Research and Education, 2014 Example 2), these results likely are sufficiently stark to merit consideration. 122 Another potential explanation is that, despite self-identifying as having been trained in business schools and/or having significant professional experience in a strategic decision-making role, many participants were not well-equipped to understand the nuances of this study’s design.  For example, of those participants in the high power condition—who arguably had little (or at least comparatively less) to lose from citing ‘company policy’ and offering the claimant customer the standard maximum of $250 to close the claim—the average offer made before the Internet risk was introduced (i.e., Initial Offer) was $4469.74 (an amount that exceeded even the full cost to repair the damaged violin by nearly $1500).  This amount was statistically equivalent to that offered by those in the low power condition, F(1,156) = 244, p = .12.  Although this finding is perhaps encouraging from a business ethics perspective given that the repaired violin was said to be worth significantly less than it had been undamaged because of its historic significance, I am uncertain if it corresponds well with offers that would be made by business-minded decision makers.  In other words, given its highly-contextualized manipulations and measures, this particular study might not have been well aligned with the Mechanical Turk data panel.  This is an empirical question that I intend to answer by repeating the study post-dissertation, likely utilizing a sample of full-time managers in an executive MBA program. Such a sample should be well attuned to the study’s nuances. Further research to test Study 4 hypotheses is important because to the extent that Internet-mediated collective response raises perceived risk of engaging in blatant acts of discrimination and reduces propensities toward behavioral expressions of inequality, it can attenuate interaction patterns that perpetuate inequality within social structures. 123 Chapter  4: Testing Impact on Social Action In Section 1.6 of this thesis, I highlighted a catch-22: Successful social action relies both on action by those directly affected by a social issue but also on action by those who are not directly affected (e.g., Arrow, 1951; Downs, 1957; Mueller, 1979; Olson, 1965), but, as research has shown, those who are not directly affected are often unwilling to participate, even if they are sympathetic toward the cause (Ratner & Miller, 2001).  The practical significance of this can be extreme, as then-US Secretary of State Hillary Clinton (2011) described in an address to the United Nations—without advocacy efforts and supportive social action by allies within the majority, subordinated minorities can never achieve the critical mass necessary to assure that even basic human rights are upheld.   Thus, studies to examine factors predicting social action that include both online and offline contexts for action by vested parties and third parties are an essential component of this dissertation research.  I report two such studies below.  The first explores the capacity for virtualized context to impact social action decisions.  The second dives deeper, testing the capacity for virtualized context to promote social action by lowering direct costs for participation and by altering the influence of other inputs to social action decision making.  124 4.1 Study 520 Study 5 builds from Ratner and Miller’s (2001) research that tested, within offline contexts only, predictors of proclivity to undertake action to dissuade lawmakers from enacting legislation redirecting funding for promising research for a cure to a (fictitious) gastrointestinal disease to a largely ineffective billboard campaign encouraging seatbelt use. People have historically shied from social action participation in general, and Ratner and Miller’s work showed aversion to be especially strong among those individuals who were not directly affected by the issue at hand (or, in their language, which I adopted here, vested in that issue), even despite sympathy for the cause. This makes sense from a rational decision-making perspective, because those directly affected by a social issue are those with the most to gain (or lose); their perceived benefit from undertaking action should be greatest. That study also demonstrated, however, that fear of being socially punished for violating the social norm of self-interest (i.e., an expectation, especially among North Americans, that people should only act concordant to their own self-interests) played a pivotal role in determining action proclivity. That too is consistent with rational decision-making; forecasted social punishment comprises additional cost of action.                                                   20 This study was reported in a conference paper that received honorable mention for best paper in organizational theory at the 2012 ASAC annual conference.  That paper was a collaborative work with Marc-David L. Seidel.  A full manuscript is currently in preparation for the peer review process.  125 Yet those engaged in social action in online contexts seem often not to have a vested interest in the causes they promote. In reviewing successful change.org online petition campaigns, Reno (2013) highlights that the more than 171,000 people who signed a petition calling on the resulted in the South African government intervening to stop the “corrective rape” of its lesbian citizens lived in over 175 different countries. Non-South African signatories acted on the cause despite not having vested interest in it, as did any non-lesbian signatories living in South Africa. Although people lacking vested interest in a cause undoubtedly sometimes act for the benefit of that cause in offline contexts as well, the sheer volume and ubiquity of online social action participation suggests that important new insights might be gained by considering the influence of vestedness in an issue and online action context in tandem as determinants of social action participation. 4.1.1 Study 5 Hypotheses I treat variables from Ratner and Miller’s (2001) research as a baseline model in the current study. This baseline includes attitude-related predictors (i.e., valence and strength of attitude toward focal issue) as well as consequence-related predictors (i.e., vestedness; the anticipated response of others to one’s actions). While their study included only social action occurring in offline contexts, I also include action occurring online. An online social act likely requires the expenditure of less time, effort, and money than a comparable offline social act, which is to say that, all else equal, online action should be easier to engage in relative to offline action. Compare, for example, the time, energy, and expense involved in driving even across town to attend a community meeting to discuss a social issue to that involved in participating in an online community discussion of that same issue. I expect 126 people to be more willing to engage in social actions in online contexts than in offline contexts, all else equal.  Hypothesis 10: People’s willingness to engage in social advocacy is greater in an online action context than in an offline action context. I also incorporate additional factors that should be theoretically relevant to social action decisions in both online and traditional contexts. First, people’s utility assessments, or their perceptions of the likelihood that their action will make achieving the desired benefit from that action possible, should predict willingness to advocate (cf. Klandermans, 1984a; Vroom, 1964). Presuming that one central goal of social action participation will be, at least in most cases, to advance the focal issue or cause, believing that one’s action will make a meaningful contribution to that effort should encourage action (Klandermans, 1984b). Conversely, the prospect of engaging in social action that has little likelihood of advancing one’s cause should elicit comparatively less willingness to act (cf. Brunsting & Postmes, 2002; Svinicki & McKeachie, 2011). This prediction is consistent with the social identity perspective on social action, which contends that people will not act against a social structure unless they perceive it to be, at least to some degree, unstable (Tajfel & Turner, 1979), and with psychological research on self-efficacy (Bandura, 1977; Schunk, 1984; Stajkovic & Luthans, 1998). People who believe that the status quo will remain intact, even if they act, have little motive to do so (Martorana et al., 2005).  Hypothesis 11: People’s willingness to engage in a social advocacy is positively related to their perceptions of the utility of that action in accomplishing a desired social goal.  127 Next, another important factor that people consider when assessing actions they might undertake is how they will feel, or what affective experience they expect to have, as a result of acting or not acting. Would-be actors forecast the affective state that will follow a behavioral act, and this anticipated affect helps to determine their courses of action (Richard, van der Pligt, & de Vries, 1996; Rivis, Sheeran, & Armitage, 2009). In some cases anticipated affect corresponds to people’s favorable or unfavorable judgment of the behavior in general, but that is not always the case. For example, one might have a favorable attitude toward spending a fun and lively night drinking copious amounts of alcohol with friends but anticipate great regret when getting ready for work the next morning (Richard et al., 1996). To the extent that people anticipate engaging in a social action will make them feel good—positive anticipated affect—they should be more willing to advocate than when they anticipate experiencing negative feelings as a result of undertaking the action.  Hypothesis 12: People’s willingness to engage in a social advocacy is positively related to the positivity of their anticipated affect from doing so. Also likely to impact these decisions is the extent to which potential social actors expect that others who matter to them will be supportive of their action (Klandermans, 1984a). People pervasively desire harmony, approval, and support from others and especially from those who are psychologically close to them (Baumeister & Leary, 1995). To the extent that people expect that taking social action will elicit support from those who matter to them, the perceived benefit of taking that action should be augmented. The prospect of little or no social support, on the other hand, should make taking action comparatively less appealing. Moreover, social support has been conceptualized as a buffering resource from 128 which people can draw when facing stressful situations (S. Cohen & Wills, 1985), as when they encounter obstacles that might otherwise prevent them from achieving desired goals. As such, anticipated social support should also raise people’s belief that it is within their power to succeed in their action and thus achieve the positive outcomes they seek. When prospective social actors anticipate that they will have access to this resource, they should be more willing to act than when they anticipate this resource will be withheld. For both these reasons, I expect that willingness to engage in a social action will be buoyed by expectations of social support for that action.  Hypothesis 13: People’s willingness to engage in a social advocacy is positively related to their expectations of social support, should they do so.  4.1.2 Study 5 Procedures and Sample  I collected responses from 381 participants in the Mechanical Turk data panel. As in previous studies, I assured response validity and verified that no more than one response was associated with any given Mechanical Turk ID.  Participants were 35 years (SD = 11) old, on average, and 53 percent of them were female. Races represented included Caucasian (78 percent), Asian (9 percent), Hispanic (4 percent), Black (3 percent), and Other Race (6 percent). Participants had an average of 5 years of post-secondary education and 13 years of work experience.   I randomly assigned participants to one of four conditions that comprised a 2 (Vested vs. Not Vested in issue) x 2 (Online vs. Offline context) between-subjects design. In each condition, the participant read a vignette (adapted from Ratner & Miller, 2001, Study 4) describing a (fictitious) proposition under consideration by the US Congress that would 129 reduce funding from promising research on a cure for a gastrointestinal disease in order to fund a largely ineffective seatbelt usage campaign. The vignette explained that the disease affected only one sex, and I manipulated vestedness by varying the affected sex (vested = 1, not vested = 0).  Half of the participants of each sex were directed to a vignette in which males were affected by the gastronomical disease; the others were directed to a vignette in which females were affected. The vignettes, which appeared as news stories, did not otherwise vary. After reading a vignette, participants indicated their attitude toward, and perceived vestedness in (as a manipulation check), the issue and their willingness to advocate on the issue (Ratner & Miller, 2001) (A) to friends and (B) as part of a community discussion (cf. Brunsting and Postmes, 2002, in which advocacy was directed only toward socially distant entities via letter/email or petition/e-petition).  I varied the mechanisms through which advocacy would occur depending on online vs. offline action context condition. Those in the offline condition (online action context = 0) responded in regard to advocating to a group of friends in person and attending a community meeting about the issue and advocating there. Those in the online condition (online action context = 1) considered virtualized analogues of those actions: advocating to friends on an online social network such as Facebook and joining an online community group (e.g., a Facebook group) dedicated to the issue and advocating there. All participants indicated they understood what each of the actions they were presented entailed. 130 4.1.3 Study 5 Measures Dependent variable.  Participants responded to single item measures of willingness to undertake each advocacy act presented to them (1 = unwilling completely, 7 = willing completely) (Ratner & Miller, 2001). Baseline model variables. I used predictors from Ratner and Miller, 2001, as a baseline model. Participants used a sliding scale to indicate attitude toward the issue (-10 = oppose strongly, 10 = support strongly). Scores greater than zero were coded as agree (1) and scores less than zero were coded as disagree (-1). The absolute values of the raw score comprised strength of attitude. I measured others’ confusion about the participant engaging in the given social act as the average of the participant’s expected response from (A) members of the vested sex and (B) others participating in the movement (1= not at all confused, 6=completely confused; α=.87), should he or she advocate. A single item measured the extent to which participants expected others involved in the movement would value them undertaking the given social action (1 = not at all, 6= extremely).    Utility of action. I standardized and averaged responses from two items each to yield a measure of perceived utility of advocating one’s position [to a group of your friends; as part of a discussion of the issue at your local community center] [to your friends on Facebook; as part of a Facebook discussion group dedicated to this issue] (cf., Klandermans, 1984a). One of these measured perceived impact of action on the likelihood that one’s side would prevail (1 = reduced dramatically, 11 = improved dramatically). The other asked, “Will you and others who share your opinion be able to influence Congress to act in favor of your 131 view, if you were to personally do the following?” (1 = certainly no, 7 = certainly yes). This measure demonstrated sufficient reliability (α=.78).  Anticipated affect. Participants used sliding markers to indicate their anticipated affect (Richard et al., 1996) from each advocacy act they were presented. As participants slid the marker upward, the affect conveyed by a yellow “smiley face” graphic changed from a flat line (the center point, coded 3), to a weak smile (coded 4), to a strong smile (coded 5). As participants slid the line marker downward, the graphic’s conveyed affect changed from a flat line to a weak frown (coded 2), to a strong frown (coded 1).  Anticipated social support. Responses to the prompt: “How supportive would the people in your life whose opinions matter to you be of you advocating your position on this issue [to/as part of…]?” for each act considered (1 = very unsupportive, 6 = very supportive) measured anticipated social support (cf. Klandermans, 1984a; Klandermans, 1984b). 4.1.4 Study 5 Results As verified through the item, “To what extent does the proposed budget change affect you?,” the vestedness manipulation produced the intended effect, with females indicating higher vestedness in the female vested condition and males indicated higher vestedness in the male vested condition, F(1, 373) = 216.36, p < .001; the vestedness manipulation was equally effective among both sexes. Most participants (85.71 percent) were opposed to the proposition. Attitude strength did not differ by vestedness, F(1, 372) = .04, ns, nor did the opposition:support ratio, χ2 = 3.48, ns. Correlations are reported in Tables 10 and 11. Table 10 relates to advocating to friends, either online or in person, whereas Table 11 relates to advocating in the context of a community discussion, either online or in person. 132 Table 10: Summary Statistics and Correlations: Advocating One’s Position on a Social Issue to a Group of Friends (Study 5) Variable     M  SD   1   2   3   4   5   6   7 1. Vested in Issue  0.49 0.50   --- 2. Female  0.53 0.50 -0.03   --- 3. Attitude -0.71 0.70  0.10  0.02   --- 4. Attitude Strength  6.90 3.09 -0.01 -0.02 -0.46***   --- 5. Others' Confusion  1.94 1.25  0.05 -0.11*  0.31* -0.26***   --- 6. Value Input  3.93 1.18  0.11* -0.01  0.12*  0.11* -0.04   --- 7. Utility of Action  5.58 1.38  0.16*  0.03  0.20**  0.03  0.11*  0.54***   --- 8. Anticipated Social Support  4.72 1.03  0.10*  0.05 -0.11*  0.22* -0.20**  0.42*** 0.36*** 9. Anticipated Affect  3.96 0.84  0.12*  0.08 -0.01  0.16** -0.11*  0.47*** 0.40*** 10. Online Action Context  0.48 0.50 -0.01 -0.04  0.09 -0.01  0.12* -0.02 0.09 11. Willingness to Advocate  5.80 1.57  0.11*  0.05 -0.05  0.21*** -0.15**  0.37*** 0.31*** Variable 8 9 10         8. Anticipated Social Support   --- 9. Anticipated Affect  0.44***   --- 10. Online Action Context -0.12* -0.02   --- 11. Willingness to Advocate  0.58***  0.47*** -0.20**  Note. n = 381. Vested is coded as no vested self-interest (0) and vested self-interest (1). Female is coded as male (0) and female (1). Attitude is coded as disagree (-1) and agree (1). Online Action Context is coded as offline context (0) and online context (1).  * p < 0.05. ** p < 0.01. *** p < 0.001.   133 Table 11: Means, Standard Deviations, and Correlations: Advocating One’s Position on a Social Issue within the Context of a Community Discussion (Study 5) Variable M SD 1 2 3 4 5 6 7 1. Vested in Issue  0.49 0.50 2. Female  0.53 0.50 -0.03 3. Favorable Attitude toward Issue -0.71 0.70  0.10  0.02 4. Attitude Strength  6.90 3.09 -0.01 -0.02 -0.46*** 5. Others' Confusion  1.80 1.26  0.07 -0.13*  0.36*** -0.27*** 6. Value Input  4.31 1.17  0.04  0.09  0.04  0.08 -0.08 7. Utility of Action  5.79 1.42  0.12*  0.04  0.14**  0.00  0.13*  0.45*** 8. Anticipated Social Support  4.51 1.23  0.05  0.07 -0.02  0.19** -0.15**  0.45***  0.43*** 9. Affect  3.79 0.99  0.19**  0.04  0.03  0.12*  0.02  0.42***  0.46*** 10. Online Action Context  0.48 0.50 -0.01 -0.04  0.09 -0.01  0.07 -0.05 -0.06 11. Willingness to Advocate  4.87 1.85  0.17**  0.04  0.03  0.19** -0.01  0.37***  0.42*** Variable 8 9 10         8. Social Support 9. Affect 0.48*** 10. Online 0.02 0.00 11. Willingness to Advocate 0.52*** 0.58*** 0.24***  Note. n = 381. Standard errors are in parentheses. Vested is coded as no vested self-interest (0) and vested self-interest (1). Female is coded as male (0) and female (1). Favorable Attitude toward Issue is coded as disagree (-1) and agree (1). Online Action Context is coded as offline context (0) and online context (1).   * p < 0.05. ** p < 0.01. *** p < 0.001.   134 Willingness to advocate to friends.   Results from hierarchical OLS regressions predicting willingness to advocate to a group of friends, either in person or online, are shown in Table 12. The base model (Model 1) included predictors consistent with those included in Ratner and Miller’s (2001) analyses. I added utility to the baseline variables in Model 2, R2-change = .02, p < .01.  The remaining predictors were added in Model 3, R2-change = .22, p < .001.  Of the baseline predictors in Model 1, strength of attitude toward the issue and expectations that taking action would be valued by others were found to be significant predictors; being vested in the issue, gender, valence of attitude toward the proposition, and anticipations of others’ confusion were not.  In the full model (i.e., Model 3), no base model variable remained significant.  The data for advocating to friends did not support my expectation (Hypothesis 10) that people would be more willing to act in virtualized contexts. In fact, participants whose actions were to take place online were actually less, not more, willing to act than were participants whose actions were to take place offline, β = -.48, p < .001 (Model 3).  I found some support for my prediction that perceived utility of action would bolster willingness to advocate (Hypothesis 11) in this portion of the data. Utility expectation was a significant and positive predictor when added to the base model in Model 2, β = .30, p < .01, but was not a significant predictor in the full model (Model 3). Thus, although perceived utility of action was predictive of greater willingness to advocate to friends, its significance was attenuated when all other predictors were considered simultaneously. This data provided full support for 135 Hypotheses 12 and 13, however. Anticipated affect positively predicted willingness to act, β =.42, p < .001, as did anticipated social support, β =.60, p < .001.  Table 12: Ordinary Least Squares Regression Predicting Willingness to Advocate to a Group of Friends (Study 5)                        Model Variable   1a 2b   3c       Vested in Issue  0.25  0.21 0.12  (0.15) (0.15) (0.13)   Female  0.15  0.12  0.03   (0.15) (0.15) (0.13)   Favorable Attitude toward Issue -0.02 -0.07  0.10   (0.13) (0.13) (0.11)   Attitude Strength  0.07*  0.06*  0.05   (0.03) (0.03) (0.02)   Others' Confusion -0.12 -0.14* -0.02   (0.06) (0.06) (0.06)   Value Input  0.46***  0.35***  0.07   (0.07) (0.08) (0.07)   Utility of Action   0.30**  0.09     (0.10) (0.09)   Online Action Context   -0.48***      (0.13)   Anticipated Positive Affect   0.42***    (0.09)   Anticipated Social Support    0.60***      (0.07)   Constant  3.52***  4.05***  0.96   (0.37) (0.47) (0.48)   R2  0.18***  0.20***  0.42***    Note. n = 368. Standard errors are in parentheses.   a Model 1 includes, as a baseline, predictors consistent with Ratner and Miller’s (2001) study. b Model 2 tests the contribution of utility of action when added to the baseline model in predicting the dependent variable.  c Model 3 includes all theorized predictors.  * p < 0.05. ** p < 0.01. *** p < 0.001.    136 Willingness to advocate within a community discussion. Table 13 presents the results of hierarchical OLS regressions predicting participants’ willingness to advocate as part of either a local community discussion or an online community discussion. Model 1 comprises the baseline model. Model 2 includes the addition of expected utility of action as a predictor, R2-change = .07, p < .001, and Model 3 includes the addition of the remaining predictors, R2-change = .24, p < .001.  Of the baseline predictors in Model 1, vestedness, strength of attitude, and anticipation that others would value one’s action were significant predictors. Gender, attitude, and anticipation of others’ confusion were not. After the effects of all focal predictors had been accounted for (Model 3), the only base model predictors that remained significant were vestedness and strength of attitude toward the issue. Being vested played a relatively weaker role than did some other variables, however. In fact, when I standardized coefficients in a post-hoc analysis, the weight of the coefficient for vestedness was the lowest of all significant predictors, β = .08, t = 1.96, whereas the weight of the coefficient for online context was the greatest, β = .25, t = 6.44.  As Table 13 reflects, data for advocacy within community discussions were consistent with all Study 5 predictions.  First, as expected (and in stark contrast to the results reported above predicting willingness to advocate to friends), participants for whom community advocacy was to take place online were significantly more willing to act than were participants for whom these advocacy efforts were to take place offline; this provides support for Hypothesis 10.   137 Table 13: Ordinary Least Squares Regression Predicting Willingness to Advocate within the Context of a Community Discussion (Study 5) Model Variable   1a   2b   3c   Vested in Issue  0.54** 0.45**  0.29*   (0.18) (0.17) (0.14)   Female  0.07 0.02  0.01   (0.18) (0.17) (0.14)   Favorable Attitude toward Issue  0.25 0.16  0.05   (0.15) (0.14) (0.12)   Attitude Strength  0.12*** 0.11***  0.06*   (0.03) (0.03) (0.03)   Others' Confusion  0.03 -0.02  0.02   (0.08) (0.08) (0.06)   Value Input  0.55*** 0.32***  0.08   (0.08) (0.08) (0.07)   Utility of Action   0.63***  0.28**     (0.11) (0.10)   Online Action Context    0.91***      (0.14)   Anticipated Positive Affect    0.66***      (0.09)   Social Support   0.36***    (0.07)   Constant  1.55** 2.65*** -0.54***   (0.46) (0.48) (0.47)   R2  0.19*** 0.26***  0.50***    Note. n = 361. Standard errors are in parentheses.  a Model 1 includes, as a baseline, predictors consistent with Ratner and Miller’s (2001) study.  b Model 2 tests the contribution of utility of action when added to the baseline model in predicting the dependent variable. c Model 3 includes all theorized predictors.   * p < 0.05. ** p < 0.01. *** p < 0.001.  Next, the extent to which would-be actors anticipated that their efforts would have utility in advancing their cause and willingness to advocate were positively related, β = .28, p < .01, in support of Hypothesis 11.  Finally, positivity of anticipated affect, β = .66, p < .001, and anticipated social support, β = .36, p < .001, both 138 predicted greater willingness to advocate, providing evidence in support of Hypotheses 12 and 13, respectively. 4.1.5 Study 5 Discussion A focal conclusion from Ratner and Miller (2001) was that vestedness in a social issue is a key predictor of proclivity to engage in advocacy. Consistent with that conclusion, Study 5 showed being vested in the focal issue to be a significant predictor of willingness to advocate for one’s position on a social issue within the context of a community discussion. However, when additional predictors were added, vestedness played a much weaker role. In the full model predicting willingness to advocate in community discussions, being vested explained relatively little of the variance, compared to that explained by online context, for example. Moreover, being vested was not a significant predictor of willingness to advocate one’s position on a social issue to a group of friends, in either the baseline model or the model that included my additional predictors. Thus, although vestedness in a social issue can play a role in determining people’s willingness to engage in social action, its influence does not appear to be universal.  Study 5 also showed that perceived utility of action, anticipated social support, and anticipated affect associated with taking action are all meaningful predictors of willingness to advocate for a social issue. These factors, along with being vested in the issue at hand, appeared to combine differently in different situations, however. One reason for this may be that the strength of potential advocates’ social ties with intended advocacy audiences likely differ by situation and affect the perceived costs 139 and benefits of acting in that situation. People might behave differently in the presence of trusted friends or family members, for example, than when in the presence people whom they do not know well and whose reactions are thus harder to forecast. One might safely assume that social tie strength varied across the audiences for the advocacy acts considered in this study, but I did not test that assumption in this study, nor were there means for me to isolate its effect.  Results also suggested that online versus offline contexts for social action can exert significant direct influence on people’s willingness to undertake advocacy21, but my design did not allow that influence to be assessed within individuals. Moreover, this study did not investigate individual attributes beyond vestedness in the issue that are likely to influence advocacy decisions or how the influence of these attributes and/or other predictors of propensity to advocate might vary when an individual considers acting in an online versus offline context. With these limitations in mind, I undertook a second study.   4.2 Study 6 Study 6 expands and refines Study 5 in several ways. First of all, it broadens the potential downside of social action to include more anticipated risks. Risk refers to potential actors’ “subjective anticipation or expectation of a cost that they might incur                                                   21 Post-hoc investigation of the predictive efficacy of interactions between online/offline context and each of the other focal predictors showed none significant. This was true both for advocating to friends and in a community discussion. 140 as a result of their movement participation (e.g., being arrested, paying a fine, being beaten, tortured, or killed)” (Wiltfang & McAdam, 1991, p. 989). Risks, as potential future costs, depend not only on advocates’ actions but also on others’ responses to those actions (Wiltfang & McAdam, 1991). A general postulate underlying Study 6 is that social risks play an important role in determining people’s willingness to outwardly advocate their positions. I test the influence of social-risk related factors (i.e., strength of relationship with audience, favorability of society’s general attitude toward the issue, and stigma associated with the group vested in the issue) that might affect the favorability of social action participation outcomes.  4.2.1 Study 6 Hypotheses Based on my assessment of Study 5 results and social network theory, I expect that potential advocates’ perceived strength of social ties with intended audiences will influence willingness to advocate. Stronger social ties are characterized by greater openness and diversity of expression than are weaker social ties (Marsden & Campbell, 1984). People are also likely to accumulate idiosyncrasy credits in their relationships with stronger social ties, and idiosyncrasy credits can buffer any negative response that might arise if the audience should receive the advocacy act negatively (cf. Hollander, 1958). Strength of tie also predicts like-mindedness and potential acceptance (McPherson et al., 2001).  Hypothesis 14: People’s willingness to engage in a social advocacy is positively related to the strength of social ties perceived with the intended audience. 141 Moreover, the social risks associated with advocating an issue should vary in accordance with social attitudes toward the issue. Consider, for example, the response one might anticipate if lobbying for the same tax-funded travel abroad scholarship program for the benefit of (a) university honors students versus (b) convicted felons; or, the backlash that one might experience if advocating for those same university honors students to ‘benefit’ from a tax-funded illegal drug experimentation program, instead. I expect that, in general, people will be more willing to advocate for an issue when they perceive that social attitudes toward that issue are more favorable than when they perceive attitudes as less favorable (cf. Briscoe & Safford, 2008).  Hypothesis 15: Peoples’ willingness to engage in social advocacy is positively related to the extent to which they perceive that the general attitude toward the focal issue within their society is favorable. Somewhat relatedly, I also anticipate that the strength of stigma, if any, associated with vested groups in social issues will influence willingness to advocate. Social groups differ in the extent to which they are seen as tainted or otherwise undesirable within society (i.e., are stigmatized). Although most of the research on stigmas has focused on those who experience it directly, Goffman (1963) suggested that by associating oneself with individuals affected by a stigma, one can have that stigma “spread” to them like a communicable disease (p. 30)—stigma by association. Because of the norm of self-interest (Ratner & Miller, 2001), this may be especially likely for those who advocate on issues related to stigmatized groups. Those who are members of vested stigma groups would heighten society’s focus on that group 142 membership through their action (cf. Ferree, 2004) and those not members of vested stigma groups could ‘contract’ those groups’ stigmas by acting on issues related to them. Thus, people’s willingness to advocate for issues should be inversely related to the stigma associated with the group vested in those issues. Hypothesis 16: People’s willingness to engage in a social advocacy is negatively related to the societal social stigma of the group vested in the social issue. Next, some individuals are likely to calculate the costs and benefits of engaging in action for a given social issue differently than others (Feather, 1982). In Study 5 I tested a pivotal individual-level predictor, being vested or unvested in the issue. Here I delve deeper, testing the effects of potential advocates having family members among the vested group, being more or less proactive persons, and being more or less prone to experiencing guilt.  I expect that one’s family being a member of the group vested in a social issue will promote willingness to advocate on that issue. Aquino and colleagues (e.g., O'Reilly & Aquino, 2011; Reed & Aquino, 2003) argued that people consider some individuals as sufficiently close in psychological space to fall within their circle of moral regard and others as not. Those who fall within a person’s circle of moral regard are extended moral in-group status (Reed & Aquino, 2003), meaning that the person will be more likely to consider actions involving them through an other-benefitting, rather than self-centric, cognitive framework and may even be willing to bear socio-psychological discomfort for their benefit (O'Reilly & Aquino, 2011; Skarlicki & Turner, 2014). For even those with the most restrictive circles of moral regard, family 143 and close friends are expected to be included (Lewin, 1951; Reed & Aquino, 2003). As such, among non-vested potential advocates, the negative weights of socio-psychological risks should be greater when no family member is affected by the issue than when one or more family members is affected. Among vested potential advocates, the shared fate of vested family members should also heighten willingness to act because acting would promote not only self-interests but also the interests of others they have deemed worthy of concern.  Hypothesis 17: People’s willingness to engage in a social advocacy is greater when one has a family member(s) among the group vested in the social issue. Additionally, people differ in proactive personality, or personal disposition to create change in one’s environment (Bateman & Crant, 1993), and willingness to engage in a social action should be greater among more proactive individuals.  The prototypic proactive person shows initiative, identifies opportunities to change things for the better, and takes action; one who is not proactive adapts to the environment faced and is unlikely to identify, much less seize, opportunities to promote change in that environment (Crant, 2000). This relatively stable personal trait has been shown positively related to social entrepreneurship intentions (Priento, 2011), and I anticipate that it will function similarly within social advocacy decisions. In essence, given a group of similar individuals who hold similar views on a social issue, those who are the type who actively promote change in their environments should perceive a greater likelihood of achieving their desired favorable outcomes from social action and assign 144 positive weight to the act of advocating, itself, and thus be more positively disposed to do so (cf. Brunstein & Gollwitzer, 1996).  Hypothesis 18: People’s willingness to engage in a social advocacy is positively related to their proactivity. Next, I also expect guilt proneness, or propensity to experience guilt in everyday life, to affect willingness to advocate a given issue. Guilt is a member of the affective states collectively known as moral emotions. These emotions are linked to the interest or welfare of others and include gratitude, pride, moral anger, shame, and guilt, among others, and can provide motivational force to do good and avoid doing bad (Haidt, 2003; Kroll & Egan, 2004). Whereas shame is maladaptive—an often debilitating self-blaming of one’s person after a transgression—guilt is generally more functional (Tangney, Stuewig, & Mashek, 2007). Guilt is not related to a negative evaluation of one’s person but, rather, is a negative evaluation of one’s actions as violations of communal responsibility; it is experienced as self-generated pangs of conscience that can serve as calls to action (Tangney, Rowland, Flicker, & Barlow, 1996).  Those who are guilt prone are likely to anticipate guilt in response to a range of potential behaviors in social contexts (Tangney, 1990; Tangney et al., 2007), and guilt fosters other-oriented empathy and inhibits regrettable behavior affecting others (Haidt, 2003). When presented with a call to advocate for a social cause, individuals who are high in guilt proneness should forecast greater negative outcomes if they do not act. They should be more willing than otherwise similar individuals in the same 145 situation to act, and thus avoid a subsequent undesirable mental state they might otherwise experience.  Hypothesis 19: People’s willingness to engage in a social advocacy is positively related to their guilt-proneness. Finally, as I alluded to in Study 5, online context should impact willingness to undertake social action in part because of different direct costs of participation in online versus offline environments. However, context is also likely to have higher-order implications for social action decision-making. It is widely accepted that contexts affect decisions, behavior, and outcomes in social contexts (Johns, 2006; Payne, 1982), and calculations of the likely costs and benefits of advocacy are likely to differ between online and offline contexts. For example, people have been shown to evaluate risks less severely in online contexts (Debatin et al., 2009; Drennan & Previte, 2006). Congruent with this, in this study I explore the capacity for online versus offline context to alter the influence of the other willingness-to-advocate predictors (i.e., two-way interactions). I also anticipate that the moderating effects of online versus offline context will differ depending on whether potential advocates are vested in the issue and thus have the most to gain from advocacy (i.e., three-way interactions).  Hypothesis 20: People’s willingness to engage in a social advocacy is negatively related to direct costs of participation (i.e., money, time, and effort that must be expended in order to undertake that action). 146 Hypothesis 21: The influence of factors affecting people’s willingness to engage in social advocacy will differ depending on whether the act is to occur in an online or offline context Hypothesis 22: The moderating influences of online vs. offline context proposed in Hypothesis 21 will differ depending on whether potential actors are vested in the focal issue (i.e., three-way interactions among Online Context, Vested, and each of the other predictors).  4.2.2 Study 6 Overview To collect data for testing these hypotheses, I conducted an experiment with participants recruited both from the general Canadian population and from community organizations serving four socially disadvantaged and/or stigmatized groups. Participants learned about potential government actions that would negatively affect one of those four groups. They then reported their attitudes toward those actions and their willingness to engage in activism in a variety of forms, both online and offline, in response. Finally, participants responded to a number of socio-psychological and social network measures.   4.2.3 Study 6 Procedures and Sample Each participant in this study read and responded to two different scenarios.  These scenarios described a (fictitious) proposal before the Canadian Parliament to finance the expansion of a generally ineffective seatbelt awareness program by reallocating funds from existing community health programs (cf. Ratner & Miller, 2001) solely benefitting one of four groups: (1) First Nations/Aboriginal Canadians; 147 (2) Lesbian, Gay, Bisexual, and Transgendered—LGBT—Canadians; (3) Canadian cancer survivors; or (4) Canadians living with HIV/AIDS. These focal vested groups were chosen because I expected variance among them in social stigma associated with the group (expected to be lowest for Canadians affected by cancer and highest for Canadians affected by HIV/AIDS) (Kurzban & Leary, 2001) and societal support for a health program solely benefitting the group (expected to be lower regarding LGBT and Aboriginal Canadians and higher regarding Canadians affected by HIV/AIDS or cancer).  In total, 150 people participated in this study (n = 150). Because vestedness can play an important role in motivating social action participation, I partnered with community organizations to recruit 25 vested participants per focal group; I recruited an additional 50 participants from among the general population of Canadian Mechanical Turk data panelists.22  Participants were 35 years (SD = 12.5) old, on average, and 73 identified as male, 66 as female, nine as transgendered, and two did not indicate gender. Most (102) identified as white, 28 as Aboriginal/First Nations, 12 as Asian/ Pacific Islander, and eight did not identify or identified as Other race.  For                                                   22 In order to verify the veracity of Mechanical Turk as a sample pool, during the analysis stage of this study, I ran a post-hoc model predicting the criterion that included a dichotomous community partner sample (0) versus Mechanical Turk sample (1) indicator; its influence was not significant, p = .73. 148 highest level of education completed, 58 percent reported four year-degrees or beyond, 18 percent two-year degrees, and 24 percent high school/equivalent.   General population participants were serially presented two scenarios completely at random. Vested group participants were serially presented with two scenarios in random order: (A) a scenario in which funding for the ineffective billboard program was to come at the expense of a community health program exclusively serving their vested group and (B) an otherwise identical scenario in which the community health program exclusively served one of the other vested groups, at random. Appendix E provides a breakdown of sample source for each issue considered.  Following each scenario, participants indicated willingness to advocate on the proposal in five different forms (presented in random order by participant and scenario). These included: to friends in person, via online social network posting, in a local community meeting, in an online community discussion, and via Twitter.  Thus, this study had capacity to yield 1500 cases (150 participants x 2 scenarios x 5 advocacy acts).  However, for each advocacy act, participants had the option of selecting “I’m not sure what that means/ I don’t use that technology,” and I treated cases for which participants selected this option as missing. This reduced the case sample by 136, largely from Twitter (84 cases), leaving the sample to be analyzed at 1364 cases (case sample: N = 1364).   149 4.2.4 Study 6 Measures Independent variables. Vested in issue is a dichotomous indicator of whether the participant indicated membership in the group directly affected by the issue considered in the scenario (0 = non-vested, 1= vested). Online action context is a dichotomous indicator of whether an advocacy act considered within a given case occurs in an offline context (coded 0) or online context (coded 1).  After reading each scenario, participants completed the following measures.   Manipulation check.  A single item, “To what extent would the proposed budget reduction directly affect a group of which you strongly feel a part?” (0 = not at all, 10 = greatly), comprised a manipulation check for vestedness in the issue described in the scenario. I did not include a manipulation check for online action context. Attitude toward issue.  Two items indicated participant attitude on each of the two issues each considered: “To what extent do you agree with the proposed shift in funding…” (-3 = strongly disagree, 3 = strongly agree) and “To what extent do you support the proposed shift in funding…” (-3 = oppose strongly, 3 = support strongly) (cf. Ratner & Miller, 2001). I averaged these two items, r = 0.89, p < .001, and if the resulting value was negative, I set valence of attitude to -1; if 0, to 0; and if positive, to 1. The averaged value’s absolute value comprised strength of attitude.   Dependent variable.  To indicate willingness to act, participants responded to a panel of items headed with, “Please indicate how willing you would be to engage in each of the following activities in support of your position, either for or against the proposed shift in funding from the community health program focused on [focal 150 group] to the seatbelt billboard campaign.”  The items were presented in randomized order and included each of the five advocacy acts: “Bringing the proposal to the attention of your friends when you are together in person and advocating your position on it to them”, “Posting information about the proposal and advocating your position on it to your ‘friends’ or ‘followers’ on Facebook, Myspace, Google+, or a similar social connectivity website”, “Attending a meeting about the proposal at your nearest community center and advocating your position on the proposal there”, “Accessing a Facebook Group or similar online group that focuses on the proposal and advocating your position on the proposal there”, and “Tweeting about the proposal and advocating your position on it through Twitter” (1 = unwilling completely, 7 = willing completely).23                                                     23 I was mindful of the length of the study and, where possible, used single items for measures that related to specific advocacy acts (and were thus presented to participants 10 times each across two scenarios).  In advancing the C-OAR-SE procedure for scale development, Rossiter (2002) proposed that a single-item is appropriate when the object of a measure can be conceptualized as singular and concrete (e.g., a specific issue, action, group, etc.) or if the associated attribute (e.g., attitude toward, willingness to, feeling of, etc.) can be conceptualized as concrete.  Bergkvist and Rossiter  (2007) tested that assertion and found no difference in the predictive validity of single-item and multiple-item measures in such situations (see also, Boland, Brucks, & Nielsen, 2012; Nagy, 2002; Robins, Hendin, & Trzesniewski, 2001; Russell, Weiss, & Mendelsohn, 1989).   151 Direct cost of action. Participants indicated the cost of engaging in each of the five advocacy acts in terms of (A) money, (B) time, and (C) effort required (0 = none or a very insignificant amount, 9 = a very significant amount). Acts were presented in random order. I summed the three values for each advocacy act and then standardized that value.  Utility of action. “If you were to [each of the five advocacy acts serially, in random order], would this action meaningfully contribute to a larger effort to influence lawmakers to act in favor of your view?” (1 = no, definitely not, 5 = yes, definitely so), comprised the utility measure (cf. Klandermans, 1984a). Anticipated affect. Anticipated affect (Richard et al., 1996) was reported in response to this prompt: “If you were to [each of the five advocacy acts serially, in random order], how would that make you feel?” (1 = very unpleasant, 5 = very pleasant) (cf. Russell et al., 1989). Anticipated social support. “If you were to [each of the five advocacy acts serially, in random order], how supportive would the people in your life whose opinions matter most to you be?” (1 = not at all supportive, 5 = very supportive), comprised the anticipated social support measure (cf. Klandermans, 1984a). Social factors.  After responding to the measure above for both scenarios presented to them, participants answered questions regarding their perception of social attitude toward programs for the benefit of the vested groups in the scenarios that they had read.  The average of three items comprised this measure (average scale reliability=.82): “Of Canadians in general, how many would… |1| agree that 152 everything possible should be done to help [vested groups from scenarios serially, in random order]; |2| agree that special programs should be in place to advance the interests of…; |3| vote in favor of a new tax to fund health research that would exclusively benefit…” (0 = almost none, 6 = almost all).  All participants also indicated the social stigma associated with membership in each of the four vested groups.  They were instructed to make three quick, intuitive assessments of three items based on Goffman’s (1963) treatment of social stigma (cf. Peters, Burraston, & Mertz, 2004): “On average, Canadians believe that [membership in each of the four vested groups serially, in random order] is… |1| not at all disgraceful (1) to very disgraceful (7); |2| acceptable (1) to unacceptable (7); |3| not at all discrediting (1) to very discrediting (7).” Responses were averaged by vested group within subsample, average scale reliability = .88, and those averages were then population-weighted to yield a social stigma value for each scenario. The weighting factors, which corresponded to approximate percentages of the overall Canadian population comprised by each vested group, were as follows: Cancer-vested, .023 (Canadian Cancer Society, 2013); Aboriginal-vested, .037 (Statistics Canada, 2013); HIV-vested, .002 (Public Health Agency of Canada, 2011); LGBT-vested, .017 (Statistics Canada, 2004); General population participants, .921.   Social tie strength.  Participants then responded to series of four items with a shared trunk that characterized their relationships with those who would be their audiences in each of the five advocacy acts considered (presented in random order).  The trunk read, “Think about [advocacy audience, serially, in random order]. Answer 153 the following with them in mind.” Items were anchored |1| I hardly know them (1) to They are my close friends (5); |2| I interact with them rarely or never (1) to I interact with them more than once per week (5); |3| I would be completely uncomfortable confiding personal problems to them (1) to I would be completely comfortable confiding personal problems to them (5); and |4| We discuss very few topics (1) to We discuss a great many topics (5) (Marsden & Campbell, 1984).  I averaged these items; mean scale reliability was .88.   Individual differences.  Subsequently, participants completed individual difference measures. Ten items from Bateman and Crant’s (1993) proactive personality scale came first (averaged; scale reliability=.89).  A sample item was, “Wherever I have been, I have been a powerful force for constructive change” (1 = disagree strongly, 7 = agree strongly).  Next came nine items from Harder and Lewis’s (1987) PFQ2 (see also, Harder & Zalma, 1990) that were averaged to comprise the guilt proneness measure, scale reliability = .86. An example item asked: “How often do you experience each of the following feelings?: Regret” (1 = never, 4 = continuously or almost continuously).  Finally, age, gender identity, race, sexuality, and cancer- and HIV/AIDS status were reported. 4.2.5 Study 6 Results I first tested if the vestedness manipulation functioned as intended. Average vestedness level reported was 5.08, SE = 1.08. I performed a maximum likelihood multilevel regression predicting vestedness based on whether or not participants had reported being a member of the group affected in each respective scenario, paneling by participant. Being a member of the affected group predicted greater vestedness than not being a member of the affected group, β = 4.82, p < .001, as expected.  154 I next tested if social stigma and social attitude differed between conditions as expected and found that they had. Figure 9 displays population weighted means and 95 percent confidence intervals of social stigma perceptions for focal groups affected in the scenarios. The strongest stigma was associated with Canadians affected by HIV/AIDS (M = 3.81, SD = 1.52), followed by LGBT (M = 3.24, SD = 1.59) and Aboriginal (M = 2.81, SD = 1.72) Canadians, whose stigma levels were statistically equivalent to one another; the lowest stigma was associated with Canadians affected by cancer (M = 1.89, SD = 1.19).   Figure 9. Population weighted mean social stigma perceptions for focal groups affected (i.e., vested groups) in scenarios, with 95% confidence intervals (Study 6). Figure 10 displays mean perceived societal attitudes toward social programs solely benefitting these groups. These perceptions were most favorable for Canadians affected by cancer (M = 5.22, SD = 0.91), followed by those affected by HIV/AIDS (M = 3.81, SD = 1.08); the least favorable perceptions were for Aboriginal (M = 3.45, 155 SD = 1.10) and LGBT (M = 3.37, SD = 1.02) Canadians, which again did not differ significantly from one another.  Figure 10. Mean perceived societal attitude toward programs solely benefitting focal vested groups in scenarios, with 95% confidence intervals (Study 6).  Next, I next tested for variance in perceived strength of social ties with advocacy audience across the advocacy acts that participants considered. As Table 14 shows, the strongest social ties were, on average, with the friends to whom one might advocate in person, whereas the second strongest were with those to whom one might advocate via postings to one’s online social networking page. Among the other three advocacy acts, outbound ties were marginally stronger with those who would participate in an online community discussion about the issue than with those who would participate in a local community discussion; there was no significant difference in tie strength between Twitter followers and either of these advocacy audiences. 156 Means and standard deviations for outbound tie strength, by advocacy act, are given in that table, along with the results of paired t-tests comparing them.  Table 14: Outbound Social Tie Strength by Advocacy Act (Study 6) Advocacy act 1. 2. 3. 4. 5. 1. To friends, in person M= 3.96  SD= .79     2. Via online social network posting  Mdiff = 1.07 SE= .10 t= 11.01*** M= 2.90  SD= .92    3. In a local community meeting  Mdiff = 1.97 SE= .10 t= 19.18*** Mdiff = .97  SE= .10 t= 9.48*** M= 1.95  SD=1.01   4. In an online community discussion Mdiff = 1.85 SE= .10 t= 17.93*** Mdiff = .85  SE= .09 t= 9.40*** Mdiff =-.14 SE= .07 t= 1.94† M= 2.08  SD=1.03  5. Via Twitter Mdiff = 1.76 SE= .11 t= 11.33*** Mdiff = .84  SE= .14 t= 6.16*** Mdiff = -.12 SE= .12 t= 0.98 Mdiff = .07 SE= .12 t= 0.60  M= 2.05  SD=1.00 Note. Means and standard deviations for outbound tie strength by advocacy act are shown on the diagonal. Mean differences between matched pairs are presented below the diagonal, along with standard errors. t-values for mean differences between act pairs are presented below the diagonal. † p<.10. * p < 0.05. ** p < 0.01. *** p < 0.001.  Table 15 reports the descriptive statistics and correlations for Study 6 variables, along with scale reliability estimates where applicable. The data from this study comprised several cases per participant, creating a threat for within subject dependence to bias the results of my prediction testing. To determine if that threat was significant in my data, I followed the approach outlined by Rabe-Hesketh and Skrondal (2012), running a likelihood-ratio test comparing a null model predicting criterion variance with an otherwise null model that clustered cases by participant. .157 Table 15: Summary Statistics and Correlations among Study Variables: Willingness to Advocate One’s Position on a Social Issue (Study 6) Variable    M SD     1     2     3     4     5     6 1.   Willingness to Advocate   4.80 2.07   ---       2.   Online Action Context   0.60 0.49  0.004   ---      3.   Vested in Issue   0.37 0.48  0.278***  0.000   ---     4.   Family Vested in Issue   0.59 0.49  0.360***  0.000  0.620***   ---    5.   Favorable Societal Attitude: Issue   4.03 0.69  0.101***  0.000 -0.078**  0.103***   (.82)   6.   Social Stigma: Vested Group   2.94 0.70 -0.620***  0.000  0.043† -0.170*** -0.670***   (.88)  7.   Favorable Attitude toward Issue  -0.80 0.58 -0.081**  0.000 -0.088*** -0.117*** -0.033 -0.067**  8.   Strength of Attitude toward Issue   2.45 0.79  0.268***  0.000  0.221***  0.226***  0.111***  0.030  9.   Perceived Utility: Action   2.73 1.09  0.399*** -0.192***  0.159***  0.235***  0.099*** -0.016  10. Anticipated Social Support    3.62 0.92  0.470*** -0.155***  0.159***  0.234***  0.085***  0.003  11. Anticipated Positive Affect   3.85 0.97  0.435*** -0.010**  0.273***  0.324***  0.128***  0.005  12. Social Tie Strength   2.74 1.11  0.310*** -0.141***  0.088***  0.116***  0.055* -0.040  13. Proactivity   5.09 0.98  0.275***  0.000  0.161***  0.176***  0.005 -0.001  14. Guilt Proneness   2.07 0.51 -0.072**  0.000 -0.058* -0.039 -0.039  0.062*  15. Age 34.82 12.09  0.084**  0.000  0.087***  0.050†  0.005†  0.047†  16. Direct Action Costs   0.00 0.69 -0.092*** -0.858***  0.000  0.000  0.000  0.000                     (table continues)      158          Variable     7     8      9     10     11     12     13     14     15 7.  Favorable Attitude toward Issue   (.89)   8.  Strength of Attitude toward Issue -0.437***  (.89)        9.  Perceived Utility: Action  0.006 0.080**   --- 10. Anticipated Social Support  -0.169*** 0.187*** 0.537***   --- 11. Anticipated Positive Affect -0.175*** 0.285*** 0.422***  0.563***   ---     12. Social Tie Strength  0.095*** 0.009 0.138***  0.151***  0.165***  (.88)    13. Proactivity -0.188*** 0.143*** 0.142***  0.185***  0.167***  0.097***  (.89)   14. Guilt Proneness -0.009 0.002 0.012 -0.050† -0.059* -0.068*** -0.351***  (.86)  15. Age -0.060* 0.105*** 0.119***  0.033 -0.015  0.033 -0.038 -0.159***   --- 16. Direct Action Costs  0.000 0.000 0.221***  0.138***  0.090*** -0.150***  0.000  0.000 0.000  Note. Scale reliability estimates are presented on the diagonal in parentheses.  † p<.10. * p < 0.05. ** p < 0.01. *** p < 0.001.                 159 Table 16: Multilevel Maximum Likelihood Analysis, Main Effects Model Predicting Willingness to Advocate (Study 6)   Willingness to Advocate  Model 1  Model 2   βua SEb z 95% CI  βua SEb z 95% CI Fixed part                 Intercept, β0 -3.43*** 0.95 -3.62 -5.29 -1.57  -2.62** 0.96 -2.74 -4.49 -0.74 Variable       Online Action Context, β1  0.35*** 0.08 4.64 0.20 0.50  -0.52** 0.17 -3.14 -0.85 -0.20 Vested in Issue, β2  0.42*** 0.12 3.57 0.19 0.65   0.42*** 0.12 3.60 0.19 0.64 Family Vested in Issue, β3  0.28* 0.13 2.16 0.03 0.53   0.28* 0.13 2.17 0.03 0.53 Direct Action Costs, β4  -0.71*** 0.12 -5.83 -0.94 -0.47 Anticipated Positive Affect, β5  0.16* 0.07 2.48 0.03 0.29   0.17** 0.06 2.64 0.04 0.30 Societal Stigma, β6 -0.06 0.08 -0.75 -0.23 0.10  -0.06 0.08 -0.77 -0.23 0.10 Guilt Proneness, β7  0.18 0.18 0.97 -0.18 0.54   0.17 0.18 0.90 -0.20 0.53 Anticipated Social Support, β8  0.44*** 0.06 6.80 0.31 0.57   0.43*** 0.06 6.76 0.31 0.56 Social Tie Strength, β9  0.36*** 0.04 10.29 0.29 0.43   0.20*** 0.04 4.64 0.12 0.29 Favorable Societal Attitude: Issue, β10  0.15† 0.08 1.83 -0.01 0.32   0.15† 0.08 1.86 -0.01 0.31 Perceived Utility of Action, β11  0.19*** 0.05 3.67 0.09 0.29   0.23*** 0.05 4.47 0.13 0.33 Proactive Personality, β12  0.36*** 0.10 3.81 0.18 0.55   0.38*** 0.10 3.94 0.19 0.56 Favorable Attitude toward Issue, β13  0.46*** 0.12 3.94 0.23 0.69   0.48*** 0.12 4.16 0.25 0.71 Strength of Attitude, β14  0.62*** 0.08 7.87 0.47 0.78   0.63*** 0.08 8.00 0.47 0.78 Age, β15  0.00 0.01 0.62 -0.01 0.02   0.00 0.01 0.63 -0.01 0.02 Random part            √νc 0.94       0.95     √εd 1.30       1.28     Derived estimates           ρ 0.35       0.35     R2e 0.39***         0.40***                                                                                                                                                                                             (table continues)160 Note. N = 1364 cases nested within 150 participants. Online Action Context coded as Offline (0) and Online (1). Vested in Issue is coded as Non-Vested (0) and Vested (1).  a Unstandardized coefficient. b Average estimated standard error of the coefficient. c Estimated standard deviation for random intercept. d Estimated standard deviation for the residual. e Percentage of variance in propensity to advocate one’s position on a social issue explained by the model.  † p < .10. * p < .05. ** p < .01. *** p < .001.  The portion of variance in willingness to advocate attributable to clustering effects was significantly greater than zero, χ2 = 547.18, p < .001. Thus, I utilized a multilevel approach in testing my predictions in order to obtain unbiased results (Rabe-Hesketh & Skrondal, 2012; Raudenbush & Bryk, 2002).  Specifically, I utilized Stata’s xtreg multilevel regression command and paneled the data by participant number. Results from multilevel regressions predicting willingness to advocate on social issues from study variables are reported in Table 16. Model 1 excludes direct action costs, which had an exceptionally strong correlation with online/offline action context, r = -0.86, p < .001; Model 2 includes it.  I assessed the efficacy of these predictive models using an approach recommended by Snijders and Bosker (2012). Specifically, the coefficient of determination for a two-level model (i.e., its pseudo- or derived- R2) equals the proportional reduction of the estimated total unexplained (i.e., residual) variance between a theorized model and a null (i.e., intercept only) model (Rabe-Hesketh & Skrondal, 2012). Models 1 and 2 explained 39 percent, p < .001, and 40 percent, p < .001, of variance in willingness to act, respectively. Unless otherwise specified, results presented below refer to Model 2.  4.2.5.1 Main Effects Reiterating Study 5 findings, perceived utility of action, β = .23, p < .001 (Hypothesis 11), positivity of anticipated affect, β = .17, p < .01 (Hypothesis 12), and anticipated social 161 support, β = .43, p < .001 (Hypothesis 13), each exerted positive influence on willingness to advocate. Within Model 1, online action context exerted positive influence on willingness to advocate as well, β = .35, p < .001 (Hypothesis 10); when I partialed out the influence of direct action cost in Model 2, however, the influence of online action context became negative, β = -.52, p < .01. Consistent with Ratner and Miller’s (2001) findings, those who were members of the vested group in a social issue were more willing to advocate the issue than those who were not, β = .42, p < .001. Consistent with Hypothesis 14, participants indicated greater willingness to engage in an advocacy act when they had stronger (vs. weaker) social ties with the audience for that act, β=.20, p < .001.  As Hypothesis 15 predicted, perceived favorability of social attitudes toward focal issues was positively related with willingness to advocate, β = .15, p < .10, although its influence was significant only if the directional nature of my prediction was taken into account (i.e., one-tailed significance testing) (Fisher, 1971 [1935]). Societal stigma associated with vested groups did not affect willingness to act, β = -.06, ns.  Thus, this data did not provide support for Hypothesis 16.  As expected, having a family member directly affected by a social issue (Hypothesis 17) promoted advocacy, β = .28, p < .05, as did proactivity (Hypothesis 18), β = .38, p < .001.  Guilt-proneness (Hypothesis 19) did not, however, β = .17, ns. Finally, consistent with one of my central predictions in this research (and Hypothesis 20), direct cost of social action participation was negatively related with willingness to act, β = -.71, p < .001.  162 4.2.5.2 Contextualized Effects   Each result reported above was unconditional, which is to say it reflected the influence of the focal predictor averaged over all levels of the other predictors in the model. I expected conditional effects depending on (A) whether the advocacy act was to occur online (i.e., two-way interactions)—Hypothesis 21—and (B) whether the advocacy act was to occur online and whether the potential advocate was vested in the issue at hand (i.e., three-way interactions)—Hypothesis 22. To test for this, I performed an additional multilevel regression in which I regressed willingness to advocate on a model including the terms for the three-way interactions along with all lower order terms.  As I report in Table 17, this model accounted for 45 percent, p < .001, of variance in willingness to advocate (derived R2: Snijders & Bosker, 2012) . The additional criterion variance explained by the interaction terms was greater than zero, χ2 = 142.75, p < .001, which is to say that the six percent increase in derived R2 over the main effects model was statistically significant. Because the interpretation of lower order interaction effects is qualified by a significant higher-order interaction involving those same predictors, I first address the portion of the results involving three-way interactions. Potential advocates’ strength of attitude on the issue, β = -0.60, p < .05, and proactive personality score, β = -0.38, p < .05, as well as anticipated social support for, β = -0.45, p < .05, and anticipated affect from, β = 0.69, p < .01, engaging in the social action exerted differential influence in determining willingness to advocate depending on the combined influence of online context and vestedness in the issue.   163 Table 17: Multilevel Maximum Likelihood Analysis of Context Effects, Predicting Willingness to Advocate (Study 6)   Willingness to Advocate     βua   SEb    Z    95% Confidence       Interval Fixed part           Intercept -5.53*** 1.29 -4.28 -8.06 -3.00 Variable Online Action Context (OAC)  0.31 1.52 0.20 -2.66 3.28 Vested in Issue (Vested)  6.98** 2.13 3.27 2.80 11.16 Favorable Attitude toward Issue  0.52*** 0.15 3.59 0.24 0.81 Strength of Attitude  0.56*** 0.10 5.54 0.36 0.76 Utility of Action  0.12 0.08 1.45 -0.04 0.27 Anticipated Positive Affect  0.27** 0.10 2.69 0.07 0.46 Anticipated Social Support  0.24* 0.10 2.41 0.04 0.43 Family Vested in Issue  0.40* 0.17 2.36 0.07 0.73 Favorable Societal Attitude: Issue  0.26† 0.15 1.79 -0.02 0.54 Societal Stigma: Vested Group  0.06 0.14 0.45 -0.21 0.34 Social Tie Strength w/ Audience  0.36*** 0.08 4.78 0.21 0.51 Guilt Proneness  0.78*** 0.21 3.68 0.36 1.19 Proactivity  0.46*** 0.11 4.18 0.24 0.67 Age  0.02* 0.01 2.31 0.00 0.04 Direct Action Costs -0.41* 0.18 -2.30 -0.76 -0.06 Interaction: OAC x Variable Vested  2.60 2.54 1.02 -2.38 7.58 Favorable Attitude toward Issue  0.11 0.16 0.67 -0.20 0.41 Strength of Attitude  0.16 0.11 1.43 -0.06 0.39 Utility of Action  0.20† 0.10 1.93 0.00 0.39 Anticipated Positive Affect -0.30* 0.12 -2.48 -0.53 -0.06 Anticipated Social Support  0.35** 0.13 2.68 0.09 0.60 Family Vested in Issue -0.19 0.20 -0.94 -0.58 0.20 Societal Attitude: Issue  0.28 0.18 1.58 -0.07 0.63 Societal Stigma: Vested Group  0.15 0.17 0.88 -0.18 0.48 Social Tie Strength w/ Audience -0.24* 0.11 -2.27 -0.45 -0.03 Guilt Proneness -0.70*** 0.19 -3.75 -1.07 -0.33 Proactivity  0.04 0.09 0.42 -0.15 0.23 Age -0.02** 0.01 -2.81 -0.04 -0.01 Direct Action Costs  0.43 1.26 0.34 -2.03 2.89           (table continues)  164      βua   SEb    z    95% Confidence Interval Interaction: Vested x Variable Favorable Attitude toward Issue -0.48† 0.27 -1.78 -1.00 0.05 Strength of Attitude -0.10 0.22 -0.46 -0.52 0.32 Utility of Action  0.15 0.13 1.19 -0.10 0.41 Anticipated Positive Affect -0.11 0.17 -0.64 -0.43 0.22 Anticipated Social Support  0.26 0.16 1.64 -0.05 0.56 Family Vested in Issue -0.80 1.15 -0.70 -3.05 1.45 Favorable Societal Attitude: Issue -0.54* 0.26 -2.10 -1.04 -0.04 Societal Stigma: Vested Group -0.50† 0.27 -1.82 -1.03 0.04 Social Tie Strength w/ Audience -0.23† 0.13 -1.80 -0.48 0.02 Guilt Proneness -0.75** 0.26 -2.93 -1.25 -0.25 Proactivity -0.12 0.15 -0.78 -0.40 0.17 Age  0.00 0.01 -0.38 -0.03 0.02 Direct Action Costs -0.38 0.30 -1.26 -0.96 0.21 Interaction: OAC x Vested x Variable Favorable Attitude toward Issue -0.58† 0.33 -1.76 -1.22 0.06 Strength of Attitude -0.60* 0.26 -2.26 -1.11 -0.08 Utility of Action -0.14 0.17 -0.80 -0.48 0.20 Anticipated Positive Affect  0.69** 0.21 3.31 0.28 1.09 Anticipated Social Support -0.45* 0.22 -2.10 -0.87 -0.03 Favorable Societal Attitude: Issue -0.27 0.30 -0.88 -0.86 0.33 Societal Stigma: Vested Group -0.01 0.32 -0.05 -0.64 0.61 Social Tie Strength w/ Audience  0.21 0.18 1.21 -0.13 0.56 Guilt Proneness -0.03 0.32 -0.09 -0.67 0.61 Proactivity -0.38* 0.18 -2.12 -0.73 -0.03 Age  0.00 0.01 -0.05 -0.03 0.03 Direct Action Costs  0.49 2.15 0.23 -3.72 4.70 Random part √νc  0.91 √εd  1.22 Derived estimates ρ  0.36 R2e  0.45         Note. N = 1364 cases nested within 150 participants. Online Action Context coded as Offline (0) and Online (1). Vested in Issue is coded as Non-Vested (0) and Vested (1).  a Unstandardized coefficient. b Average estimated standard error of the coefficient. c Estimated standard deviation for random intercept. d Estimated standard deviation for the residual. e Percentage of variance in propensity to advocate one’s position on a social issue explained by the model.   † p < .10. * p < .05. ** p < .01. *** p < .001.  165 The nature of these interactions are presented in Figures 11 through 14, respectively. Simple slopes (b), difference in simple slopes (∆b), and 95 percent confidence intervals (CI95) were estimated using the Stata margins post-estimation command (UCLA Institute for Digital Research and Education, 2013b).  Delving deeper into these significant three-way interactions, I found that the influence of attitude strength was positive within offline contexts for both those vested, b = 0.46, CI95 = [0.06, 0.87], and those not vested, b = 0.56, CI95 = [0.36, 0.77], in focal issues. Its influence within online contexts was also positive for those not vested, b = 0.72, CI95 = [0.54, 0.91], but did not differ from zero for vested participants, b = 0.03, CI95 = [-0.34, 0.40]. Figure 11 depicts how the simple slopes for attitude strength’s influence differed depending on whether the potential advocate was vested within online contexts, ∆b = -0.69, CI95 = [-1.08, -0.31] but not within offline contexts, ∆b = 0.10, CI95 = [-0.31, 0.52].  As shown though Figure 12, the effect of proactive personality differed within online contexts, depending on whether the potential advocate was vested in a focal issue; it also differed among those who were vested, depending on whether advocacy was to take place online. Specifically, within online contexts proactive personality exerted positive influence for those not vested, b = 0.50, CI95 = [0.29, 0.71], but no meaningful influence among those vested in focal issues, b = 0.00, CI95 = [-0.22, 0.15], ∆b =-.50, CI95 = [-0.77, -0.22]. Within offline contexts, proactive personality was positively related to the criterion for both those vested, b = 0.34, CI95 = [0.05, 0.63], and those not vested, b = 0.46, CI95 = [0.24, 0.67]; the strength for those relationships was statistically equivalent, ∆b = 0.12, CI95 = [-0.17, 0.40], however. 166 Among those vested, the influence of proactive personality was more positive within offline contexts than within online contexts, ∆b = 0.34, CI95 = [0.04, 0.64]. Figure 13 displays the contextualized effects of anticipated social support for vested and unvested participants. Within offline contexts, anticipated social support exerted more positive influence on willingness to advocate among those vested, b = 0.49, CI95 = [0.24, 0.75], than among those not vested, b = 0.24, CI95 = [0.04, 0.43], ∆b = -0.26, CI95 = [-0.56, 0.05]. In online advocacy contexts, the influence of anticipated social support was again positive among both those vested, b = 0.39, CI95 = [0.12, 0.65], and not vested, b = 0.58, CI95 = [0.39, 0.77], but the difference in its influence did not differ significantly between them, ∆b = 0.19, CI95 = [-0.51, 0.13]. Among those not vested, anticipated social support exerted stronger positive influence in online contexts than in offline contexts, ∆b = -0.35, CI95 = [-0.59, -0.09].  Finally, Figure 14 shows the nature of the three-way interaction involving anticipated affect. Within online advocacy contexts, the influence of anticipated affect on willingness to advocate was positive among those vested, b = 0.55, CI95 = [0.30, 0.80], and insignificant among those not vested, b = -0.03, CI95 = [-0.21, 0.15], and the difference between those relationships was significant, ∆b = -0.58, CI95 = [-0.87, -0.29]. Within offline contexts, the influence of anticipated affect was positive among those not vested, b = 0.27, CI95 = [0.07, 0.46], and insignificant among those vested, b = 0.16, CI95 = [-0.12, 0.44], but the difference in its effects between the two groups was unsubstantial, ∆b = 0.11, CI95 = [-0.22, 0.43]. For those vested in the focal issue, the influence of anticipated affect was more positive in online 167 contexts than in offline contexts, ∆b = 0.39, CI95 = [0.05, 0.72]. For those not vested, the opposite was true, ∆b = -0.30, CI95 = [-0.53, -0.06].  Figure 11. Nature of Online x Vested x Attitude Strength interaction predicting willingness to advocate (Study 6). Left: Offline. (A) Non-vested simple slope: 0.56, 95% CI [0.36, 0.77]; (B) Vested simple slope: 0.46, 95% CI [0.06, 0.87]. (A) – (B) = 0.10, 95% CI [-0.32, 0.52]. Right: Online. (C) Non-vested simple slope: 0.72, 95% CI [0.54, 0.91]; (D) Vested simple slope: 0.03, 95% CI [-0.34, 0.40]. (C) – (D) = -0.69, 95% CI [-1.08, -0.31].  Between: (A) – (C) = -0.16, 95% CI [-0.38, 0.06]. (B) – (D) = 0.43, 95% CI [-0.04, 0.90]. 1234567Willingness to Advocate0 1 2 3Strength of Attitude toward ProposalNon-Vested VestedOffline1234567Willingness to Advocate0 1 2 3Strength of Attitude toward ProposalNon-Vested VestedOnline168  Figure 12. Nature of Online x Vested x Proactivity interaction predicting willingness to advocate (Study 6). Left: Offline. (A) Non-vested simple slope: 0.46, 95% CI [0.24, 0.67]; (B) Vested simple slope: 0.34, 95% CI [0.05, 0.63]. (A) – (B) = 0.12, 95% CI [-0.17, 0.40]. Right: Online. (C) Non-vested simple slope: 0.50, 95% CI [0.29, 0.71]; (D) Vested simple slope: 0.00, 95% CI [-0.28, 0.28]. (C) – (D) = -0.50, 95% CI [-0.77, -0.22].  Between: (A) – (C) = -0.04, 95% CI [-0.22, 0.15]. (B) – (D) = 0.34, 95% CI [0.04, 0.64].  1234567Willingness to Advocate1 2 3 4 5 6 7Proactive PersonalityNon-Vested VestedOffline1234567Willingness to Advocate1 2 3 4 5 6 7Proactive PersonalityNon-Vested VestedOnline169  Figure 13. Nature of Online x Vested x Anticipated Social Support interaction predicting willingness to advocate (Study 6). Left: Offline. (A) Non-vested simple slope: 0.24, 95% CI [0.04, 0.43]; (B) Vested simple slope: 0.49, 95% CI [0.24, 0.75]. (A) – (B) = -0.26, 95% CI [-0.56, 0.05]. Right: Online. (C) Non-vested simple slope: 0.58, 95% CI [0.39, 0.77]; (D) Vested simple slope: 0.39, 95% CI [0.12, 0.65]. (C) – (D) = 0.19, 95% CI [-0.51, 0.13].  Between: (A) – (C) = -0.35, 95% CI [-0.59, -0.09]. (B) – (D) = 0.11, 95% CI [-0.23, 0.45].   1234567Willingness to Advocate1 2 3 4 5 6 7Anticipated Social SupportNon-Vested VestedOffline1234567Willingness to Advocate1 2 3 4 5 6 7Anticipated Social SupportNon-Vested VestedOnline170  Figure 14. Nature of Online x Vested x Anticipated Positive Affect interaction predicting willingness to advocate (Study 6). Left: Offline. (A) Non-vested simple slope: 0.27, 95% CI [0.07, 0.46]; (B) Vested simple slope: 0.16, 95% CI [-0.12, 0.44]. (A) – (B) = 0.11, 95% CI [-0.22, 0.43]. Right: Online. (C) Non-vested simple slope: -0.03, 95% CI [-0.21, 0.15]; (D) Vested simple slope: 0.55, 95% CI [0.30, 0.80]. (C) – (D) = -0.58, 95% CI [-0.87, -0.29].  Between. (A) – (C) = 0.30, 95% CI [0.06, 0.53]. (B) – (D) = -0.39, 95% CI [-0.72, -0.05].     1234567Willingness to Advocate1 2 3 4 5 6 7Anticipated AffectNon-Vested VestedOffline1234567Willingness to Advocate1 2 3 4 5 6 7Anticipated AffectNon-Vested VestedOnline171 I turn now to the lower-order interactions. I expected the influence of factors predicting people’s propensity to advocate for their position on a social issue to differ depending on whether that action was to occur in an online or offline context. Each of the following were found to exert differential influence on willingness to advocate in online versus offline contexts (Table 16): social tie strength, β = -0.24, p < .05; guilt proneness, β = -0.70, p < .001; age, β = -0.02, p < .01; and anticipated social support, β = 0.35, p < .01, and anticipated affect, β = -0.30, p < .05, both of which were involved in significant three-way interactions, as described above. The interaction between online/offline context and perceived utility of action was marginally significant as a predictor, β = 0.20, p < .10. In the context of the model reported in Table 17, the effect of the interaction between online/offline context and each predictor was determined with all other variables held constant at their means and vestedness held constant at value zero. For the predictors not involved in significant three-way interactions, vestedness had been shown not to significantly affect the influence of their interaction with online/offline context in predicting willingness to advocate. For anticipated social support and anticipated affect, unqualified effects are reported in the three-way interaction results above. Exploring these interactions with the margins post-estimation command, I found that participants’ perceptions of the strength of their social ties with advocacy audiences exerted positive influence on willingness to act within offline contexts, b = 0.28, CI95 = [0.15, 0.40], but not within online contexts, b = 0.12, CI95 = [-0.01, 0.24]. Similarly, greater guilt proneness predicted greater willingness to advocate within offline contexts, b = 0.50, CI95 = [0.12, 0.89], but not within online contexts, b = -0.21, CI95 = [-0.59, 0.17].  (This finding 172 provides qualified support for my prediction that guilt proneness promotes willingness to act).  The same pattern also emerged for potential advocates’ age, with higher age predictive of greater willingness to advocate within offline contexts, b = 0.02, CI95 = [0.01, 0.03], but not within online contexts, b = -0.00, CI95 = [-0.01, 0.01]. Table 18 presents a summary of factors that significantly affect (i.e., constrain) willingness to advocate, by action context and actor vestment. As shown in that table, many constraints to offline advocacy were not meaningful for either vested or non-vested actors online or, in some cases, were not meaningful for either group. This suggests there are fewer barriers to social action in online contexts.  Table 18: Factors Affecting Willingness to Advocate in Offline and Online Contexts among Social Actors Who are Not Vested and Vested in the Focal Issue (Study 6)  Offline Advocacy Online Advocacy                                      Actors: Not Vested Vested Not Vested Vested Variable     Favorable Attitude toward Issue * * * * Utility of Actiona * * * * Family Vested in Issue * * * * Favorable Societal Attitude: Issue † † † † Direct Action Cost * * * *      Social Tie Strength w/ Audience * *   Guilt Proneness * *   Age * *        Strength of Attitude * * *  Anticipated Positive Affect * *  * Anticipated Social Support * * *b * Proactivity * * *   Note.  Factors affecting willingness to advocate offline but not online comprise potential constraints on social action participation attenuated in online action context.   a Effect significant in main effects model (see Table 16) but not in the conditional effects model (see Table 17). b Among those not vested, effect stronger in online context.   † p < .10. * p < .05. 173 4.2.6 Study 6 Discussion and Post-Hoc Analyses Study 6 examined the influence of several factors that I expected to influence people’s willingness to engage in social action. The social actions considered in Study 6 included advocating one’s position on a social issue to friends in person, via one’s online social networking webpage (e.g., Facebook wall), at a local community meeting, as part of an online community discussion, and via Twitter. Each participant in the study indicated their willingness to engage in each of these advocacy acts for two different issues, and participants were purposively recruited from community groups serving vested groups to ensure variance in issue vestedness. The social issues considered involved medical benefits directly benefitting Aboriginal people (higher social contentiousness, lower courtesy stigma), people living with HIV (lower social contentiousness, higher courtesy stigma), people living with cancer (lower social contentiousness, lower courtesy stigma), and LGBT people (higher social contentiousness, higher courtesy stigma).  I utilized a multilevel maximum likelihood approach in testing my predictions because cases were nested within participants necessitating a multilevel data analysis approach. Like Study 5, which utilized a between groups factorial design and hierarchical regression analysis, Study 6 demonstrated that vestedness is a significant predictor of propensity to advocate for one’s position on a social issue. Also like Study 5, Study 6 demonstrated that perceived utility of action, anticipated social support, and positivity of anticipated affect associated with taking action were all meaningful predictors of propensity to act on a social issue.  174 Study 6 expanded the model predicting willingness to engage in social action to also include perceived social tie strength, social attitude toward the issue, social stigma associated with the various groups vested in the issues considered, having family members vested in the issue, proactive personality score, guilt-proneness, and direct action costs (Table 16, Model 2). Each of these, aside from social stigma and guilt proneness, exerted significant main effects on the criterion. The direction of these significant main effects was each positive except for direct action costs, which exerted negative influence, as expected.  Direct action cost had a strong negative correlation (-.89, p < .001) with online context; the perceived costs of action were much lower online than offline.  In the main effects model predicting willingness to advocate, online context exerted positive influence absent the direct action cost variable was not included (Table 16, Model 1) but negative influence when it was included (Table 16, Model 2).  In a post-hoc analysis, I utilized the Stata ml_mediation command for assessing mediation in multi-level data to determine the indirect effect, direct effect, and total effect of online context in the main effects model (UCLA Institute for Digital Research and Education, 2013a); this approach was adapted from Krull and MacKinnon (2001). In so doing, I treated the other predictors in Model 2 as covariates, engaged the maximum likelihood estimation option, and estimated 95 percent confidence intervals for the effect coefficients through bootstrapping with 1000 replications.  As would be expected given Model 2 results, the indirect effect of online context on willingness to advocate through direct action costs was positive (β = 0.87; CI95 = [.58, 1.13]) and the direct effect of online context on willingness to advocate was negative (β = -.52; CI95 175 = [-.77, -.24]). Importantly, however, the total effect of online context on willingness to advocate was positive (β = .35, CI95 = [.22, .48]).   These findings still beg a question, however. What was it about advocating in online contexts that, did it not require less time, money, and effort investment, would make it less alluring than advocating in offline contexts?  To gain insight into this, I conducted a series of additional post-hoc multi-level mediation analyses, as above, treating each Model 2 factor that related to the action itself (i.e., anticipated affect following the action, anticipated social support for engaging in the action, perceived social tie strength with advocacy audience, and perceived utility of the action) as mediator, in turn.  I controlled for the influence of all Model 2 variables (including direct cost) other than the focal mediator in each run.  I found that social tie strength mediated a significant portion of online context’s negative main effect on willingness to advocate (β = -.47, CI95 = [-.64, -.29]) but anticipated social support (β = -.05, CI95 = [-.10, .01]), perceived utility of action (β = .03, CI95 = [-.01, .08]), and anticipated affect (β = .00, CI95 = [-.02, .02]) each did not.  Thus, in summary, the ability to engage in social action online encouraged social advocacy efforts by reducing the perceived required investment (time, effort, money) required to participate.  Its benefit was reduced, however, predominantly because people did not share as strong of ties with online advocacy audiences as with offline advocacy audiences.  However, there was still more to this story, because the Internet’s effect on social advocacy decisions also manifested at a higher-order, contextualizing level.  Study 6 also tested the interactions among online action context and each of the other predictor variables (i.e., two-way interactions), as well as three-way interactions for each that also included 176 vestedness in the focal issue (see Table 17), and results indicated that the influence, or importance, of several factors determining one’s willingness to engage in social advocacy varied in online versus offline contexts. For example, as I summarized in Table 18, factors such as potential advocates’ age, guilt proneness, and perceived strength of social ties with advocacy audiences that mattered for offline advocacy decisions did not matter for online advocacy decisions.  In other words, those potential requirements for social action offline were eliminated by online context for action.   In conclusion, the combined results from this study provide support for two of my central assertions.  First, online context influenced proclivity toward social advocacy directly by making it easier, cheaper, and less time-consuming to participate. Second, online context influenced proclivity toward social advocacy indirectly by altering (and in many cases eliminating) the influence of other factors that determined how willing people were to engage in social advocacy.   4.3 Chapter 4 Discussion Much of the past research on social action implicitly makes dated assumptions about communication mechanisms, social interaction spaces, and means of organizing.  It does not consider the recent spread of technology that mediates communication, interaction, and concerted action in virtualized forms. Yet online social action is becoming increasingly commonplace, and “empirical studies of online activism are surprisingly scarce” (K. Lewis et al., 2014).  This research investigates how online versus offline context for action impacts willingness to undertake persuasive forms of social action, both directly and indirectly, by 177 altering the importance of a number of other factors that theory and research suggest play an important role in the cognitive evaluations that leading people to undertake social action.  Results from Study 5, a 2 x 2 between-groups experiment with a US sample (n = 381), suggested that the effect of a key predictor of offline advocacy—vestedness, or being directly affected by an issue—differed for parallel online and offline advocacy acts.  In Study 6, I utilized a nested research design (n = 150, cases: N = 1364), with participants recruited from community organizations serving Canadians affected by cancer; Canadians affected by HIV/AIDS; First Nations Canadians; and LGBT Canadians.  Each group was vested in one of four parallel scenarios, and participants each indicated willingness to engage in five advocacy acts in response to (A) the scenario in which their group was vested and (B) one of the other three scenarios at random.  Additionally, Canadians not vested in these focal issues each responded to two scenarios at random.  In all cases, I randomized presentation order, both for scenarios and advocacy acts.  Multilevel regression results showed that for persuasion-based social action such as advocacy to friends and communities, direct costs of participation are significantly lower in online contexts and suggests that people are significantly more willing to engage in comparable social acts in online versus offline environs.  The results also indicated, however, that the influence of online social action context goes well beyond the direct effect of making things easier, cheaper, or less time consuming as a popular term for describing online social action—slacktivism—would suggest. Instead, online social action context also reduced the importance of other focal predictors of social action participation, such as strength of attitude toward the issue, individual proactivity, and 178 anticipated affect. It fully attenuated the roles of potential social actors’ age and proneness to experience guilt, as well as their perceived strength of social ties with their potential advocacy audiences. This research comprised both an experiment and a pseudo-experiment in which some participants were purposively recruited through community organizations to promote variance in a key predictor of social action participation, vestedness. These two approaches allowed me a high degree of control in manipulating and isolating the influence of my focal predictors, promoting internal validity. I also found considerable consistency across the two, despite differences in design and data analysis approach, increasing my confidence in the findings.  Manipulations included social issues relating to a variety of vested groups—men/women, Aboriginal peoples, LGBT individuals, those affected by cancer, and those affected by HIV/AIDS. This served to enrich my data set and test societal factors such as social stigma surrounding vested groups and perceptions of society’s general attitude on focal issues, both of which I expected to influence social action decisions (and the latter of which did). This variety in stimuli also helps to increase the external validity of my findings, making the results more generalizable.   There is ample opportunity to refine and expand upon the current research, however. Van Stekelenburg and Klandermans (e.g., 2007; 2013) advanced a dual-pathways to social action motivation model in which one path goes through ideology, another goes through efficacy, and both converge through group-based anger as the central mediator linking them to motivation. If the current research were overlaid with that mode, it would fall exclusively in 179 the efficacy pathway. Going forward, I will integrate the ideology pathway (roughly analogous to morality in the Van Zomeren, 2013, model) and more thoroughly investigate the role of emotion, particularly as a potential mediator.    180 Chapter  5: Conclusion 5.1 Limitations and Future Directions 5.1.1 Theoretical Boundaries The updates to theory on inequity attenuation that I have proposed in this thesis are not without boundaries.  First and foremost, in this research I seek to reconsider the boundary conditions for action against the more powerful in an era of Internet-mediated social interaction rather than to explain inaction or collusion.  There are, of course, risks to acting against those with more power and more resources, regardless of the context for that action, and the same mechanisms that make the Internet conducive for acts of challenge would also likely facilitate the organization and execution of hierarchy-sustaining acts.  In most cases, however, I contend that through strategic utilization of the Internet, actors can spread or reduce that risk and thus should be more apt to challenge the status quo.  Future research can assess this likelihood.  Next, I argued that the Internet should reduce the effects of fear of punishment as a condition of conformity because it allows a large number of widely dispersed actors to cooperate, thus both increasing the probability of success and spreading the risk of failure.  It is also possible, however, that large numbers could result in diffusion of responsibility.  For example, a would-be actor might often observe large-scale technology-mediated collective social action and, as a result, cultivate an expectation that “they” will act in sufficient numbers such that there is no need for “me” to incur even minimal cost or risk of punishment by joining in that action.  Testing the extent to which this is true and the conditions that make it more or less likely to occur will be an important avenue for 181 subsequent research, because diffusion of responsibility can deter people from acting even on deeply held beliefs (see, e.g., Latane & Darley, 1968).    Relatedly, there also exists the risk of information overload (Goldhaber, 1997), as well as the risk of desensitization, arising from the online calls to social action that help to generate the large numbers of actors necessary increase likelihood of success and lower cost and risk of participation. For example, a constant deluge of requests to support various causes, advance equality, and/or combat the perceived wrongs of powerful individuals, groups, and organizations might result in information that a person ordinarily would act on being overlooked.  Such deluge also ultimately might result in a gradual desensitization to all such calls.  On the other hand, research from a social identity perspective suggests that people are more willing to engage in collective challenge when they feel they cannot work within the current system to transcend the standing associated with their in-groups (Ellemers et al., 1993; Lalonde & Silverman, 1994). Thus, to the extent that these calls to action crystalize within would-be actors a perception that the status quo systematically prevents ‘people like them’ from advancing, their willingness to act together should increase.  Future research exploring and expanding upon these possibilities should provide valuable insights for social action theorists, as well as social movement organizers, marketers, and strategists.  Next, although the Internet should encourage third party action to promote social fairness in general, third parties may become aware of acts of hierarchy-perpetuating acts that that view as illegitimate and grievance and still not engage in countervailing action. Past studies suggest that these individuals are likely to engage in a rationalizing processes that allows them to alleviate cognitive dissonance and maintain their positive self-views (Ashforth 182 & Anand, 2003; Batson, Thompson, & Seuferling, 1999). Through these rationalizing processes, third parties may conclude, for example, that the acts were justified because of the personal attributes or actions of those harmed (i.e., blaming the victim) (Furnham, 2003; Lerner & Miller, 1978; W. Ryan, 1971) or that the outcome received was not actually so bad after all (i.e., denial of harm) (Ashforth & Anand, 2003).  Moreover, because people have a cognitive need to demonstrate consistency in thought and action (Festinger, 1957), these rationalizations may have the added effect of giving rise to subsequent behavior that is decidedly unfavorable to those harmed in hierarchy-sustaining acts (Skarlicki & Kulik, 2005).  Third parties may, for example, distance themselves from, or publically disparage victims or, in some cases, even act in support of the harm-doer (e.g., Batson, Fultz, & Schoenrade, 1987; Brockner & Greenberg, 1990; Stotland, 1969).  Research to determine if these responses are more or less likely in an online context and to explore aspects of communication that encourage or dissuade them would sharpen my model.   Additionally, testing the capacity for common attributes of information disseminated online to make such disengagement more or less likely should have capacity to yield theoretically and practically significant findings.  Examples of such attributes might include perceived general consensus (cf. Mutz, 1998) (e.g., as from 10,000,000 ‘Likes’ of a posting condemning a power-holder’s actions—there is no ‘Dislike’ option); immediacy of action (e.g., ‘join this online movement now’ vs. ‘come to the rally on Saturday’) (Ajzen, 1985; J. Weber & Gillespie, 1998), visibility of action combined with self-presentation motives (e.g., calls for action via Facebook wall posting or public Tweet) (cf. Kristofferson et al., 2014); 183 and vividness/graphicness of information (e.g., cellphone video posted online vs. what could be shown on the evening news) and the emotional reactions it elicits (Douglas, Lyon, & Ogloff, 1997).24 Additionally, it is possible that social action and other forms of resistance undertaken online dissuades actors from engaging in more substantial and ostensibly more meaningful acts (e.g., Gladwell, 2010).  Lewis et al. (2014), for example, analyzed the donation and recruitment activity of the 1.2 million individuals who joined the Save Darfur Facebook page, and finding that few donated to the cause or actively recruited other members, concluded that, “in the case of the Save Darfur campaign, Facebook conjured an illusion of activism rather than facilitating the real thing.”  Perhaps long-term engagement, high visibility action, and expenditures of great amounts of resources—“meaningful support” (cf. Kristofferson et al., 2014) or “the real thing” (cf. Lewis et al., 2014)—is superior, as the catchy but decidedly normative slacktivism label given to the alternative seems to imply. However, it is also possible that a crowdsourced approach to resistance and change in which mass numbers of people do or give a little—for most, perhaps nothing beyond lending their                                                   24 In this study, 120 mock jurors read a detailed trial transcript of a murder trial. The proportion of participants returning a guilty verdict was twice as large among participants provided with photographs of the murder in addition to the written transcript than among those provided only with the written transcript (the control). Non-control participants were more likely to report emotional distress, but participants equally felt that they had acted fairly.  184 click legitimacy to the effort—is equally or, perhaps in some cases, more effective.  Further theorizing and research around that question focused on that question should prove valuable.     5.1.2 Feeding off Grassroots? Formal Advocacy Organizations and (Apparently) Organic Online Social Movements Another opportunity for future research would contribute to an emerging collection of research has begun to examine if (e.g., Bortree & Seltzer, 2009; H. R. Edwards & Hoefer, 2010; Greenberg & MacAulay, 2009) and how (e.g., Guo & Saxton, 2014) social advocacy organizations leverage social media in their causes.  Although work to date has focused on the visible actions of formal social advocacy organizations, those organizations might also strategically utilized technology behind the scenes to facilitate what appear to be fully “concerned citizen movements.” A number of studies in different countries and focused on a number of social issues illustrate that many times people view non-government organizations—entities such as formal advocacy organizations that are neither part of the government nor part of the for-profit business sector—with mistrust or outright hostility (Dave, 2012; Girgen, 2008; Kennelly, 2011; Prakash & Gugerty, 2010; Sissons, 2005; Spasser, 2013; Stiles, 2002).  Moreover, the term “activist” can have a very negative connotation, being analogous to “rabble rouser” or, more bluntly, “trouble maker” (Kennelly, 2011, p. 49).   As such, the grassroots aura of what are, and/or appear to be, the organic responses of everyday people might promote participation and, at least in some cases, bolster the likelihood of success relative to a movement visibly driven by “activists”.  Research could 185 investigate the extent to which backstage action by formal advocacy groups occurs in online social movements and the impacts of that action.  In additional research, the big-picture tradeoffs between traditional mobilization pathways engaged by activist organizations and such a grass-roots approach should also be considered.  For example, as I overviewed in Chapter 1 (1.5.1.1), media outlets may be more attentive to the issues claims of (formal) organizations (see, Andrews & Caren, 2010), which means that organic-feeling movements may receive less media coverage, and (favorable) media coverage can lubricate social change.  On the other hand, to the extent that those who hear about a cause view it as genuine (versus a case of trumped-up politicking) because it is associated with an organic-feeling movement, they may be more likely to sympathize with, and ultimately act on, that cause.  Further, in some cases, movements that are or appear to be grassroots-driven do attain coverage in offline media outlets.  I suggested above (1.5.3) that this crossover likely occurs through the inertia generated by mass endorsement (i.e., click legitimacy).  Formal investigation of that proposition would be valuable for understanding the potential tradeoffs between evidently activist organization led movements versus evidently organic movements. 5.1.3 Virtualized Free Space This dissertation argues that the Internet facilitates easier access to free space.  The development of a Free Space Index is a focal contribution of this dissertation, but future empirical research is necessary to determine if, in fact, people experience free spaces online.  Moreover, research is needed to determine the extent to which these spaces are accessible to and utilized by different groups within societies.  As I described in Section 1.5, material access to the Internet is not universal, even in North America where it is generally prevalent.  186 Instead, what is popularly known as the digital divide segregates those who cannot access the Internet’s promise from those who can (e.g., DiMaggio et al., 2001; Norris, 2001).   As I introduced in Chapter 1, this segregation line has traditionally been seen as a matter of material access (van Dijk, 2012)—the rich can afford Internet-enabled hardware, the poor cannot; those in metropolitan areas have high-speed Internet available, those in rural areas do not; etc.  Although growing availability of inexpensive, Internet-enabled mobile hardware such as tablet computers, iPods, and especially smart phones, along with wide-spread, low-cost, high-speed Internet connectivity through wireless networks is arguably closing this gap (e.g., A. C. T. Smith, 2010; Wray, 2009), there are access limitations beyond those of material access.  As digital divide theorist Jan van Dijk (2005) explicated, people also differ in their capacity to access the Internet in at least three other ways beyond material access disparities.  The first difference in capacities to access the Internet beyond material access relates to mental access.  This refers to differences in predisposition toward accessing the Internet because of psychological and attitudinal (i.e., mental) reasons ranging from individual interest in the Internet and Internet-related hardware to levels of anxiety toward, or comfort with, technology.  Simply put, some people are more mentally predisposed toward accessing the Internet than are others, in general.  It might also be the case that those who are most downtrodden within society are those who are the most distrustful of the Internet, with their anxiety functionally precluding them from accessing Internet-mediated free spaces. Next, people can differ in skills access – some people have adequate education and/or social support to develop sufficient digital skills to utilize the Internet to a higher degree than 187 those who have not had the opportunity to develop those skills.  These skills are multifaceted and relate both to medium and content.  According to van Dijk (2012), medium-related skills include operational skills (ability to functionally operate a digital medium; knowing “what buttons to push”) and formal skills (ability to handle a medium’s formal structures in order to access content). Content-related skills include information skills (ability to search, select, and evaluate information in digital form); communication skills (ability to convey information, contact others, create online identities, give opinions, draw attention, etc.); content-creation skills (ability to create a design and implement a plan to contribute content online); and strategic skills (ability to leverage digital media to achieve specific goals) (van Dijk, 2012).  It seem that content-related skills would be particularly important in creating and/or accessing Internet-mediated free spaces and that strategic skills, in combination with content-creation and communication skills, would be key in creating successful action from attitudes and intentions born within those free spaces. Finally, people can differ in usage access.  From a social structuring perspective, it is not only the use of the Internet in a general sense that matters but also the uses to which it is put, and as a result of social structures already in place, distributions of socially significant or meaningful Internet use does not appear to be even.  For example, in 2010 van Deursen and van Dijk reported that for the first time in history Dutch people with low levels of education were using the Internet more hours in their leisure time than were the more educated (van Dijk, 2012).  Upon further inspection, however, those authors found that of all the possible uses to which the Internet might be put, online gaming and chatting were the only ones consuming more hours for the less educated than for the more educated.  Different patterns 188 of Internet usage among various groups within societies should differentially expose them to Internet-mediated free spaces.  As such, understanding how these groups tend to use the Internet should provide a good deal of insight into why some might utilize the Internet to promote their cause to a greater extent than do others.     Beyond expanding on and testing the implications of the above, future research should also assess such things as the frequency with which free spaces are experienced online versus offline, both in general and by different populations and sub-populations.  The frequency with which acts to challenge power or counteract abuses of power arise from free spaces, both in general and in regard to specific types of inequality attenuating action, should be determined.  Moreover, individual, communal, structural, and situational mediators and moderators of the relationship between free spaces and challenge acts should be investigated.   For example, an online interaction space that is generally considered a free space and would ordinarily be at risk of spawning challenge acts might be especially dangerous. This point is well illustrated by a 2013 interview reported in Esquire (i.e. Junod, 2014) in which Glenn Greenwald, a member of the core group of journalists who first worked with Edward Snowden to report the widespread monitoring of electronic communications by the US National Security Administration (NSA) (e.g., MacAskill, Borger, & Greenwald, 2013), said this of the protected space: It’s the place that uniquely enables us to explore limits, to test boundaries, to engage in novel and creative ways of thinking and being. Only if we feel free of the kind of judgmental eyes of others are we able to try different things out, 189 to experiment, to evolve, to free ourselves of mores that are imposed on us or conventional orthodoxies about how we’re supposed to behave and think. […] The promise of the Internet has always been that it was gonna [sic] be this unprecedentedly potent instrument of liberation and democratization. That it would empower people to band together to work against oppression. That it would let you explore things and meet people who you wouldn’t otherwise get to know in completely free and unconstrained ways. And what has happened instead is that we face the threat that it’s the exact opposite—that instead the Internet could become the most potent and odious tool of human control and oppression in human history. (p. 60) These points are well made, to a point.  It is important to remember, however, that resistance can be enacted against a variety of powerful entities including not only the state but also firms, markets, cultural systems, and powerful groups or individuals (see, e.g., de Bakker, den Hond, King, & Weber, 2013). The extent to which online resistance and resistance-building should be subject to any special dangers as compared to that occurring offline, it seems, should be a function of the potential target meeting six criteria: having the (1) capacity, (2) ability, and (3) inclination to electronically monitor online communications and being able to (4) analyze data from the myriad and diverse communications that occur online, (5) identify potential threats to itself based on that analysis, and (6) identify the true identities and/or locations of resisters once a threat has been identified. Nation states that have a strong technology infrastructures in place, a legal environment allowing widespread surveillance and data mining, and a well-trained, organized, and motivated group of analysts 190 dedicated to the task likely exemplify such resistance targets.25 In some cases (e.g., if resistance against other targets comprise illegal activity), nation states may also extend their capabilities to other targets. To the extent that those six criteria are met by the target or are likely to be extended to it by another powerful entity, free space that is fully “off the grid” may be safer and more conducive to resistance and successful change efforts. Moreover, if the resistance target meets those criteria without resisters realizing it, they may feel a false sense of invulnerability within the online environment that makes failure and/or punishment especially likely. 5.1.3.1 Identifying and Characterizing Virtualized Free Space Another promising avenue for future research is to explore likely free spaces in virtualized contexts.  One could execute a study in which participants evaluate popular virtual interaction spaces in North America (e.g., Facebook, Google+, Myspace, Youtube, Twitter, Linkedin, Reddit…) using the Free Space Index.  Data from this study would allow for the ranking of these spaces in terms of freedom for resistance for the study’s focal social issues, providing some insight into which are the most fruitful digital wombs for resistance.  The sample for this study could be recruited from multiple cultures so as to allow for                                                   25 However, as was highlighted in 1.5.2, Internet-savvy subversives (i.e., those with high skills access in this domain) can conceal their identities and physical locations in a number of ways, some of which are quite sophisticated. This makes criterion 6 less certain even among entities meeting criteria 1 through 5 and means that skills access may be the ultimate determinant of online free space for resistance against these entities. 191 comparisons of Free Space Index responses based on both cultural differences.  So, too, could comparisons be made based on participants’ individual differences (e.g., gender identity, racial identity, generation/age, socio-economic status, sexual orientation, education level, comfort with technology, trust in authority) to identify differences in the ways that various groups assess free space in general and to determine if different groups have divergent perceptions of the freeness of each focal interaction space in the study.  In conducting such a study, care should be taken to ensure that no result is reported in such a way as to perpetuate stereotypes or encourage stereotypical thinking.  5.2 Implications  This dissertation has important implications for organizational structuring.  As I briefly alluded to in framing this research, the patterns of domination, legitimation, and significance institutionalized within structures guide actors’ interpretations of objects, events, and other actors within their environments (Giddens, 1979). To the extent that these interpretations and the actions they engender are perfect reflections of the current structure, that structure is reinforced and perpetuated (Barley, 1986).  However, the acts of power, communication, and moral sanctioning that comprise and perpetuate structures are interactions among entities with free agency, and thus, deviations from a structure’s institutional template are likely in the course of everyday life.   To the extent that deviations are momentary, isolated, and random or can be subsumed by the existing framework, they are unlikely to significantly alter the existing structure (Meyer, 1982).  However, when these “slippages” persist, “they become replicated patterns whose contours depart, perhaps ever so slightly, from former practices” (Barley, 192 1986, p. 80). As these new patterns become a part of actors’ conceptualization of “the way things are” (Berger & Luckmann, 1967, p. 61), a new template emerges for the way things should be.  As I have described, inputs from the physical realm can be input into the Internet realm, processed and acted upon as guided by Internet realm structures, and interjected back into the physical realm. The actions flowing out of the Internet realm that otherwise would not occur are objects in physical realm structures that become a part of the set of objects, actors, and interactions that (re)define those structures. Additionally, the actors in Internet realm structures are concurrently part of physical realm structures, and experiences those actors have in the Internet realm (perhaps that they would not have had in the physical realm) may lead them to see hierarchy-sustaining acts as illegitimate  whereas they otherwise would have seen them as legitimate.  Such changes to actors’ physical realm cognitive schemas and action sets would make them “new” (i.e., different) actors in those physical realm structures.  Because structures are constantly negotiated by their actors, “new” actors flowing out of the Internet realm may alter subsequent physical realm structures, as would be the case, for example, if increased threat of punishment constrains patterns of social interaction that had previously acted to sustain the prevailing structure.  These “new” actors may also model behavior that changes the action set considered to be legitimate by other actors and thus alter other actors’ action sets, as well.  In these ways, the Internet has the potential to mediate organizational and social restructuring. 193 This dissertation also has important implications for organizations, more generally.  These implications are something that United Airlines, for example, might attest because of its experience with Dave Carroll and his “United Breaks Guitars” music video (see Chapter 2 introduction). 26  Despite his tremendous power- and resource- disadvantages, Carroll leveraged technology to challenge what he perceived to be unfair treatment at the hands of the far more powerful corporation and, ultimately, to impose his will on that more powerful entity.  As that case demonstrated, the greater risk for illegitimate actions mediated by virtualization technologies had potential to alter organizational outcomes and behaviors. Although Study 4 results did not bear this out, I believe that future research in this area is definitely warranted, and I for one plan to rerun Study 4 post-dissertation with a sample of executive MBA students in whom I can have greater confidence will have the knowledge necessary to process its business language and strategic nuances. 5.3 Summary In this research, I advanced an updated model of inequality attenuation that takes into consideration changes in social interaction and communication arising from the introduction of Internet technology and the virtual free spaces that it houses.  I argued that through the Internet, the influence of traditional conditions of conformity such as resource disadvantages, fear of punishment, and the institutionalized legitimacy of the status quo is diminished.  An effective communication network is the fundamental resource required for                                                   26 This case was the basis for the Internet risk manipulation in Study 4. 194 successful action against those with more power, and while access to that resource was limited in the past, that is largely no longer the case.  Moreover, the Internet decreases the influence of resource disadvantages by making crowdsourcing possible; said differently, through the Internet, a lot of people in many different places can each contribute a bit of resources that may ultimately be sufficient to overcome the resources of the more powerful entity.   The constraining capacity of fear of punishment is diminished because the Internet allows for easy access to virtual free spaces—spaces that are protected from public observation and which have been shown capable of birthing rebellion. It is also diminished because it allows people to act anonymously. Finally, Internet-mediated networking and communication allows for large numbers of participants in widely dispersed areas to participate in action to challenge the more powerful, and this should both increase the likelihood of the challenge succeeding and spreads the risk far and wide in the case of failure.   The constraining force of the institutionalized legitimacy of the status quo, I argued, can be reduced by the Internet because free spaces allow for alternative perspectives to be cultivated and social networking and social media allows for messages framed with those alternative perspectives to be disseminated broadly.  Institutionalized legitimacy beliefs can be challenged when alternative perspectives—alternative views of how things should (or even could) be—are introduced to the system.  Also, when a message achieves significant following online, it can cross over into traditional media, presenting an opportunity for the alternative (i.e., delegitimizing) perspective to reach other potential actors, complete with legitimacy implied by a story being “news.”     195 I further argued that the Internet has important implications for decisions by those at higher levels of hierarchies to act in favor of those at lower levels.  Through the Internet, those who are at higher levels of hierarchies but who are sympathetic to the causes of those at lower levels can act for their benefit outside the view of their social circles and thus avoid much of the social punishment that would otherwise dissuade them from acting.  Moreover, the Internet allows for communication to be packaged in rich media that should allow it to elicit strong emotional reactions; emotions play an important role in translating free space interactions into challenge acts, and the emotional reactions of those with power, in particular, tend to spur them into action. Next, I described a new threat associated with engaging in hierarchy-sustaining acts that should attenuate inequality by discouraging those acts a priori.  Specifically, I argued that members of more powerful groups who engage in hierarchy-sustaining acts that are perceived as illegitimate (e.g., coercion, discrimination) are at risk of those actions being made public through Internet-mediated C-form organizations (e.g., Facebook, Twitter, YouTube). I further argued that once these behaviors become public, a collective, negative, and potentially aggressive response by members of the subordinated group and its allies can arise.   Subsequently, I reported a three-pronged empirical program (a summary list of hypotheses tested in this dissertation is presented in Appendix F).  The first prong comprised the development and validation of the Free Space Index, a necessary step in facilitating future research arising from this dissertation. The index should also be a valuable contribution in its own right as a tool that researchers in a variety of fields can also utilize.  196 In this prong, I first generated an item pool to assess the comparative safety of a given social space from six categories of risk that the literature suggests deters resistance and, with help from a group of doctoral students trained in organizational theory and briefed on the free space construct, refined those items.  I then executed three studies to further refine the items and specify and test the measurement model. In the first of these, Study 1, I constructed a matrix in which definitions of each of the six risk dimensions were listed along the top and 60 potential items were listed in the rows.  I then presented this matrix to participants (n = 56), who rated the fit of each item to each risk category.  In Study 2 (n = 78), I selected best items from those retained in Study 1 and specified and tested a measurement model, following best practices and incorporating both formative and reflective aspects to yield a hybrid index that does not suffer the same shortcomings as would a purely formative index.  In Study 3, I further assessed the validity of the index using a nomological network approach.  Study 3 data were collected at parallel social interaction spaces in Canada (n = 123), Denmark (n = 100), and the USA (n = 92), which allowed for the index to be cross-culturally validated, as well. The second prong comprised Study 4.  In Study 4, I tested a central proposition of my dissertation—that the Internet can mediate power inequality attenuation by increasing the risk associated with leveraging power illegitimately and thereby reduce the prevalence of such acts.  This data provided support for none of my Study 4 hypotheses, and although it is possible that my predictions in this domain do not align with reality, the direction of effects found were consistent with my predictions.  I proposed that the null findings for this study may be misleading.  More specifically, the sample comprised participants recruited through 197 the Mechanical Turk data panel, and although this data panel has been shown conducive to social and organizational research in general, the manipulations and measures in this study were highly contextualized and might have been better suited to a sample whose knowledge of business strategy could be more readily assured.  Alternatively, null results may have arisen from the sample being insufficiently large (n = 160 across 4 cells in a 2x2 factorial) to verify effects that were present with statistical certainty.  Post-hoc power analysis through Monte-Carlo simulation based on the Study 4 data set suggested that the likelihood of falsely confirming null hypotheses neared 90 percent (simulated power = .113).  Further research is warranted, I argued, before discarding this proposition. The third prong of empirics tested another of my central propositions:  The Internet can increase the likelihood that both directly affected and empathetic parties will undertake action to advance the interests of those in a disadvantaged or subordinated group.  Results from Study 5, a 2 x 2 between-groups experiment with a US sample (n = 381), suggested that the effect of a key predictor of offline advocacy—vestedness, or being directly affected by an issue—differed for parallel online and offline advocacy acts.  In Study 6, I utilized a nested research design (n = 150, cases: N = 1364), with participants recruited from community organizations serving Canadians affected by cancer; Canadians affected by HIV/AIDS; First Nations Canadians; and LGBT Canadians.  Each group was vested in one of four parallel scenarios, and participants each indicated willingness to engage in five advocacy acts in response to (A) the scenario in which their group was vested and (B) one of the other three scenarios at random.  Additionally, Canadians not vested in these focal issues each responded to two scenarios at random.  In all cases, I randomized presentation order, both 198 for scenarios and advocacy acts.  Multilevel regression results showed that online context promoted social action not only by reducing the effort required to act but also by attenuating the effects of a number of socio-psychological factors that could constrain willingness to act in traditional, offline contexts. Finally, I presented some boundaries for my theory and discussed the process through which changes in interaction patterns that I argue the Internet could facilitate could influence organizational structures.  Like the superhero’s mask and the supremacist’s hood, the Internet can provide a sense of protection that empowers people to act—for better or for worse—whereas they otherwise likely would not (see, e.g., publicshaming.tumblr.com).  Contrary to the “cyber-libertarian” perspective, which argues that the Internet will “inevitably stifle government restrictions, destroy hierarchical forms of authority, and free up the exchange of information and ideas worldwide” (Deibert, 2002, p. 143), I acknowledge that the negotiation of power is an iterative process (e.g., Francisco, 1996), and that electronic technologies can be used to reinforce power (e.g., Burris, Smith, & Strahm, 2000; Gallagher, 2013; Kowalski, Giumetti, Schroeder, & Lattanner, 2014) or shuffle it from one powerful institution to another (see, Benac, 2013, for one example), as well as to challenge it.  The powerful have always been able to harm the less powerful, however, whereas the capacity for the less powerful to band together with relatively little effort and at very little expense to counteract acts of harm are emergent.   Ultimately, no society or organization exists without some form of hierarchy (Sidanius & Pratto, 1999), but I argue that contemporary communication and coordination technologies presents new opportunities for restructuring—reordering, flattening, inequality 199 attenuation—to occur.  Research such as that which I have proposed that takes this into account should advance knowledge of organizational structuring, social power negotiation, social action, activism, social movements, collective action, and similar phenomena.  In conclusion, the presence of Internet technology in an organizational or social structure should help to attenuate inequality because it allows for easy access to virtual free spaces, with virtually no cost for communication and coordination, and because it otherwise enables low power actors and their more powerful allies to overcome some of the traditional conditions of conformity within hierarchical organizational structures.  The Internet should also help to attenuate inequality because it mediates a new form of countervailing action.  The increased- and new- risks of countervailing action should raise the perceived risk associated with engaging in hierarchy-sustaining acts and the broad communication through social media of past hierarchy-sustaining acts and ensuing countervailing action should decrease the perceived legitimacy and safety of subsequent action to perpetuate inequality.    200 References AAVAZ.org. (2013, October 31). The world's worst punishment for rape?  Retrieved November 12, 2013, from http://www.avaaz.org/en/justice_for_liz_loc/?twi ABC News. (2010, October 29). School board member's anti-gay rant.  Retrieved April 20, 2013, from http://abcnews.go.com/GMA/video/arkansas-school-district-board-members-anti-gay-rant-12002856 Afuah, A., & Tucci, C. L. (2012). Crowdsourcing as a solution to distant search. Academy of Management Review, 37(3), 355-375. Agre, P. (1998). The Internet and public discourse. First Monday: Peer-Reviewed Journal of the Internet, 3(3). Retrieved December 3, 2012, from http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/issue/view/90 Aiken, L. S., & West, S. G. (1991). Multiple Uegression: Testing and Lnterpreting Lnteractions. Newbury Park, CA: Sage. Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. New York: Springer. Ajzen, I. (1991). The theory of planned behavior. Theories of Cognitive Self-Regulation, 50(2), 179-211. Alberici, A. I., & Milesi, P. (2013). The influence of the Internet on the psychosocial predictors of collective action. Journal of Community & Applied Social Psychology, 23(5), 373-388. Alter, A. L., Oppenheimer, D. M., & Zemla, J. C. (2010). Missing the trees for the forest: A construal level account of the illusion of explanatory depth. Journal of Personality and Social Psychology, 99(3), 436-451. 201 Anderson, C., & Berdahl, J. L. (2002). The experience of power: Examining the effects of power on approach and inhibition tendencies. Journal of Personality and Social Psychology, 83(6), 1362-1377. Anderson, C., & Galinsky, A. D. (2006). Power, optimism, and risk-taking. European Journal of Social Psychology, 36(4), 511-536. Anderson, C., John, O. P., & Keltner, D. (2012). The personal sense of power. Journal of Personality, 80(2), 313-344. Anderson, C., & Spataro, S. (2005). Misperceiving your place: Humility and hubris in social hierarchies. In M. A. Neale, E. A. Mannix & M. C. Thomas-Hunt (Eds.), Research in managing groups and teams: Status and groups (pp. 69-92). Greenwich, CT: Elsevier. Anderson, C., Srivastava, S., Beer, J. S., Spataro, S. E., & Chatman, J. A. (2006). Knowing your place: Self-perceptions of status in face-to-face groups. Journal of Personality and Social Psychology, 91(6), 1094-1110. Andrews, K. T., & Caren, N. (2010). Making the news: Movement organizations, media attention, and the public agenda. American Sociological Review, 75(6), 841-866. Antony, M. G., & Thomas, R. J. (2010). ‘This is citizen journalism at its finest’: YouTube and the public sphere in the Oscar Grant shooting incident. New Media & Society, 12(8), 1280-1296. Arendt, H. (1970). On violence. New York: Harcourt. Arkansas Matters. (2010). Signed doc.  Retrieved April 19, 2013, from http://arkansasmatters.com/images/Multi_Media/arkansasmatters/nxd_media/dox/pdf/2010_11/Signed%20Doc.pdf 202 Arnold, J. C. (2013). How the Internet destroys community. Beliefnet: Inspiration. Spirituality. Faith. Retrieved January 3, 2014, from http://www.beliefnet.com/Faiths/Christianity/2006/03/How-The-Internet-Destroys-Community.aspx Arrow, K. J. (1951). Social choice and individual values. New Haven, CT: Yale University Press. Arthur, C. (2010, March 25). China and the Internet: Tricks to beat the online censor. The Guardian. Retrieved March 8, 2013, from http://www.guardian.co.uk/world/2010/mar/25/china-internet-how-to-beat-censorship?intcmp=239 Ashforth, B. E., & Anand, V. (2003). The normalization of corruption in organizations. Research in Organizational Behavior, 25, 1-52. Associated Press. (2010, October 27). Arkansas school official accused of anti-gay posting.  Retrieved April 19, 2013, from http://www.foxnews.com/us/2010/10/27/arkansas-school-official-accused-anti-gay-screed/ Atkins, P., & De Paula, J. (2006). Atkins' physical chemistry (Vol. 8). Oxford, UK: Oxford University Press. Atkins, P., & De Paula, J. (2012). Elements of physical chemistry. Oxford, UK: Oxford University Press. Avery, H. E. (1974). Basic reaction kinetics and mechanisms. New York: Macmillan. Azzam, T. I., Beaulieu, D. A., & Bugental, D. B. (2007). Anxiety and hostility to an "outsider," as moderated by low perceived power. Emotion, 7(3), 660. Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall. 203 Bargh, J. A., & McKenna, K. Y. A. (2004). The Internet and social life. Annual Review of Psychology, 55, 573-590. Barley, S. R. (1986). Technology as an occasion for structuring - Evidence from observations. Administrative Science Quarterly, 31(1), 78-108. Barley, S. R. (1991). Contextualizing conflict: Notes on the anthropology of disputes and negotiations. In M. Bazerman, R. Lewicki & B. H. Sheppard (Eds.), Research on negotiation in organizations (Vol. 3, pp. 165–202). Greenwich, CT: JAI Press. Barnard, C. (1968). The functions of the executive (18th ed.). Oxford, UK: Oxford University Press. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. Barton, B. (2010). "Abomination" - Life as a Bible Belt gay. Journal of Homosexuality, 57, 465-484. Bateman, T. S., & Crant, J. M. (1993). The proactive component of organizational behavior: A measure and correlates. Journal of Organizational Behavior, 14, 103-118. Batson, C. D., Fultz, J., & Schoenrade, P. A. (1987). Distress and empathy - 2 qualitatively distinct vicarious emotions with different motivational consequences. Journal of Personality, 55(1), 19-39. Batson, C. D., Thompson, E. R., & Seuferling, G. (1999). Moral hypocrisy: Appearing moral to oneself without being so. Journal of Personality and Social Psychology, 77, 525-537. 204 Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117(3), 497-529. BBC. (2011, March 13). Benefits e-petition hits 100,000. BBC News. Retrieved August 11, from http://www.bbc.co.uk/news/uk-politics-14474429 BBC. (2013, February 22). North Korea to offer mobile Internet access. BBC News. Retrieved March 11, 2013, from http://www.bbc.co.uk/news/technology-21546877 Beach, L. R., & Mitchell, T. R. (1978). A contingency model for the selection of decision strategies. Academy of Management Review, 3, 439-449. Benac, N. (2013, April 01). Obama image machine churns out rosy words, images, video as traditional press access narrows. The Associated Press. Retrieved April 2, 2013, from http://www.washingtonpost.com/politics/obama-image-machine-churns-out-rosy-words-images-video-as-traditional-press-access-narrows/2013/04/01/bc8be1b0-9ae2-11e2-9219-51eb8387e8f1_story_1.html Berdahl, J. L. (2007). Harassment based on sex: Protecting social status in the contest of gender hierarchy. Academy of Management Review, 32(2), 641-658. Berger, P. L., & Luckmann, T. (1967). The social construction of reality. Garden City, NY: Doubleday. Bergkvist, L., & Rossiter, J. R. (2007). The predictive validity of multiple-item versus single-item measures of the same constructs. Journal of Marketing Research, 44(2), 175-184. Bernhein, B. D. (1994). A theory of conformity. Journal of Political Economy, 102(5), 841-877. Bernoff, J., & Schadler, T. (2010). Empowered. Harvard Business Review, 88(7/8), 94-101. 205 Blalock, H. M. (1968). Theory building and causal inferences. In H. M. Blalock & A. Blalock (Eds.), Methodology in social research (pp. 155-198). New York: McGraw-Hill. Boland, W. A., Brucks, M., & Nielsen, J. H. (2012). The attribute carryover effect: What the “runner-up” option tells us about consumer choice processes. Journal of Consumer Research, 38(5), 872-885. Bollen, K. A., & Davis, W. R. (2009). Causal indicator models: Identification, estimation, and testing. Structural Equation Modeling, 16, 498-522. Bollen, K. A., & Jackman, R. W. (1990). Regression diagnostics: An expository treatment of outliers and influential cases. In J. Fox & J. S. Scott (Eds.), Modern methods of data analysis (pp. 257-291). Newbury Park, CA: Sage. Bollen, K. A., & Lennox, R. (1991). Conventional wisdom on measurement: A structrual equation perspective. Psychological Bulletin, 110(2), 305-314. Bollen, K. A., & Ting, K. A. (2000). Tetrad tests for causal indicators. Psychological Methods, 5(1), 3-22. Borgatti, S. P., & Foster, P. C. (2003). The network paradigm in organizational research: A review and typology. Journal of Management, 29(6), 991-1013. Bortree, D. S., & Seltzer, T. (2009). Dialogic strategies and outcomes: An analysis of environmental advocacy groups' Facebook profiles. Public Relations Review, 35, 317-319. Bourhis, R. Y. (1994). Power, gender, and intergroup discrimination - Some minimal group experiments. Psychology of Prejudice : The Ontario Symposium, 7, 171-208. 206 Boyle, C. (2014, January 22). Why 2014 is 'tipping point' for Internet: Marissa Mayer, Yahoo CEO. CNBC. Retrieved February 7, 2014, from www.cnbc.com/id/101353869 Branigan, T. (2012, April 30). Chen Guangcheng's nephew flees in fear for his life: Chen Kegui said to be in hiding as Chinese dissident's escape to US embassy reveals bitter human cost to family. The Guardian. Retrieved June 1, 2013, from http://www.theguardian.com/world/2012/apr/30/chen-guangcheng-nephew-flees Brockner, J., & Greenberg, J. (1990). The impact of layoffs on survivors: An organizational justice perspective. In J. S. Carroll (Ed.), Applied social psychology and organizational settings (pp. 45–75). Hillsdale, NJ: Erlbaum. Broverman, N. (2010). Arkansas school board member thinks fags should die. The Advocate. Retrieved April 17, 2013, from http://www.advocate.com/news/news-features/2010/10/26/arkansas-school-board-member-thinks-fags-should-die Brown, B. B., Clasen, D. R., & Eicher, S. A. (1986). Perceptions of peer pressure, peer conformity dispositions, and self-reported behavior among adolescents. Developmental Psychology, 22(4), 521. Brunstein, J. C., & Gollwitzer, P. M. (1996). Effects of failure on subsequent performance: The importance of self-defining goals. Journal of Personality and Social Psychology, 70(2), 395-407. Brunsting, S., & Postmes, T. (2002). Social movement participation in the digital age. Small Group Research, 33(5), 525-554. Bugental, D. B., Blue, J., & Cruzcosa, M. (1989). Perceived control over caregiving outcomes: Implications for child abuse. Developmental Psychology, 25, 532-539. 207 Buhrmester, M., Kwang, T., & Gosling, S. M. (2011). Amazon's Mechanical Turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3-5. Burke, P. (2014). A social history of knowledge II: From the encyclopaedia to Wikipedia: John Wiley & Sons. Burris, V. A. L., Smith, E., & Strahm, A. N. N. (2000). White supremacist networks on the Internet. Sociological Focus, 33(2), 215-235. Burt, R. S. (1992). Structural holes. Cambridge, MA: Harvard University Press. Butler, K. (2013, July 18). Jimmy Carter on NSA: 'America has no functioning democracy.' Retrieved February 3, 2014, from http://www.upi.com/blog/2013/07/18/Jimmy-Carter-on-NSA-America-has-no-functioning-democracy/6541374165224/ Cabantous, L., Gond, J. P., & Johnson-Cramer, M. (2010). Decision theory as practice: Crafting rationality in organizations. Organization Studies, 31(11), 1531-1566. Canadian Cancer Society. (2013). Cancer statistics at a glance.  Retrieved November 5, 2013, from http://www.cancer.ca/en/cancer-information/cancer-101/cancer-statistics-at-a-glance/?region=on Carroll, G. R. (1985). Concentration and specialization: Dynamics of niche width in populations of organizations. American  Journal of Sociology, 90, 1262-1283. Carroll, G. R., & Khessina, O. M. (2005). The ecology of entrepreneurship. In S. A. Alvarez, R. Agarwal & O. Sorenson (Eds.), Handbook of entrepreneurship research: Disciplinary perspectives (pp. 167-200). New York: Springer. 208 Chan, D. (1998). Functional relations among constructs in the same content domain at different levels of analysis. Journal of Applied Psychology, 83, 234-246. Change.org. (2011). South Africa: Take action to stop 'corrective rape'.  Retrieved December 6, 2011, from http://www.change.org/petitions/south-africa-take-action-to-stop-corrective-rape Chatman, J. A., Polzer, J. T., Barsade, S. G., & Neale, M. A. (1998). Being different yet feeling similar: The influence of demographic composition on work processes and outcomes. Administrative Science Quarterly, 43, 749-780. Chee-Sing, C. (2011). No place to hide. Enterprise Innovation, 7(1), 4-4. Christensen, H. S. (2011). Political activities on the Internet: Slacktivism or political participation by other means? First Monday: Peer-Reviewed Journal of the Internet, 16(2). Retrieved November 3, 2013, from http://firstmonday.org/ojs/index.php/fm/article/view/3336/2767 Clair, J. A., Beatty, J. E., & MacLean, T. L. (2005). Out of sight but not out of mind: Managing invisible social identities in the workplace. Academy of Management Review, 30(1), 78-95. Clemens, E. S., & Cook, J. M. (1999). Politics and institutionalism: Explaining durability and change. Annual Review of Sociology, 25, 441-466. Clinton, H. R. (2011). Remarks in recognition of international Human Rights Day.  Retrieved December 8, 2011, from http://www.state.gov/secretary/rm/2011/12/178368.htm 209 CNN Wire Staff. (2010a, November 1). Arkansas school board member officially resigns over anti-gay post. CNN U.S. Retrieved April 19, 2013, from http://www.cnn.com/2010/US/11/01/arkansas.anti.gay.resignation/index.html CNN Wire Staff. (2010b, October 28). Arkansas school board member to resign over anti-gay post. CNN U.S. Retrieved November 1, 2012, from www.cnn.com/2010/US/10/28/arkansas.anti.gay.resignation/index.html CNN Wire Staff. (2010c, October 28). Arkansas schools chief condemns anti-gay posts by school board member. CNN U.S. Retrieved April 19, 2013, from http://www.cnn.com/2010/US/10/28/arkansas.anti.gay.bullying/index.html Cohen, L. E. (1997). Deus ex machina?  A study of how jobs are designed. (Ph.D. Dissertation), University of California, Berkeley, CA.   Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98(2), 310-357. Cooley, C. H. (1922). Human nature and the social order. New York: Scribners. Cooper, A. (2010a, October 29). Anderson Cooper 360°: CNN: Arne Duncan on anti-bullying efforts.  Retrieved April 18, 2013, from http://www.cnn.com/video/?/video/bestoftv/2010/10/29/ac.duncan.on.bullying.intv.cnn#/video/bestoftv/2010/10/29/ac.duncan.on.bullying.intv.cnn Cooper, A. (2010b, October 28). Anderson Cooper 360°: CNN: Clint McCance talks with Anderson Cooper.  Retrieved January 10, 2011, from http://www.youtube.com/watch?v=IQ8hknilBCs 210 Cooper, A. (2010c, October 29). Anderson Cooper 360°: CNN: Ellen DeGeneres comments on Clint McCance.  Retrieved April 18, 2013, from http://www.youtube.com/watch?v=Lxe3VOpgDkI Cooper, A. (2010d, October 27). Anderson Cooper 360°: CNN: Whistleblower Anthony Turner shocked by Clint McCance anti-gay rant Retrieved January 10, 2011, from http://www.youtube.com/watch?v=Eo1NDQj9mok Corrales, J. (2010). The LGBT organizational density of world cities: The gay-friendliness index. In J. Corrales & M. Pecheny (Eds.), Sexuality in latin America: A reader on lesbian, gay, bisexual, and transgender rights (pp. 372-381). Pittsburgh, PA: University of Pittsburgh Press. Costarelli, S. (2006). The distinct roles of subordinate and superordinate group power, conflict, and categorization on intergroup prejudice in a multiethnic Italian territory. Journal of Social Psychology, 146(1), 5-13. Crant, J. M. (2000). Proactive behavior in organizations. Journal of Management, 26(3), 435-462. Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52(4), 281-302. Curtis, R. F., & Jackson, E. F. (1962). Multiple indicators in survey research. American Journal  of Sociology, 68, 195-204. Danova, T. (2013, October 2). Morgan Stanley: 75 billion devices will be connected to the internet of things by 2020. Business Insider. Retrieved February 2, 2014, from http://www.businessinsider.com/75-billion-devices-will-be-connected-to-the-internet-by-2020-2013-10 211 Dave, N. N. (2012). Queer activism in India: A story in the anthropology of ethics. Durham, NC: Duke University Press. Davidson, P., & Lieberman, D. (2010, March 16). FCC wants more access to affordable high-speed Internet. USA Today. Retrieved September 12, 2011, from http://www.usatoday.com/money/industries/technology/2010-03-16-1Afcc16_ST_N.htm de Bakker, F. G. A., den Hond, F., King, B., & Weber, K. (2013). Social movements, civil society and corporations: Taking stock and looking ahead. Organization Studies, 34(5-6), 573-593. Debatin, B., Lovejoy, J. P., Horn, A.-K., & Hughes, B. N. (2009). Facebook and online privacy: Attitudes, behaviors, and unintended consequences. Journal of Computer-Mediated Communication, 15(1), 83-108. Deibert, R. J. (2002). Dark guests and great firewalls: The Internet and Chinese security policy. Journal of Social Issues, 58(1), 143-159. DeJordy, R. (2008). Just passing through: Stigma, passing, and identity decoupling in the workplace. Group & Organization Management, 33(5), 504-531. Diamantopoulos, A., Riefler, P., & Roth, K. P. (2008). Advancing formative measurement models. Journal of Business Research, 61(12), 1203-1218. Diamantopoulos, A., & Siguaw, J. A. (2006). Formative versus reflective indicators in organizational measure development: A comparison and empirical illustration. British Journal of Management, 17, 263-282. 212 Diamantopoulos, A., & Winklhofer, H. (2001). Index construction with formative indicators: An alternative to scale development. Journal of Marketing Research, 37, 269-277. Diani, M. (2000). Social movement networks virtual and real. Information, Communication & Society, 3(3), 386-401. DiMaggio, P., Hargittai, E., Neuman, W. R., & Robinson, J. P. (2001). Social implications of the Internet. Annual Review of Sociology, 27(1), 307-336. Douglas, K. S., Lyon, D. R., & Ogloff, J. R. (1997). The impact of graphic photographic evidence on mock jurors' decisions in a murder trial: probative or prejudicial? Law and Human Behavior, 21(5), 485. Downs, A. (1957). An economic theory of democracy. New York: Harper & Row. Drennan, J., & Previte, J. (2006). Privacy, risk perception, and expert online behavior: An exploratory study of household end users. Journal of Organizational and End User Computing, 18(1), 1-22. Druckman, J. N., Fein, J., & Leeper, T. J. (2012). A source of bias in public opinion stability. American Political Science Review, 106(2), 430-454. Dunne, D. (2010). United breaks guitars: A case study. Retrieved June 1, 2013, from http://ddunne.com/wp-content/uploads/2012/07/United-Breaks-Guitars-Case-Final.pdf Earl, J. (2006). Pursuing social change online. Social Science Computer Review, 24(3), 362-377. Earl, J., Kimport, K., Prieto, G., Rush, C., & Reynoso, K. (2010). Changing the world one webpage at a time: Conceptualizing and explaining Internet activism. Mobilization: An International Quarterly, 15(4), 425-446. 213 Earl, J., McKee Hurwitz, H., Mejia Mesinas, A., Tolan, M., & Arlotti, A. (2013). This protest will be Tweeted. Information, Communication & Society, 16(4), 459-478. Edwards, H. R., & Hoefer, R. (2010). Are social work advocacy groups using Web 2.0 effectively? Journal of Policy Practice, 9, 220-239. Edwards, J. R. (1995). Alternatives to difference scores as dependent variables in the study of congruence in organizational research. Organizational Behavior and Human Decision Processes, 64, 307-324. Edwards, J. R. (2001). Multidimensional constructs in organizational behavior research: An integrative analytical framework. Organizational Research Methods, 4(2), 141-192. Edwards, J. R. (2011). The fallacy of formative measurement. Organizational Research Methods, 14(2), 370-388. Ellemers, N., Wilke, H., & Vanknippenberg, A. (1993). Effects of the legitimacy of low group or individual status on individual and collective status-enhancement strategies. Journal of Personality and Social Psychology, 64(5), 766-778. Emerson, R. M. (1962). Power-dependence relations. American Sociological Review, 27(1), 31-41. Entman, R. M. (2007). Framing bias: Media in the distribution of power. Journal of Communication, 57(1), 163-173. Evans, S. (1979). Personal politics. New York: Vintage Books. Evans, S., & Boyte, H. C. (1986). Free spaces: The sources of democratic change in America. New York: Harper & Row. Facebook.com. (2010a). Fire Clint McCance.  Retrieved December 4, 2011, from http://www.facebook.com/pages/Fire-Clint-McCance/170421922968484 214 Facebook.com. (2010b). People United Against Clint McCance.  Retrieved April 19, 2013, from https://www.facebook.com/pages/People-United-Against-Clint-McCance/162285590458739 Facebook.com. (2013). Justice for Trayvon Martin.  Retrieved December 4, 2013, from http://www.facebook.com/jforTrayvonMartin Facebook.com. (2014). Boycott 2014 Winter Olympic Games in Russia.  Retrieved January 2, 2014, from http://www.facebook.com/Boycott2014WinterOlympicGamesinRussia Fantasia, R., & Hirsch, E. L. (1995). Culture in rebellion: The appropriation and transformation of the vein in the Algerian revolution. In H. Johnston & B. Klandermans (Eds.), Social movements and culture (pp. 144-159). Minneapolis: University of Minnesota Press. Fast, N. J., & Chen, S. (2009). When the boss feels inadequate: Power, incompetence, and aggression. Psychological Science, 20, 1406-1413. Fayers, P. M., Hand, D. J., Bjordal, K., & Groenvold, M. (1997). Causal indicators in quality of life research. Quality of Life Research, 6, 393-406. Feather, N. T. (1982). Introduction and overview. In N. T. Feather (Ed.), Expectations and actions: Expectancy-value models in psychology. Hillsdale, NJ: Erlbaum. Feather, N. T., & Newton, J. W. (1982). Values, expectations, and the prediction of social action: An expectancy-valence analysis. Motivation and Emotion, 6(3), 217-244. Felin, T., & Foss, N. J. (2005). Strategic organization: A field in search of micro-foundations. Strategic Organization, 3(4), 441-455. 215 Ferree, M. M. (2004). Soft repression: Ridicule, stigma, and silencing in gender-based movements. Research in Social Movements, Conflicts and Change, 25, 85-101. Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row, Peterson. Finchilescu, G. (2010). Intergroup anxiety in interracial interaction: The role of prejudice and metastereotypes. Journal of Social Issues, 66(2), 334-351. Fisher, R. A. (1971 [1935]). The design of experiments (9th ed.). New York: MacMillan. Fiske, S. T. (1993). Controlling other people - The impact of power on stereotyping. American Psychologist, 48(6), 621-628. Folger, R., & Martin, C. (1986). Relative deprivation and referent cognitions: Distributive and procedural justice effects. Journal of Experimental Social Psychology, 22, 532-546. Fornell, C., & Bookstein, F. L. (1982). Two structural equation models: LISREL and PLS applied to consumer exit-voice theory. Journal of Marketing Research, 18, 39-50. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39-50. Fox, S., & Raine, L. (2014, February 27). The Web at 25 in the U.S.  Retrieved February 27, 2014, from http://www.pewinternet.org/2014/02/27/part-1-how-the-internet-has-woven-itself-into-american-life/ Francisco, R. A. (1996). Coercion and protest: An empirical test in two democratic states. American Journal of Political Science, 40(4), 1179. Fraser, M., & Dutta, S. (2008, November 19). Barack Obama and the Facebook election. US News and World Report. Retrieved December 23, 2011, from 216 http://www.usnews.com/opinion/articles/2008/11/19/barack-obama-and-the-facebook-election Free Press. (2013, June 30). Vladimir Putin signs anti-gay propaganda bill. The Telegraph. Retrieved February 28, 2014, from http://www.telegraph.co.uk/news/worldnews/europe/russia/10151790/Vladimir-Putin-signs-anti-gay-propaganda-bill.html French, J. R. P., & Raven, B. H. (1959). The bases of social power. In D. Cartwright & A. Zander (Eds.), Group dynamics (pp. 150-167). New York: Harper & Row. Furnham, A. (2003). Belief in a just world: research progress over the past decade. Personality and Individual Differences, 34(5), 795-817. Galinsky, A. D., Gruenfeld, D. H., & Magee, J. C. (2003). From power to action. Journal of Personality and Social Psychology, 85(3), 453-466. Gallagher, R. (2013, February 10). Software that tracks people on social media created by defence firm. The Guardian. Retrieved January 19, 2014, from http://www.guardian.co.uk/world/2013/feb/10/software-tracks-social-media-defence Gamson, W. A. (1996). Safe spaces and social movements. Perspectives on Social Problems, 8, 27-38. Gans, H. (2005). Deciding what's news: A study of CBC Evening News, NBC Nightly News, Newsweek, and Time (2 ed.). Evanston, IL: Northwestern University Press. Gans, H. (2011). Multiperspectival news revisited: Journalism and representative democracy. Journalism, 12(1), 3-13. 217 Ghose, A., & Han, S. P. (2014, in press). Estimating demand for mobile applications in the new economy. Management Science. Retrieved February 25, from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2378007 Gibson, C. B., Cooper, C. D., & Conger, J. A. (2009). Do you see what we see? The complex effects of perceptual distance between leaders and teams. Journal of Applied Psychology, 94, 62-76. Giddens, A. (1979). Central problems in social theory. Berkeley, CA: University of California Press. Gil de Zúñiga, H., Jung, N., & Valenzuela, S. (2012). Social media use for news and individuals' social capital, civic engagement and political participation. Journal of Computer Mediated Communication, 17, 319-336. Girgen, K. (2008). Constructing animal rights activism as social threat: Claims-making in the New York Times and in Congressional Hearings. (Doctor of Philosophy), Florida State University (Paper 4269), Tallahassee, FL.  Retrieved February 2, 2014, from http://diginole.lib.fsu.edu/cgi/viewcontent.cgi?article=1714&context=etd Gladwell, M. (2010). Small change. New Yorker, 86(30), 42-49. Goffman, E. (1963). Stigma: Notes on the management of spoiled identity. New York: Simon and Schuster. Goldhaber, M. H. (1997). The attention economy and the Net. First Monday: Peer-Reviewed Journal of the Internet, 2(4). Retrieved February 10, 2013, from http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/519/440 218 Goldman, B. M. (2001). Toward an understanding of employment discrimination claiming: An integration of organizational justice and social information processing theories. Personnel Psychology, 54(2), 361-386. Graham, C. C. (2012). Teachers' explicit and implicit attitude toward homosexuality: The role of internal and external motivation to respond without prejudice. (Doctorate in Social Work), University of Pennsylvania (Paper 29), Philadelphia, PA.  Retrieved February 10, 2014, from http://repository.upenn.edu/cgi/viewcontent.cgi?article=1031&context=edissertations_sp2 Greenberg, J., & MacAulay, M. (2009). NPO 2.0? Exploring the web presence of environmental nonprofit organizations in Canada. Global Media Journal-Canadian Edition, 2, 63-88. Guo, C., & Saxton, G. D. (2014). Tweeting social change: How social media are changing nonprofit advocacy. Nonprofit and Voluntary Sector Quarterly, 43(1), 57-79. Guzzo, R. A., & Shea, G. P. (1992). Group performance and intergroup relations in organizations. In M. D. Dunnette & L. E. Hough (Eds.), Handbook of industrial and organizational psychology (2nd ed., pp. 269-313). Palo Alto, CA: Consulting Psychologists Press. Haidt, J. (2003). The moral emotions. In R. J. Davidson, K. R. Scherer & H. H. Goldsmith (Eds.), Handbook of affective sciences (pp. 852–870). Oxford: Oxford University Press. Hair, J. F., Jr. (2013). Using the SmartPLS software: Assessment of measurement models.  Retrieved July 20, 2013, from http://tinyurl.com/Hair-SmartPLS 219 Hair, J. F., Jr., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis. New Jersey: Prentice Hall. Hair, J. F., Jr., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2014). A primer on partial least squares structural equation modeling (PLS-SEM). Thousand Oaks, CA: Sage. Hammond, J. (2009, July 16). United Airlines and customer generated turbulence. NielsenWire. Retrieved May 7, 2013, from http://www.nielsen.com/us/en/newswire/2009/united-airlines-and-consumer-generated-turbulence.html Handley, S. M., & Benton Jr, W. C. (2012). Mediated power and outsourcing relationships. Journal of Operations Management, 30(3), 253-267. Hannan, M. T., & Freeman, J. (1987). The ecology of organizational founding: American labor unions, 1836-1985. American Journal of  Sociology, 92(4), 910-943. Harder, D., & Lewis, S. J. (1987). The assessment of shame and guilt. In J. N. Butcher & C. D. Spielberger (Eds.), Advances in personality assessment (Vol. 6, pp. 89-114). Hillsdale, NJ: Lawrence Erlbaum. Harder, D., & Zalma, A. (1990). Two promising shame and guilt scales: a construct validity comparison. Journal of  Personality Assessment, 55(3-4), 729-745. Harlow, S., & Harp, D. (2012). Collective action on the web:  A cross-cultural study of social networking sites and online and offline activism in the United States and Latin America. Information Communication & Society, 15(2), 196-216. 220 Harvey, C. G., Stewart, D. B., & Ewing, M. T. (2011). Forward or delete: What drives peer-to-peer message propagation across social networks? Journal of Consumer Behaviour, 10(6), 365-372. Hathcoat, J., & Barnes, L. (2010). Explaining the relationship among fundamentalism and authoritarianism: An epistemic connection. International Journal for the Psychology of Religion, 20, 73-84. Haug, C. (2013). Organizing spaces: Meeting arenas as a social movement infrastructure between organization, network, and institution. Organization Studies, 34(5-6), 705-732. Hauser, R. M. (1973). Diaggregating a social-psychological model of education attainment. In A. S. Goldberger & O. D. Duncan (Eds.), Structural equation models in the social sciences (pp. 255-289). San Diego: Academic Press. Hauser, R. M., & Goldberger, A. S. (1971). The treatment of unobservable variables in path analysis. Sociology Methodology, 3, 81-117. Hawkins, R. (2011, August 4). What is the point of e-petitions? BBC News, 2011. Retrieved December 9, 2011, from http://www.bbc.co.uk/news/uk-politics-14409032 Hayes, K. (2010, October 27). Clint McCance: Anti-gay Facebook comments allegedly made by Ark. school board member. CBS News. Retrieved April 20, 2013, from http://www.cbsnews.com/8301-504083_162-20020960-504083.html Healey, K. M. (1995). Victim and witness intimidation: New developments and emerging responses (Vol. 2). Washington, DC: US Department of Justice, Office of Justice Programs, National Institute of Justice. 221 Herek, G. M. (1987). Religious orientation and prejudice: A comparison of racial and sexual attitudes. Personality and Social Psychology Bulletin, 13(1), 34-44. Herek, G. M. (1990). The context of anti-gay violence: Notes on cultural and psychological heterosexism. Journal of Interpersonal Violence, 5(3), 316-333. Herek, G. M. (2000). The psychology of sexual prejudice. Current Directions In Psychological Science, 9(1), 19-22. Herek, G. M. (2004). Beyond "homophobia": Thinking about sexual prejudice and stigma in the twenty-first century. Sexuality Research and Social Policy, 1, 6-24. Herek, G. M. (2007). Confronting sexual stigma and prejudice: Theory and practice. Journal of Social Issues, 63, 905-925. Herring, S. C., Stein, D., & Virtanen, T. (2013). Handbook of pragmatics of computer-mediated communication. Berlin: Mouton. Heussner, K. M. (2010a, October 20). Facebook goes purple for Spirit Day. ABC News. Retrieved April 19, 2013 from  http://abcnews.go.com/Technology/spirit-day-facebook-users-wear-purple/story?id=11928210#.UXK15YLR2C8 Heussner, K. M. (2010b, November 1). Official resigning over anti-gay comments. ABC News. Retrieved April 19, 2013 from  http://abcnews.go.com/Technology/cries-grow-louder-anti-gay-school-board-official/story?id=11995495 222 Hickson, D. J., Hinings, C. R., Lee, C. A., Schneck, J. M., & Pennings, J. M. (1971). A strategic contengencies' theory of interorganizational power. Administrative Science Quarterly, 16(2), 216-229. Hindman, M. (2009). The myth of digital democracy. Oxford, UK: Princeton University Press. Hinkin, T. R., & Tracey, J. B. (1999). An analysis of variance approach to content validation. Organizational Research Methods, 2(2), 175-186. Hirsh, J. B., Galinsky, A. D., & Zhong, C. B. (2011). Drunk, powerful, and in the dark: How general processes of disinhibition produce both prosocial and antisocial behavior. Perspectives on Psychological Science, 6(5), 415-427. Hollenbeck, C. R., & Zinkhan, G. M. (2006). Consumer activism on the Internet: The role of anti-brand communities. Advances in Consumer Research, 33(1), 479-485. Hollinger, R. C., & Clark, J. P. (1982). Formal and informal social controls of employee deviance. Sociological Quarterly, 23(3), 333-343. Holt, R. (2013, March 21). Twitter in numbers. The Telegraph. Retrieved January 3, 2014, from http://www.telegraph.co.uk/technology/twitter/9945505/Twitter-in-numbers.html Horney, K. (1939). New ways in psychoanalysis. New York: W. W. Norton. Hornsey, M. J., Spears, R., Cremers, I., & Hogg, M. A. (2003). Relations between high and low power groups: The importance of legitimacy. Personality and Social Psychology Bulletin, 29(2), 216-227. Howe, J. (2006, June 1). The rise of crowdsourcing. Wired. Retrieved November 18, 2012, from http://www.wired.com/wired/archive/14.06/crowds.html 223 Howe, J. (2008). Crowdsourcing: Why the power of the crowd is driving the future of business. New York: Crown Business. Howell, R. D., Breivik, E., & Wilcox, J. B. (2007). Reconsidering formative measurement. Psychological Methods, 12(2), 205-218. Huang, J., Curran, P., Keeney, J., Poposki, E., & DeShon, R. (2012). Detecting and deterring insufficient effort responding to surveys. Journal of Business and Psychology, 27(1), 99-114. Hudon, M., & Sandberg, J. (2013). The ethical crisis in microfinance: Issues, findings, and implications. Business Ethics Quarterly, 23(4), 561-589. Human Rights Campaign Foundation. (2012). 2012 Municipal Equality Index. Retrieved November 12, 2013, from http://www.hrc.org/files/assets/resources/MEI-2012_rev.pdf Hunjan, S., & Towson, S. (2007). 'Virginity is everything': Sexuality in the context of intimate partner violence in the South Asian community. In S. Das Dasgupta (Ed.), Body evidence: Intimate violence against south Asian women in America (pp. 53-67). Piscataway, NJ: Rutgers University Press. Iacobucci, D. (2010). Structural equations modeling: Fit indices, sample size, and advanced topics. Journal of Consumer Psychology, 20, 90-98. Ibarra, H. (1992). Homophily and differential returns: Sex-differences in network structure and access in an advertising firm. Administrative Science Quarterly, 37(3), 422-447. Internet World Stats. (2012, June 30). World Internet users and population stats.  Retrieved November 12, 2012, from http://www.internetworldstats.com/stats.htm 224 Jacob, D. J. (2011). Introduction to atmospheric chemistry. Princeton, NJ: Princeton University Press. John, O. P., & Benet-Martinez, V. (2000). Measurement: Reliability, construct validation, and scale construction. In H. T. Tries & C. M. Judd (Eds.), Handbook of research method in social and personality psychology (pp. 339-369). New York: Cambridge University Press. Johns, G. (2006). The essential impact of context on organizational behavior. Academy of Management Review, 31(2), 386-408. Jones, M. (2010). Demand resignation of anti-gay Arkansas school board member.  Retrieved April 19, 2013, from http://www.change.org/petitions/demand-resignation-of-anti-gay-arkansas-school-board-member Jost, J. T., & Banaji, M. R. (1994). The role of stereotyping in system-justification and the produciton of falce consciousness. British Journal of Social Psychology, 33, 1-27. Jost, J. T., Pelham, B. W., & Carvallo, M. R. (2002). Non-conscious forms of system justification: Implicit and behavioral preferences for higher status groups. Journal of Experimental Social Psychology, 38(6), 586-602. Judge, T. A., & Bretz, R. D. (1994). Political influence behavior and career success. Journal of Management, 20(1), 43. Junod, T. (2014). Glenn Greenwald: What I've learned. Esquire, 161, 60. Jutras, L. (2009, December 7). Facebook mobilizes masses-- But what for? The Globe and Mail. Retrieved February 10, 2013, from http://www.theglobeandmail.com/technology/facebook-mobilizes-masses---but-what-for/article4294760/ 225 Kang, C. (2010, April 15). Library of Congress plan for Twitter: A big, permanent retweet. The Washington Post. Retrieved November 27, 2011, from http://www.washingtonpost.com/wp-dyn/content/article/2010/04/15/AR2010041505752.html Karger, H. J. (2004). Scamming the poor. The Social Policy Journal, 3(1), 39-54. Katz, D., & Kahn, R. L. (1978). The social psychology of organizations. New York: Wiley. Kay, A. C., Gaucher, D., Peach, J. M., Laurin, K., Friesen, J., Zanna, M. P., & Spencer, S. J. (2009). Inequality, discrimination, and the power of the status quo: Direct evidence for a motivation to see the way things are as the way they should be. Journal of Personality and Social Psychology, 97(3), 421-434. Kay, A. C., & Jost, J. T. (2003). Complementary justice: Effects of "poor but happy" and "poor but honest" stereotype exemplars on system justification and implicit activation of the justice motive. Journal of Personality and Social Psychology, 85(5), 823-837. Keen, A. (2007). The cult of the amateur: How today’s Internet is killing our culture and assaulting our economy. New York: Doubleday. Keltner, D., Ellsworth, P. C., & Edwards, K. (1993). Beyond simple pessimism - Effects of sadness and anger on social perception. Journal of Personality and Social Psychology, 64(5), 740-752. Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, approach, and inhibition. Psychological Review, 110(2), 265-284. Kennedy, P. A. (2003). Guide to econometrics (5th ed.). Boston: MIT Press. 226 Kennelly, J. (2011). Citizen \outh: Culture, Dctivism, and Dgency in a Qeoliberal Hra. New York: Palgrave Macmillian. King, B. G. (2008). A Political Mediation Model of Corporate Response to Social Movement Activism. Administrative Science Quarterly, 53(3), 395-421. King, B. G. (2011). The Tactical Disruptiveness of Social Movements. Social Problems, 58(4), 491-517. Kipnis, D. (1976). The powerholders. Chicago: University of Chicago Press. Klandermans, P. G. B. (1984a). Mobilization and participation in trade union action: An expectancy-value approach. Journal of Occupational Psychology, 57(2), 107-120. Klandermans, P. G. B. (1984b). Mobilization and participation: Social-psychological expansisons of resource mobilization theory. American Sociological Review, 49(5), 583-600. Koon, D. (2010, October 29). Update: No resignation yet from Clint McCance. Arkansas Times. Retrieved April 19, 2013, from http://www.arktimes.com/ArkansasBlog/archives/2010/10/29/midland-superintendent-no-word-from-mccance Koon, D. (2011, March 2). Facebook, Clint McCance and a teaching moment about hate. The Arkansas Times. Retrieved April 19, 2013, from http://www.arktimes.com/arkansas/facebook-clint-mccance-and-a-teaching-moment-about-hate/Content?oid=1573608 Kowalski, R. M., Giumetti, G. W., Schroeder, A. N., & Lattanner, M. R. (2014, in press). Bullying in the digital age: A critical review and meta-analysis of cyberbullying 227 research among youth. Psychological Bulletin. Retrieved February 11, 2014, from http://psycnet.apa.org/psycinfo/2014-04307-001/ Krikorian, R. (2013, August 16). New Tweets per second record, and how! Twitter Engineering Blog. Retrieved February 27, 2014, from https://blog.twitter.com/2013/new-tweets-per-second-record-and-how Kristofferson, K., White, K., & Peloza, J. (2014). The nature of slacktivism: How the social observability of an initial act of token support affects subsequent prosocial action. Journal of Consumer Research, 40, 1149-1166. Kroll, J., & Egan, E. (2004). Psychiatry, moral worry, and moral emotions. Journal of Psychiatric Practice, 10, 352-360. Krull, J. L., & MacKinnon , D. P. (2001). Multilevel modeling of individual and group level mediated effects. Multivariate Behavioral Research, 36(2), 249-277. Kuhn, A. (2009, July 24). Twitter seen as tool for social change in China. NPR. Retrieved December 8, 2011, from http://www.npr.org/templates/story/story.php?storyId=106963768  Kurzban, R., & Leary, M. R. (2001). Evolutionary origins of stigmatization: The functions of social exclusion. Psychological Bulletin, 127(2), 187-208. Lalonde, R. N., & Silverman, R. A. (1994). Behavioral preferences in response to social injustice: The effects of group permeability and social identity salience. Journal of Personality and Social Psychology, 66(1), 78-85. Land, K. (1970). On estimation of path coefficients for unmeasured variables from correlations among observed variables. Social Forces, 48, 506-511. 228 Langton, N. (1987). Niche theory and social movements - A population ecology approach. Sociological Quarterly, 28(1), 51-70. Latane, B., & Darley, J. M. (1968). Group inhibition of bystander intervention. Journal of Personality and Social Psychology, 10, 215-221. Leberknight, C. S., Chiang, M., & Wong, F. M. F. (2012). A taxonomy of censors and anti-censors: Part 1 - Impacts of Internet censorship. International Journal of E-Politics, 3(2), 52-64. Lee-St. John, J. (2005, October 25). Viewpoint: Civil rights as gay rights. Time. Retrieved May 25, 2013, from http://www.time.com/time/nation/article/0,8599,1121811,00.html Lee, N. (2013). E-Government, E-Activism. Facebook nation (pp. 115-146). Boston: Springer. Lennox, R. D., & Wolfe, R. N. (1984). Revision of the self-monitoring scale. Journal of Personality and Social Psychology, 46(6), 1349-1364. Leonard, C. (2009, September 1). In defense of "slacktivism". Retrieved February 10, 2013, from http://bayercenter.wordpress.com/2009/09/01/in-defense-of-slactivism/ Lerner, M. J., & Miller, D. T. (1978). Just world research and attribution process: Looking back and ahead. Psychological Bulletin, 85(5), 1030-1051. Lewin, K. (1951). Constructs in field theory. In D. Cartwright (Ed.), Field theory in social science: Selected theoretical papers (pp. 30-42). New York: Harper. Lewis, G. B. (2003). Black-white differences in attitudes toward homosexuality and gay rights. The Public Opinion Quarterly, 67(1), 59-78. Lewis, K., Gray, K., & Meierhenrich, J. (2014). The structure of online activism. Sociological Science, 1, 1-9. 229 Lomi, A. (1995). The population and community ecology of organizational founding: Italian cooperative bans. European Sociological Review, 11, 75-98. MacAskill, E., Borger, J., & Greenwald, G. (2013, June 6). The National Security Agency: Surveillance giant with eyes on America. The Guardian. Retrieved March 21, 2014, from http://www.theguardian.com/world/2013/jun/06/national-security-agency-surveillance Machain, C. M., Clifton, M. T., & Regan, P. M. (2011). Deterring rebellion. Foreign Policy Analysis, 7(3), 295-316. MacKenzie, S. B., Podsakoff, P. M., & Jarvis, C. B. (2005). The problem of measurement model misspecification in behavioral and organizational research and some recommended solutions. Journal of Applied Psychology, 90, 710-730. Mackenzie, S. B., Podsakoff, P. M., & Podsakoff, N. P. (2011). Construct measurement and validation procedures in MIS and behavioral research: Integrating new and existing techniques. M I S Quarterly, 35(2), 293-A295. Madsen, T. L., Woolley, J. L., & Sarangee, K. (2012). Using Internet-based collaboration technologies for innovation: Crowdsourcing vs. expertsourcing. Paper presented at the 72nd Annual Meeting of the Academy of Management, Boston, MA.  Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self-reinforcing nature of power and status. The Academy of Management Annals, 2, 351 - 398. Mail Foreign Service. (2010a, October 30). 'I'm sorry I made those ignorant comments': School board member's grovelling TV apology over Facebook post urging gay youths to commit suicide. Daily Mail (UK). Retrieved April 18, 2013, from 230 http://www.dailymail.co.uk/news/article-1324954/Clint-McCances-apologises-Facebook-post-urging-gay-youths-commit-suicide.html Mail Foreign Service. (2010b, October 29). School board member’s Facebook post urged “sinful” gay youths to commit suicide. Daily Mail (UK). Retrieved April 18, 2013, from http://www.dailymail.co.uk/news/article-1324450/Clint-McCances-Facebook-post-urged-gay-youths-commit-suicide.html Mann, M. (1986). The sources of social power: A history of power from the beginning to AD 1760. Cambridge, UK: Cambridge University Press. Marsden, P. V., & Campbell, K. E. (1984). Measuring tie strength. Social Forces, 63(2), 482-501. Martin, J. (1992). Cultures in organizations: Three perspectives. Oxford, UK: Oxford University Press. Martin, T., & Fulton, S. (2012, April 15). Prosecute the killer of our son, 17-year-old Trayvon Martin.  Retrieved December 12, 2013, from http://www.change.org/petitions/prosecute-the-killer-of-our-son-17-year-old-trayvon-martin Martinez-Torres, M. R., & Diaz-Fernandez, M. C. (2013). Current issues and research trends on open-source software communities. Technology Analysis & Strategic Management, 26(1), 55-68. Martorana, P. V., Galinsky, A. D., & Rao, H. (2005). From system justification to system condemnation: Antecedents of attempts to change power hierarchies. In M. A. Neale, 231 E. A. Mannix & M. C. Thomas-Hunt (Eds.), Research in managing groups and teams: Status and groups (pp. 285-315). Greenwich, CT: Emerald Group. Mason, T. D. (1996). Insurgency, counterinsurgency, and the rational peasant. Public Choice, 86, 63-83. Matsuda, C. R., Lawrence, I., R. D., & Crenshaw, K. W. (Eds.). (1993). Words that wound: Critical race theory, assaultive speech, and the First Amendment. Boulder, CO: Westview. McAdam, D. (1986). Recruitment to high-risk activism: The case of Freedom Summer. American  Journal of Sociology, 92, 64-90. McCarthy, J., & Zald, M. N. (1973). The trend of social movements. Morristown, IN: General Leaming. McCarthy, J., & Zald, M. N. (1977). Resource mobilization and social movements: A partial theory. American Journal of Sociology, 82, 1212-1241. McDonald, S., & Ahem, K. (2000). The professional consequences of whistleblowing by nurses. Journal of Professional Nursing, 16(6), 313-321. McGee, J. S. (1958). Predatory price cutting: The Standard Oil (N. J.) case. Journal of Law and Economics, 1, 137-169. McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415-444. Mechanic, D. (1962). Sources of power of lower participants in complex organizations. Administrative Science Quarterly, 7(3), 349-364. Merton, R. K. (1936). The unanticipated consequences of purposive social action. American Sociological Review, 1(6), 894-904. 232 Meyer, A. D. (1982). Adapting to environmental jolts. Administrative Science Quarterly, 27(4), 515-537. Mintzberg, H. (1983). Power in and around organizations. Englewood Cliffs, NJ: Prentice Hall. Mollick, E. (2014). The dynamics of crowdfunding: An exploratory study. Journal of Business Venturing, 29(1), 1-16. Morozov, E. (2009, September 5). From slacktivism to activism. Foreign Policy. Retrieved January 2, 2013, from http://neteffect.foreignpolicy.com/posts/2009/09/05/from_slactivism_to_activism Morozov, E. (2011). The dark side of internet freedom: The net delusion. New York: PublicAffairs. Mueller, D. C. (1979). Public choice. Cambridge, UK: Cambridge University Press. Mukherjee, D. (2010). An exploratory study of older adults' engagement with virtual volunteerism. Journal of Technology in Human Services, 28(3), 188-196. Mutz, D. C. (1998). Impersonal influence: How perceptions of mass collectives affect political attitudes. Cambridge, UK: Cambridge University Press. Nagy, M. S. (2002). Using a single-item approach to measure facet job satisfaction. Journal of Occupational and Organizational Psychology, 75(1), 77-86. Ng, S. H., & Cram, F. (1988). Intergroup bias by defensive and offensive groups in majority and minority conditions. Journal of Personality and Social Psychology, 55(5), 749-757. Norris, P. (2001). Digital divide: Civic engagement, information, poverty, and the Internet in democratic societies. New York: Cambridge University Press. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw Hill. 233 O'Mahony, S. (2007). What does it mean to be community managed? Journal of Management and Governance, 2, 139-150. O'Mahony, S., & Bechky, B. A. (2008). Boundary organizations: Enabling collaboration among unexpected allies. Administrative Science Quarterly, 53(3), 422-459. O'Reilly, J., & Aquino, K. (2011). A model of third parties’ morally-motivated responses to mistreatment in organizations. Academy of Management Review, 36(3), 526-543. Oliveira, M. (2013, July 19). Facing backlash, Tim Hortons apologizes for blocking gay and lesbian news website on WiFi. The Globe and Mail. Retrieved July 25, 2013, from http://www.theglobeandmail.com/report-on-business/tim-hortons-faces-backlash-for-blocking-gay-and-lesbian-news-website-on-wifi/article13320893/ Oliver, P. E. (1989). Bringing the crowd back in: The nanoorganizational elements of social movements. Research in Social Movements, Conflicts and Change, 11, 1-30. Olmstead, K., Jurkowitz, M., Mitchell, A., & Enda, J. (2013, October 11). How Americans get TV news at home. PewResearch Journalism Project. Retrieved March 1, 2014, from http://www.journalism.org/2013/10/11/how-americans-get-tv-news-at-home/ Olson, M. (1965). The logic of collective action. Cambridge, MA: Harvard University Press. Ostroff, C., Atwater, L. E., & Feinberg, B. J. (2004). Understranding self-other agreement: A look at rater and ratee characteristics, context, and outcomes. Personnel Psychology, 57, 333-375. Ostroff, C., & Schmitt, N. (1993). Configurations of organizational effectiveness and efficiency. Academy of Management Journal, 36(6), 1345-1361. 234 Parmerlee, M. A., Near, J. P., & Jensen, T. C. (1982). Correlates of whistleblowers' perceptions of organizational retaliation. Administrative Science Quarterly, 27(1), 17-34. Parsons, O. A., & Schneider, J. M. (1974). Locus of control in university students from Eastern and Western societies. Journal of Consulting and Clinical Psychology, 42(3), 456. Payne, J. W. (1982). Contingent decision behavior. Psychological Bulletin, 92(2), 382-402. Peters, E. M., Burraston, B., & Mertz, C. K. (2004). An emotion-based model of risk perception and stigma susceptibility: Cognitive appraisals of emotion, affective reactivity, worldviews, and risk perceptions in the generation of technological stigma. Risk Analysis, 24(5), 1349-1367. Petter, S., Straub, D., & Rai, A. (2007). Specifying formative constructs in information systems research. M I S Quarterly, 31(4), 623-656. Polletta, F. (1999). 'Free spaces' in collective action. Theory and Society, 28, 1-38. Postmes, T. (2007). The psychological dimensions of collective action, online. In A. N. Joinson, K. Y. A. McKenna, T. Postmes & U. D. Reips (Eds.), The Oxford handbook of Internet psychology (pp. 165-184). Oxford, UK: Oxford University Press. Postmes, T., & Brunsting, S. (2002). Collective Action in the Age of the Internet: Mass Communication and Online Mobilization. Social Science Computer Review, 20(3), 290-301. Prakash, A., & Gugerty, M. K. (2010). Advocacy organizations and collective action. Cambridge, UK: Cambridge University Press. 235 Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. F. (1994). Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67(4), 741-763. Preece, J., & Maloney-Krichmar, D. (2003). Online communities. In J. Jacko & A. Sears (Eds.), Handbook of human-computer interaction (pp. 596-620). Mahwah, NJ: Lawrence Erlbaum. Priento, L. C. (2011). The influence of proactive personality on social entrepreneurial intentions among African-American and Hispanic undergraduate students: The moderating role of hope. Academy of Entrepreneurship Journal, 17(2), 77-96. Public Health Agency of Canada. (2011, November 8). Canada AIDS survey report.  Retrieved September 12, 2013, from http://www.phac-aspc.gc.ca/aids-sida/publication/survreport/2011/dec/index-eng.php Rabe-Hesketh, S., & Skrondal, A. (2012). Multilevel and logitudinal modeling using STATA (3rd ed. Vol. I). College Station, TX: Stata Press. Ragins, B. R. (2008). Disclosure disconnects: Antecedents and consequences of disclosing invisible stigmas across life domains. Academy of Management Review, 33(1), 194-215. Ragins, B. R., & Cornwell, J. M. (2001). Pink Triangles: Antecedents and consequences of perceived workplace discrimination against gay and lesbian employees. Journal of Applied Psychology, 86(6), 1244-1261. Ragins, B. R., Singh, R., & Cornwell, J. M. (2007). Making the invisible visible: Fear and disclosure of sexual orientation at work. Journal of Applied Psychology, 92(4), 1103-1118. 236 Raine, L., Purcell, K., & Smith, A. (2011, January 18). The social side of the Internet | Pew Internet & American life project.  Retrieved 4 November, 2013, from http://www.pewinternet.org/Reports/2011/The-Social-Side-of-the-Internet.aspx Rao, H., & Dutta, S. (2012). Free spaces as organizational weapons of the weak: Religious festivals and regimental mutinies in the 1857 Bengal Native Army. Administrative Science Quarterly, 57 (4), 625-668. Ratner, R. K., & Miller, D. T. (2001). The norm of self-interest and its effects on social action. Journal of Personality and Social Psychology, 81(1), 5-16. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods. Thousand Oaks, CA: Sage Publications. Reed, A., & Aquino, K. (2003). Moral identity and the expanding circle of moral regard toward out-groups. Journal of Personality and Social Psychology, 84(6), 1270-1286. Reno, J. (2013, December 20). A new model for global change. International Business Times . Retrieved February 25, 2014, from http://www.ibtimes.com/changeorg-petitions-cover-issues-sarah-palin-human-trafficking-rape-lesbians-south-africa-new-model Richard, R., van der Pligt, J., & de Vries, N. (1996). Anticipated affect and behavioral choice. Basic & Applied Social Psychology, 18 (2), 111-129. Richman, L. S., & Leary, M. R. (2009). Reactions to discrimination, stigmatization, ostracism, and other forms of interpersonal rejection: A multimotive model. Psychological Review, 116(2), 365-383. Ringle, C. M., Wende, S., & Will, A. (2005). SmartPLS (Version 2.0.M3). Hamburg, Germany. Retrieved May 1, 2011, from http://www.smartpls.de 237 Rivis, A., Sheeran, P., & Armitage, C. J. (2009). Expanding the affective and normative components of the theory of planned behavior: A meta-analysis of anticipated affect and moral norms. Journal of Applied Social Psychology, 39(12), 2985-3019. Roberts, T. (2010). MSNBC: Thomas Roberts talks with former classmate of Clint McCance.  Retrieved April 19, 2013, from http://www.youtube.com/watch?v=V6zyq8UvnYQ Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg self-esteem scale. Personality and Social Psychology Bulletin, 27(2), 151-161. Rossiter, J. (2002). The C-OAR-SE procedure for scale development in marketing. International Journal of Research in Marketing, 19, 305-335. Rotter, J. B. (1966). Generalized expectancies of internal versus external control of reinforcements. Psychological Monographs, 80, 1-28. Ruedy, N. E., Moore, C., Gino, F., & Schweitzer, M. E. (2013). The cheater’s high: The unexpected affective benefits of unethical behavior. Journal of Personality and Social Psychology, 105(4), 531–548. Runciman, W. G. (1966). Relative deprivation and social justice. London, UK: Routledge & Kegan Paul. Rusbult, C. E. (1980). Commitment and satisfaction in romantic associations: A test of the investment model. Journal of Experimental Social Psychology, 16, 172-186. Russell, J. A., Weiss, A., & Mendelsohn, G. A. (1989). Affect Grid: A single-item scale of pleasure and arousal. Journal of Personality and Social Psychology, 57(3), 493-502. 238 Ryan, C., & Gamson, W. A. (2006). The art of reframing political debate. Context, 5(1), 13-18. Ryan, W. (1971). Blaming the victim. New York: Pantheon. Sandelowski, M., Lambe, C., & Barroso, J. (2004). Stigma in HIV-positive women. Journal of Nursing Scholarship, 36(2), 122-128. Saunders, C., Rutkowski, A. F., van Genuchten, M., Vogel, D. R., & Orrego, J. M. (2011). Virtual space and place: Theory and test. M I S Quarterly, 35(4), 1079-1098. Scheff, T. J. (1988). Shame and conformity: The deference-emotion system. American Sociological Review, 53(3), 395-406. Schneider, B., Salvaggio, A. N., & Subirats, M. (2002). Climate strength: a new direction for climate research. Journal of Applied Psychology, 87(2), 220-229. Schunk, D. H. (1984). The self-efficacy perspective on achievement behavior. Educational Psychologist, 19, 199-218. Scott, J. (1990). Domination and the arts of resistance: Hidden transcripts. New Haven, CT: Yale University Press. Seidel, M. D. L., & Stewart, K. (2011). An initial description of the C-form. Research in the Sociology of Organizations, 33, 37-72. Shah, D. V., Cho, J., Eveland, W. P., & Kwak, N. (2005). Information and expression in a digital age - Modeling Internet effects on civic participation. Communication Research, 32(5), 531-565. 239 Shah, D. V., Cho, J., Nah, S., Gotlieb, M. R., Hwang, H., Lee, N. J., . . . McLeod, D. M. (2007). Campaign ads, online messaging, and participation: Extending the communication mediation model. Journal of Communication, 57(4), 676-U113. Shanock, L. R., Baran, B. E., Gentry, W. A., Pattison, S. C., & Heggestad, E. D. (2010). Polynomial Regression with Response Surface Analysis: A Powerful Approach for Examining Moderation and Overcoming Limitations of Difference Scores. Journal of Business and Psychology, 25(4), 543-554. (Erratum published in Journal of Business and Psychology, vol. 29,  p. 161, 2014). Shirky, C. (2008). Here comes everybody. New York: Penguin. Shover, N. (1996). Great pretenders: Pursuits and careers of persistent thieves. Boulder, CO: Westview Press. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420–428. Sidanius, J. (1993). The psychology of group conflict and the dynamics of oppression: A social dominance perspective. In S. Iyengar & W. J. McGuire (Eds.), Explorations in political psychology (pp. 183–219). Durham, NC: Duke University Press. Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and oppression. New York: Cambridge University Press. Simon, C. (2010). Clint McCance resigns: The back story.  Retrieved April 19, 2013, from http://www.hrc.org/blog/entry/clint-mccance-resigns-the-back-story 240 Sissons, M. E. (2005, July 4). Egypt: Margins of repression- State limits on nongovernmental organization activism.  Retrieved February 27, 2014, from http://www.hrw.org/reports/2005/07/03/egypt-margins-repression Skarlicki, D. P., & Kulik, C. T. (2005). Third-party reactions to employee (mis)treatment: A justice perspective. In B. M. Staw & R. M. Kramer (Eds.), Research in organizational behavior: An annual series of analytical essays and critical reviews (Vol. 26, pp. 183-229). New York: JAI-Elsevier. Skarlicki, D. P., & Turner, R. A. (2014). Unfairness begets unfairness: Victim derogation bias in employee ratings. Organizational Behavior and Human Decision Processes, 124(1), 34-46. Smith, A. (1976 [1776]). An inquiry into the nature and causes of the wealth of nations (Vol. 1 and 2). Oxford, UK: Oxford University Press. Smith, A. C. T. (2010, July 7). Mobile Access 2010. Pew Internet & Life Project. Retrieved March 11, 2013, from http://pewinternet.org/Reports/2010/Mobile-Access-2010.aspx?r=1 Smith, A. C. T. (2012, June 26). Cell Internet Use 2012. Pew Internet & Life Project. Retrieved February 27, 2014, from http://www.pewinternet.org/files/old-media/Files/Reports/2012/PIP_Cell_Phone_Internet_Access.pdf Smith, P. B., Trompenaars, F., & Dugan, S. (1995). The Rotter locus of control scale in 43 countries: A test of cultural relativity. International Journal of Psychology, 30(3), 377-400. Snijders, T. A. B., & Bosker, R. J. (2012). Multilevel analysis: An introduction to basic and advanced multilevel modeling (2nd ed.). London, UK: Sage. 241 Snow, D. A., Zurcher, L. A., & Sheldon, E.-O. (1980). Social networks and social movements: A microstructural approach to differential recruitment. American Sociological Review, 45(5), 787-801. Snyder, M., & Omoto, A. M. (2007). Social action. In A. W. Kruglanski & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (pp. 940- 961). New York: Guilford. Soprano, A., Crielaard, B., & Piacenza, F. (2010). Measuring operational and reputational risk: A practitioner's approach. Hoboken, NJ: Wiley. Spasser, A. J. (2013). Winning hearts and minds: Using "ag-gag" outrage and corporate rebranding to achieve a public image makeover for the animal rights movement. (Master of Public Affairs), University of San Francisco (Paper 4), San Francisco.  Retrieved February 27, 2014, from http://repository.usfca.edu/capstone Spencer, J. N., Bodner, G. M., & Rickard, L. H. (2010). Chemistry: structure and dynamics. Hoboken, NJ: John Wiley & Sons. Sprouse, J. (2011). A validation of Amazon Mechanical Turk for the collection of acceptability judgments in linguistic theory. Behavior Research Methods, 43(1), 155-167. Stajkovic, A. D., & Luthans, F. (1998). Self-efficacy and work-related performance: A meta-analysis. Psychological Bulletin, 124(2), 240-261. Stalder, F. (2006). Manuel Castells: The theory of the network society. Cambridge, UK: Polity Press. Stark, E. (2007). How men entrap women in personal life. Oxford, UK: Oxford University Press. StataCorp. (2013). Working with categorical data and factor variables Stata user's guide (Vol. 13, pp. 351-367). College Station, TX: Stata Press. 242 Statistics Canada. (2004, June 15). Canadian community health survey. The Daily. Retrieved September 12, 2013, from http://statcan.gc.ca/daily-quotidien/040615/dq040615b-eng.htm Statistics Canada. (2010, May 10). Internet use by individuals, by selected characteristics.  Retrieved 31 March, 2013, from http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/comm35a-eng.htm Statistics Canada. (2013, March 12). Aboriginal peoples. 2006 Census release topics. Retrieved September 12, 2013, from http://www12.statcan.ca/census-recensement/2006/rt-td/ap-pa-eng.cfm Stewart, M. (2010, October 29). Dr. Phil slams Arkansas official for 'non-apology apology'. ac360 blogs. Retrieved April 19, 2013, from http://ac360.blogs.cnn.com/2010/10/29/dr-phil-slams-arkansas-official-for-non-apology-apology/ Stiles, K. (2002). International support for NGOs in Bangladesh: Some unintended consequences. World Development, 30(5), 835-846. Stoessel, F. (2008). Thermal safety of chemical processes: Risk assessment and process design. Hoboken, NJ: John Wiley & Sons. Stotland, E. (1969). Exploratory ingestigations of empathy. In L. Berkowitz (Ed.), Advances in Hxperimental Vocial Ssychology (Vol. 3). New York: Academic Press. Straub, D. W., Boudreau, M.-C., & Gefen, D. (2004). Validation guidelines for IS positivist research. Communications of the AIS, 13, 380-427. 243 SuchIsLifeVideos. (2010). MSNBC: Clint McCance (Midland, Arkansas school board member) hails gay suicides.  Retrieved April 19, 2013, from http://www.youtube.com/all_comments?v=oE76ypLgNeE Svinicki, M., & McKeachie, W. J. (2011). McKeachie's teaching tips: Strategies, research, and theory for college and university teachers (13th ed.). Belmont, CA: Wadsworth Cengage. Tajfel, H., & Turner, J. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (Vol. 2). Chicago: Nelson-Hall. Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel & W. G. Austin (Eds.), Psychology of Lntergroup Uelations (pp. 7-24). Chicago: Nelson-Hall. Takei, G. (2010). George Takei calls out anti-gay Arkansas school board member.  Retrieved February 10, 2013, from http://www.youtube.com/watch?v=UACK93xF-FE Tam, D. (2013, January 30). Facebook by the numbers: 1.06 billion monthly active users. CNET News. Retrieved May 10, 2013, from http://news.cnet.com/8301-1023_3-57566550-93/facebook-by-the-numbers-1.06-billion-monthly-active-users/ Tangney, J. P. (1990). Assessing individual differences in proneness to shame and guilt: Development of the self-conscious affect and attribution inventory. Journal of Personality and Social Psychology, 59(1), 102-111. Tangney, J. P., Rowland, S. M., Flicker, L., & Barlow, D. H. (1996). Are shame, guilt, and embarrassment distinct emotions? Journal of Personality and Social Psychology, 70(6), 1256-1269. 244 Tangney, J. P., Stuewig, J., & Mashek, D. J. (2007). Moral emotions and moral behavior. Annual Review of Psychology, 58, 345-372. Tarrow, S. G. (1998). Power in movement: Social movements and contentious politics. Cambridge, UK: Cambridge University Press. Taylor, I., Walton, P., & Young, J. (1973). The new criminology. New York: Harper and Row. Thomas, R. J. (1994). What machines can't do: Politics and technoloy in the industrial enterprise. Berkeley, CA: University of California Press. Tiedens, L. Z., Ellsworth, P. C., & Mesquita, B. (2000). Stereotypes about sentiments and status: Emotional expectations for high- and low-status group members. Personality and Social Psychology Bulletin, 26(5), 560-574. Tilly, C. (1995). Popular contention in Great Britain, 1758–1834. Cambridge, MA: Harvard University Press. Tomassoni, T. (2012, March 16). Petitions are going viral, sometimes to great success. NPR News. Retrieved February 19, 2013, from http://www.npr.org/2012/03/16/148556371/petitions-are-going-viral-sometimes-to-great-success Tor. (2013). obfsproxy.  Retrieved March 2, 2013, from www.torproject.org/projects/obfsproxy.html.en Trochim, W. M. K., & Donnelly, J. P. (2008). Research methods knowledge base (3 ed.). Mason, OH: Cengage Learning. Turner, R. A., & Schabram, K. F. (2012). The bases of power revisited: An interpersonal perceptions perspective. Journal of Organizational Psychology, 12(1), 9-18. 245 Turner, V. W. (1986). Dewey, Dilthey, and drama: An essay in the anthropology of experience. In V. W. Turner & E. M. Bruner (Eds.), The anthropology of experience (pp. 33-44). Champaign, IL: University of Illinois Press. U. S. Census Bureau. (2013). Educational attainment in the United States: 2013 - Detailed Tables. Retrieved March 1, 2014, from www.census.gov/hhes/socdemo/education/data/cps/2013/tables.html UCLA Institute for Digital Research and Education. (2013a). Stata FAQ: How can I perform mediation with multilevel data? (Method 1).  Retrieved March 1, 2014, from http://www.ats.ucla.edu/stat/stata/faq/ml_mediation.htm UCLA Institute for Digital Research and Education. (2013b). Using Stata margins command to understand multiple interactions.  Retrieved November 10, 2013, from http://www.ats.ucla.edu/stat/stata/faq/margins_interaction.htm UCLA Institute for Digital Research and Education. (2014). Stata FAQ: How can I do power and robustness analyses for factorial anova? (Stata 11).  Retrieved January 18, 2014, from http://www.ats.ucla.edu/stat/stata/faq/factorialsim.htm Ulaga, W., & Eggert, A. (2006). Value-based differentiation in business relationships: Gaining and sustaining key supplier status. Journal of Marketing, 70(1), 119-136. van Dijk, J. A. G. M. (2005). The deepening divide: Inequality in the information society. Thousand Oaks, CA: Sage. van Dijk, J. A. G. M. (2012). The evolution of the digital divide - The digital divide turns to inequality of skills and usage. In J. Bus, M. Crompton, M. Hildebrandt & G. Metakides (Eds.), Digital enlightenment yearbook 2012. Fairfax, VA: IOS Press. 246 van Laer, J., & van Aelst, P. (2010). Internet and social movement action repertoires. Information, Communication & Society, 13(8), 1146-1171. Van Stekelenburg, J., & Klandermans, B. (2007). Individuals in movements: A social psychology of contention. In C. M. Roggeband & B. Klandermans (Eds.), The handbook of social movements across disciplines (pp. 157–204). New York: Springer. Van Stekelenburg, J., & Klandermans, B. (2013). The social psychology of protest. Current Sociology, 61(5-6), 886-905. Van Zomeren, M. (2013). Four core social‐psychological motivations to undertake collective action. Social and Personality Psychology Compass, 7(6), 378-388. Van Zomeren, M., Postmes, T., & Spears, R. (2008). Toward an integrative social identity model of collective action: A quantitative research synthesis of three socio-psychological perspectives. Psychological Bulletin, 134(4), 504-535. Vasi, I. B., & King, B. G. (2012). Social movements, risk perceptions, and economic outcomes: The effect of primary and secondary stakeholder activism on firms' perceived environmental risk and financial performance. American Sociological Review, 77(4), 573-596. Veenstra, A. S., Iyer, N., Hossain, M. D., & Park, J. (2014). Time, place, technology: Twitter as an information source in the Wisconsin labor protests. Computers in Human Behavior, 31, 65-72. Vickers, J. (2005). Abuse of market power. The Economic Journal, 115(504), F244-F261. Vogt, P. J. (2013, December 16). White House e-petitions work more often than you'd think. On The Media. Retrieved December 26, 2013, from 247 http://www.onthemedia.org/story/white-house-e-petitions-actually-work-sometimes/ Vorauer, J. D., Main, K. J., & O'Connell, G. B. (1998). How do individuals expect to be viewed by members of lower status groups? Content and implications of meta-stereotypes. Journal of Personality and Social Psychology, 75 (4), 917-937. Vroom, V. H. (1964). Work and motivation . New York: Wiley. Wade, J. B., Swaminathan, A., & Saxon, M. S. (1998). Normative and resource flow consequences of local regulations in American brewing industry, 1845-1918. Administrative Science Quarterly, 43 , 905-935. Warman, M. (2012, January 21). 50 billion devices online by 2020. The Telegraph . Retrieved March 1, 2013, from http://www.telegraph.co.uk/technology/internet/9051590/50-billion-devices-online-by-2020.html Watson, J. (2010, November 5). Exclusive interview: Clint McCance whistleblower Anthony Turner. Velvet Park Media . Retrieved April 19, 2013, from http://www.velvetparkmedia.com/blogs/exclusive-interview-clint-mccance-whistleblower-anthony-turner Weber, J., & Gillespie, J. (1998). Differences in ethical beliefs, intentions, and behaviors: The role of beliefs and intentions in ethics research revisited. Business and Society, 37 (4), 447-467. Weber, M. (1947). The theory of social and economic organization . Oxford: Oxford University Press. 248 Weiwei, A. (2012, April 16). China's censorship can never defeat the Internet. The Guardian. Retrieved March 12, 2013, from http://www.guardian.co.uk/commentisfree/libertycentral/2012/apr/16/china-censorship-internet-freedom West, A. (2013, July 3). 14 disturbing things Snowden has taught us (so far). Global Post. Retrieved August 7, 2013, from http://www.globalpost.com/dispatch/news/politics/130703/edward-snowden-leaks West, E. (1995). Organization building in the wake of ethnic conflict: A comparison of three groups. Social Forces, 4, 1333-1363. White, D. S., Gunasekaran, A., Shea, T. P., & Ariguzo, G. C. (2011). Mapping the global digital divide. International Journal of Business Information Systems, 7(2), 207-219. White, K., & Peloza, J. (2009). Self-venefit versus other-benefit marketing appeals: Their effectiveness in generating charitable support. Journal of Marketing, 73(4), 109-124. Whitley, B., & Lee, S. (2000). The relationship of authoritarianism and related constructs to attitudes toward homosexuality. Journal of Applied Social Psychology, 30, 144-170. Whitson, J. A., Liljenquist, K. A., Galinsky, A. D., Magee, J. C., Gruenfeld, D. H., & Cadena, B. (2013). The blind leading: Power reduces awareness of constraints. Journal of Experimental Social Psychology, 49(3), 579-582. Wike, R. (2014, February 13). Emerging nations embrace Internet, mobile technology: Cell phones nearly ubiquitous in many countries. PewResearch Global Attitudes Project. Retrieved February 25, 2014, from 249 http://www.pewglobal.org/2014/02/13/emerging-nations-embrace-internet-mobile-technology/ Wilkinson, I., & Kipnis, D. (1978). Interfirm use of power. Journal of Applied Psychology, 63(3), 315-320. Wiltfang, G. L., & McAdam, D. (1991). The costs and risks of social activism: A Study of sanctuary movement activism. Social Forces, 69(4), 987-1010. Winer, B. J. (1971). Statistical principles in experimental design. New York: McGraw-Hill. Wingrove, J. (2010, October 19). Calgary's Naheed Nenshi becomes Canada's first Muslim mayor. The Globe and Mail. Retrieved November 20, 2012, from http://www.theglobeandmail.com/news/politics/calgarys-naheed-nenshi-becomes-canadas-first-muslim-mayor/article1215182/?page=all Wray, R. (2009, December 22). Africa sees massive growth in mobile web usage. The Guardian. Retrieved March 11, 2013, from http://www.guardian.co.uk/technology/2009/dec/22/mobilephones-internet Wrenn, E. (2009, July 24). The sweet music of revenge: Singer pens YouTube hit after United Airlines breaks his guitars... and shares plunge 10%. The Daily Mail. Retrieved May 5, 2013, from http://www.dailymail.co.uk/news/article-1201671/Singer-Dave-Carroll-pens-YouTube-hit-United-Airlines-breaks-guitar--shares-plunge-10.html Wright, B. R. E., Caspi, A., Moffitt, T. E., & Paternoster, R. (2004). Does the perceived risk of punishment deter criminally prone individuals? Rational choice, self-control, and crime. Journal of Research in Crime and Delinquency, 41(2), 180-213. 250 Wright, S. C., Taylor, D. M., & Moghaddam, F. M. (1990). Responding to membership in a disadvantaged group - From acceptance to collecitve protest. Journal of Personality and Social Psychology, 58(6), 994-1003. Xie, B. (2008). Civic engagement among older Chinese Internet users. Journal of Applied Gerontology, 27(4), 424-445. Yang, G. (2013). The power of the Internet in China: Citizen activism online. New York: Columbia University Press. Yao, G., Wu, C.-H., & Yang, C.-T. (2008). Examining the content validity of the WHOQOL-BREF from respondents' perspective by quantitative methods. Social Indicators Research, 85(3), 483-498. Zald, M. N., & McCarthy, J. D. (1979). The dynamics of social movements: Resource mobilization, social control, and tactics. Cambridge, MA: Winthrop. Zickuhr, K., & Madden, M. (2012a, June 6). Older Adults and Internet Use. PewResearch Internet Project. Retrieved February 14, 2014, from www.pewinternet.org/2012/06/06/main-report-15/ Zickuhr, K., & Madden, M. (2012b, June 6). Online activities. PewResearch Internet Project. Retrieved November 4, 2012, from www.pewinternet.org/Reports/2012/Older-adults-and-internet-use/Main-Report/Online-activities.aspx Zickuhr, K., & Smith, A. (2012, March 13). Digital differences. PewResearch Internet Project. Retrieved March 31, 2013, from www.pewinternet.org/Reports/2012/Digital-differences/Main-Report/Internet-adoption-over-time.aspx  251 Appendix A Instructions and Information Provided to Participants (Study 2) Base instructions Below, you will be presented with two somewhat similar scenarios in which the same person with the same beliefs behaves in the same ways but in two different settings. You will rate the likelihood of various things happening as a result of the person’s actions in each. Please read the scenarios carefully.   Condition 1 [Scenario A: General Social Context] Suppose that a person living in a country ruled by a harsh totalitarian regime expressed a desire to see current rulers set aside or overthrown and a representative democracy put in place. Suppose that the person expressed these things within a very public place where the people present are representative of that society as a whole. Think about outcomes the person might experience as a result. Then, please indicate the likelihood that the person would immediately or eventually experience each outcome shown as a result of his or her actions. [Scenario B: Potential Free Space] Suppose that the same person who lives in a country ruled by a harsh totalitarian regime expressed his or her desire to see current rulers set aside or overthrown just as in the first scenario you read but instead of doing so in a very public place, the person expressed these things at a secret meeting of people all known to have very negative views of the ruling regime. This meeting takes place in a remote location and it is virtually certain that no one outside the group knows about the meeting. Think about outcomes the person might experience as a result. Then, please indicate the likelihood that 252 the person would immediately or eventually experience each outcome shown as a result of his or her actions. Condition 2  [Scenario A: General Social Context] Suppose that a person said that there were some really unethical things going on at work and that the managers were corrupt and abusive, especially toward employees with disabilities. But, the person has no hard evidence to offer to substantiate those claims. Further suppose that the person said these things within a very public context representative of your society as a whole. Think about outcomes the person might experience as a result. Then, please indicate the likelihood that the person would immediately or eventually experience each outcome shown as a result of his or her actions. [Scenario B: Potential Free Space] Suppose that the same person said that there were some really unethical things going on at work and that the managers were corrupt and abusive, especially toward employees with disabilities, and had no hard evidence to offer to substantiate those claims, but instead of doing so in the manner described in the first scenario you read, the person anonymously posted this information on discussion boards hosted by the local newspaper, radio station, and television news stations. Think about outcomes the person might experience as a result. Then, please indicate the likelihood that the person would immediately or eventually experience each outcome shown as a result of his or her actions. Condition 3  [Scenario A: General Social Context] Suppose that a high school student works part-time as a vocal performer for a local religious organization. This person is anatomically male and has always gone by a traditionally male name and worn stereotypically male clothing. 253 Suppose this person now expresses that she considers herself to be female in gender and does so within a very public context representative of your society as a whole. Think about outcomes the person might experience as a result. Then, please indicate the likelihood that she would immediately or eventually experience each outcome shown as a result of her actions. [Scenario B: Potential Free Space] Suppose that the same high school student who works part-time as a vocal performer for a local religious organization, is anatomically male, and has always gone by a traditionally male name and worn stereotypically male clothing expresses that she considers herself to be female in gender but instead of doing so in a very public place, she expresses this during a meeting of the gay, lesbian, bisexual, and transgendered (LGBT) youth group at her local LGBT community resource center. Think about outcomes the person might experience as a result. Then, please indicate the likelihood that she would immediately or eventually experience each outcome shown as a result of her actions. Condition 4 [Scenario A: General Social Context] Suppose that a person recorded a television ad to run in primetime on a local station in his or her community. Within that ad, the person condemned the treatment that Native Peoples / members of First Nations tribes within the country had received throughout history and emphatically called for ½ of all government workers to be replaced by Native Peoples / members of First Nations tribes so as to atone for this mistreatment. The community in which the ad ran is representative of your society as a whole. Think about outcomes the person might experience as a result. Then, please 254 indicate the likelihood that she would immediately or eventually experience each outcome shown as a result of her actions. [Scenario B: Potential Free Space] Suppose that the same person condemned the treatment that Native Peoples / members of First Nations tribes within the country had received throughout history and emphatically called for ½ of all government workers to be replaced by Native Peoples / members of First Nations tribes so as to atone for this mistreatment but instead of doing so in the manner described in the first scenario you read, the person expressed these views in person to people attending the “Celebrating Native American Heritage Festival.” Think about outcomes the person might experience as a result. Then, please indicate the likelihood that she would immediately or eventually experience each outcome shown as a result of her actions. Condition 5 [Base info] Imagine that a person lives in a country ruled by a repressive and highly authoritarian military regime, and the only newspapers, radio stations, and television outlets in the country are government controlled. In one area of the country, citizens are protesting the high unemployment and government corruption that they say is ruining their lives. The person has captured stirring photos of the local protest, including a photo of a young man who set himself on fire, and desperately wants to spread awareness of what’s going on in hopes the protest will spread and ultimately result in a new, fairer government. With that in mind: [Scenario A: General Social Context] Suppose that in his or her effort to promote an uprising against the current government, the person prints out thousands of posters and 255 pamphlets, including the pictures and descriptions of the protests and travels from town to town throughout the country, hanging the posters and distributing the pamphlets in town squares, markets, and other public spaces and encouraging those in those spaces to help spread the word. Think about outcomes the person might experience as a result. Then, please indicate the likelihood that she would immediately or eventually experience each outcome shown as a result of his or her actions. [Scenario B: Potential Free Space] Suppose that this same person makes an to promote an uprising against the current government, but instead of doing so in the manner described in the first scenario you read, the person shares the images and stories through Facebook, Twitter, and/or other social media outlets and asks “friends” and “followers” there to help spread the word. Think about outcomes the person might experience as a result. Then, please indicate the likelihood that she would immediately or eventually experience each outcome shown as a result of his or her action.  256 Appendix B Results from Item Selection Analysis, by Risk Category Item Pool (Study 2). In the tables below, representativeness comprises the item-test correlation for each item’s safety score and the composite of safety scores for items within its pool. A safety score comprises the difference in the likelihood of the outcome comprising an item if resistance occurs in a posited free space versus that if resistance occurs in a more general, public social context.  Overlap comprises the correlation between each item’s safety score and composites of safety scores for items in each risk category pool of which it is not a part. Items with representativeness less that .70 or/and average overlap in excess of .70 were considered poor exemplars of their respective risk categories and eliminated. Those values are shown with strike-through.  The text for each item is reported in Table 2 (found in Section 1.8.4).   ‘Institutional’ Items  IN1 IN2 IN3 IN4 IN5 IN6 IN7 Representativeness 0.834 0.876 0.872 0.836 0.895 0.918 0.859 Overlap          Strong Ties 0.478 0.465 0.446 0.306 0.524 0.447 0.490   Collateral 0.691 0.694 0.716 0.642 0.811 0.747 0.697   Societal, Active 0.668 0.663 0.676 0.556 0.774 0.649 0.723   Societal, Passive 0.513 0.484 0.408 0.405 0.608 0.422 0.453   Professional 0.560 0.533 0.486 0.522 0.607 0.479 0.522 Average Overlap 0.582 0.568 0.546 0.486 0.665 0.549 0.577 Representativeness/Avg. Overlap 1.433 1.543 1.595 1.720 1.347 1.673 1.489   257  ‘Societal, Active’ Items  SA1 SA2 SA3 SA4 SA5 SA6 SA7 Representativeness 0.750 0.920 0.891 0.831 0.935 0.784 0.746 Overlap          Strong Ties 0.622 0.731 0.789 0.645 0.732 0.544 0.566   Collateral 0.636 0.731 0.712 0.768 0.722 0.663 0.699   Institutional 0.510 0.645 0.613 0.746 0.656 0.747 0.661   Societal, Passive 0.720 0.743 0.774 0.607 0.714 0.593 0.529   Professional 0.723 0.663 0.719 0.619 0.671 0.615 0.542 Average Overlap 0.642 0.702 0.721 0.677 0.699 0.632 0.599 Representativeness/Avg. Overlap 1.168   1.227 1.336 1.240 1.245    ‘Societal, Passive’ Items  SP1 SP2 SP3 SP4 SP5 SP6 Representativeness 0.753 0.561 0.692 0.801 0.761 0.708 Overlap         Strong Ties 0.774 0.470 0.500 0.586 0.714 0.530   Collateral 0.677 0.239 0.242 0.492 0.803 0.698   Institutional 0.536 0.177 0.091 0.342 0.607 0.691   Societal, Active 0.839 0.300 0.350 0.564 0.739 0.708   Professional 0.651 0.351 0.456 0.681 0.742 0.657 Average Overlap 0.696 0.307 0.328 0.533 0.721 0.657 Representativeness/Avg. Overlap 1.082   1.502  1.078        258  ‘Strong Ties’ Items  ST1 ST2 ST3 ST4 ST5 ST6 ST7 Representativeness 0.818 0.830 0.740 0.819 0.535 0.557 0.667 Overlap          Collateral 0.731 0.716 0.569 0.721 0.104 0.126 0.494   Institutional 0.618 0.577 0.442 0.516 0.077 0.071 0.285   Societal, Active 0.734 0.758 0.627 0.740 0.268 0.255 0.560   Societal, Passive 0.723 0.725 0.599 0.620 0.418 0.415 0.649   Professional 0.660 0.661 0.600 0.597 0.280 0.238 0.564 Average Overlap 0.693 0.687 0.567 0.639 0.229 0.221 0.510 Representativeness/ Avg. Overlap 1.180 1.208 1.305 1.282      ‘Professional’ Items  PR1 PR2 PR3 PR4 PR5 PR6 PR7 PR8 Representativeness 0.785 0.770 0.825 0.895 0.544 0.852 0.507 0.828 Overlap           Strong Ties 0.551 0.504 0.534 0.545 0.485 0.590 0.549 0.639   Collateral 0.706 0.548 0.532 0.710 0.375 0.702 0.373 0.677   Institutional 0.521 0.504 0.380 0.493 0.324 0.584 0.304 0.550   Societal, Active 0.627 0.583 0.589 0.647 0.390 0.711 0.433 0.711   Societal, Passive 0.533 0.595 0.603 0.712 0.595 0.642 0.578 0.712 Average Overlap 0.587 0.547 0.528 0.621 0.434 0.646 0.447 0.657 Representativeness/Avg. Overlap 1.336 1.409 1.565 1.441  1.319  1.259        259  ‘Collateral’ Items  CO1 CO2 CO3 CO4 CO5 CO6 CO7 Representativeness 0.873 0.788 0.797 0.866 0.711 0.720 0.716 Overlap          Strong Ties 0.632 0.516 0.515 0.708 0.372 0.391 0.682   Institutional  0.691 0.596 0.711 0.612 0.623 0.717 0.535   Societal, Active 0.757 0.693 0.608 0.718 0.517 0.542 0.754   Societal, Passive 0.649 0.591 0.497 0.678 0.349 0.374 0.766   Professional 0.659 0.573 0.515 0.741 0.493 0.542 0.701 Average Overlap 0.677 0.594 0.569 0.691 0.471 0.513 0.688 Representativeness/Avg. Overlap 1.289 1.327 1.401 1.253 1.510 1.403 1.042 260 Appendix C Customizable FSIx Questionnaire27  Instructions: On EACH of the 12 lines below you will mark TWO (2) responses, for a total of 24.   COLUMN 1: In COLUMN 1, please indicate the likelihood that you would immediately or eventually experience each outcome shown as a result of your actions, if you were to [RESISTANCE ACTIVITY]28 assuming that did this [IN/AT/ON EMBEDDING SOCIAL CONTEXT].29    COLUMN 2: In COLUMN 2, please indicate the likelihood that you would immediately or eventually experience each outcome shown if your actions took place INSTEAD [IN/AT/ON FOCAL CONTEXT].30  On each line below, circle 1 answer per column (2 answers per line) to indicate your answers. COLUMN 1: If your actions took place  [IN/AT/ON] [PROTOTYPIC SOCIAL CONTEXT] COLUMN 2: If your actions took place  [IN/AT/ON] [FOCAL CONTEXT] Very UNLIKELY Unlikely Somewhat Unlikely Somewhat Likely  Likely  Very LIKELY Very UNLIKELY Unlikely Somewhat Unlikely Somewhat Likely  Likely  Very LIKELY 1 Being threatened with physical harm by others within society 1 2 3 4 5 6 1 2 3 4 5 6 2 Being seen as unfit for leadership positions within society 1 2 3 4 5 6 1 2 3 4 5 6 3 Being viewed less favorably by your employer 1 2 3 4 5 6 1 2 3 4 5 6 4 Being made to feel unwelcome or unwanted by family members 1 2 3 4 5 6 1 2 3 4 5 6 5 People you care about being the targets of harmful words or actions because of their association with you 1 2 3 4 5 6 1 2 3 4 5 6 6 Having to pay fines or financial penalties to public authorities 1 2 3 4 5 6 1 2 3 4 5 6 7 Having conflict with your relationship partner—partner/ spouse, boyfriend/girlfriend  1 2 3 4 5 6 1 2 3 4 5 6 8 Being jailed or physically punished by public authorities 1 2 3 4 5 6 1 2 3 4 5 6 9 Being treated as if you are immoral/tainted 1 2 3 4 5 6 1 2 3 4 5 6 10 Family, friends, or other valued social relations being watched by the authorities because of their association with you 1 2 3 4 5 6 1 2 3 4 5 6 11 Losing the support of valued business relations  1 2 3 4 5 6 1 2 3 4 5 6 12 Being held against your will by those other than public authorities 1 2 3 4 5 6 1 2 3 4 5 6                                                   27 One should bear in mind the topical specificity of free space discussion above when customizing this instrument. 28 E.g., (as in study 3), “openly advocate for gay, lesbian, bisexual, transgendered, and queer/questioning—LGBTQ—persons to receive preference in hiring and university admissions decisions to make up for past and present discrimination against them” 29 E.g., “in a very public context that is representative of society in [British Columbia/Denmark/Tennessee] as a whole” 30 The potential free space: E.g., “at a Gay Pride Festival event in this city” 261 Appendix D Study 4 Stock Information Appendix D.1   Call for Action (Email from customerDVSUHVHQWHGWRSDUWLFLSDQWV)  Hello.  My name is Pat Paston, and I am a professional musician. My livelihood as a musician requires me to travel extensively, so I fly very often.  I’ve had few problems until ten months ago when my band, Pat & Pals, flew Global Airlines from Edmonton, Alberta, to Little Rock, Arkansas, USA, to start a concert tour.  At the boarding gate, an announcement was made that oversized carry-ons would not fit and would have to be checked for pickup at the baggage carrousel in Little Rock. I told the agent that my carry-on was an $18,000 violin, made in 1733 by master violin maker Antonius Stradiuarius, and that I wanted to keep it with me. In reply, he said that would not be possible but not to worry because he would affix it with bright “Fragile” and “High Priority” tags.  Begrudgingly, I surrendered it to his care.  Waiting to deplane in Little Rock, one of my band mates saw my violin and our other instruments being thrown around recklessly by Global baggage handlers. I immediately tried to tell the flight attendant who cut me off saying: “Don’t talk to me about that. Talk to the lead agent inside.”  My bandmate and I, along with another passenger who overheard us and had also witnessed the rough handling, tried to explain the situation to the lead agent, but she replied that she was an “acting” lead agent and disappeared into the crowd.  I spoke to a third Global employee at the gate, and she dismissed me saying, “but hun, that’s why we make you sign the waiver.”  I explained that I didn’t sign a waiver and that no waiver would excuse what was happening outside. She said to take it up with the luggage desk if there were damages.  When the luggage was finally delivered—more than 30 minutes later-I was upset to find no “Fragile” or “High Priority” tags on my violin case. I immediately took my violin out of both its hard outer case and an inner padded protector sleeve. I was devastated to find a huge crack in the soundboard. With a sad heart, I took my ruined, historic violin to the luggage desk.   The luggage desk agent acknowledged the damage and opened a claim (#GL394-13-2345) but advised me that any repairs would be 262 covered only up to $200.  I explained again the value of the violin and emphasized that (1) I had been forced to check it despite my opposition, (2) it did not have the promised tags placed on it, and (3) it had been thrown around like a sack of potatoes despite its obvious fragility (even without the tags--the case is violin shaped!).  “Sorry,” he said, “Global Airlines policy.”  He went on to explain that I should contact the Global luggage line if I wanted to discuss my claim further and provided me with a 1-800 number.  I called the 1-800 number when it opened the next morning and was told that I would need to submit a receipt for repairs before any appeal would be considered.  I explained that the only way to repair the instrument was to replace the entire top of the violin and that this was a 300 year old violin made by a master craftsman that we were talking about!  It cannot be ‘repaired’; it can only be replaced (at great cost!). The agent informed me that only appeals for repair costs, not replacement, would be considered.  I was BY NO MEANS satisfied with this but figured it was better than nothing, which seemed to be my only other option. I had the violin repaired and submitted the receipt for the costs, which totaled $3,000. I also submitted a notarized statement from the president of the World Violin Museum attesting to the $18,000 value of the violin without damages or repairs and estimating its current, repaired value at, ironically, $3,000.  Over a month later, I received a call telling me that my appeal had been approved and that a check for 250—TWO HUNDRED AND FIFTY!??!— dollars would be mailed within 6 weeks. The agent went on to explain that “Global Airlines policy” limits appeals to $250. When I asked why I hadn’t been told that before I had repairs (that I never wanted to have made to begin with!) made, her response was “Well, did you ever ask?”. I asked to speak to a supervisor. She said that would not be possible because no one else in the organization besides the Operations Manager had authority to override Luggage Line decisions or to exceed the $250 limit and asked if she should dispatch the $250 check to close the claim. I said…NO!  After an extensive search of the Global Airlines website, I finally found your email address (but not your phone number...). I have been battling Global for months trying to resolve this. A Stradiuarius violin with the entire top replaced is no longer a Stradiuarius violin. So, now my $18,000 violin no longer exists, plus I am out another $3,000 for repairs that I never wanted. 263 Despite Global being fully and completely responsible for this, your company wants to send me a check for $250 and close the claim. This is simply NOT ACCEPTABLE!  You can email me or give me a call at 123-456-7890 if you would like any further information.  Otherwise, I await your response.  Thank you,  Pat Paston   Appendix D.2   Global Baggage Claims PolicyDVSUHVHQWHGWRSDUWLFLSDQWV  X.11. Checked Baggage Damage Claims for Commercial Flights Although we at Global Airlines do our best to ensure that no damage occurs to checked baggage entrusted to us by our commercial customers, damages may sometimes occur.  If presented with evidence of damages by a customer, a Customer Service Representative at the destination airport shall complete Form 17A and notify the customer that the cost for repairs to the item(s) will be reimbursed up to a maximum of $200 per customer, per flight. If the item(s) cannot be repaired, the cost of a comparable replacement (i.e., of similar size, quality, value, etc.) will be reimbursed up to a maximum of $200 per customer, per flight.      X.11.a. Customer Appeals In some instances, customers may argue that $200 is insufficient to cover the cost of repair or replacement of damaged item(s). In these instances, Customer Service Representatives shall refer the customer to the Global Luggage Line (1-800-LUGGAGE). Luggage Line representatives may exercise discretion in authorizing reimbursements up to a maximum of $250 for repair of replacement of damaged item(s), per customer, per flight. Repair or replacement costs in excess of $250 per customer, per flight may have a strong and negative impact on the profitability of that flight and must be approved by the Operations Manager or a member of executive management (i.e., COO, CTO, CFO, CEO).    Appendix D.3 United Breaks Guitars    Music video available  without subtitles: http://youtu.be/5YGc4zOqozo      http://www.davecarrollmusic.com/ubg/song1/  with English subtitles: http://youtu.be/Cudz3Qofrdo  264 Lyrics by Dave Carroll31 As performed by Sons of Maxwell  “I flew United Airlines on my way to Nebraska. The plane departed, Halifax, connecting in Chicago's O'Hare. While on the ground, a passenger said from the seat behind me,  ‘My God, they're throwing guitars out there.’  The band and I exchanged a look, best described as terror, At the action on the tarmac, and knowing whose projectiles these would be. So before I left Chicago, I alerted three employees Who showed complete indifference towards me.  United...       (United...) You broke my Taylor Guitar. United...       (United...) Some big help you are.  You broke it, you should fix it. You're liable, just admit it. I should've flown with someone else Or gone by car. 'Cause United breaks guitars.  When we landed in Nebraska, I confirmed what I'd suspected: My Taylor'd been the victim of a vicious act of malice at O'Hare.  So began a year-long saga, of pass the buck, ‘Don't ask me,’  And ‘I'm sorry, sir, your claim can go nowhere.’ So to all the airline’s people, from New York to New Deli, Including kind Mrs. Irlweg, who says the final word from them is ‘no.’                                                   31 These lyrics are also available online. On April 5, 2014, I created an archive of one website on which they are posted.  It is perpetually accessible at the following url: http://web.archive.org/web/20140405203807/http://www.lyricsmode.com/lyrics/s/sons_of_maxwell/united_breaks_guitars.html 265  I heard all your excuses,  And I've chased your wild gooses And this attitude of yours, I say, must go.  United...       (United...) You broke my Taylor Guitar. United...       (United...) Some big help you are.  You broke it, you should fix it. You're liable, just admit it. I should've flown with someone else Or gone by car. 'Cause United breaks guitars.  Well, I won't say that I'll never fly with you again,  'Cause, maybe, to save the world, I probably would,  But that won't likely happen,  And if it did, I wouldn't bring my luggage 'Cause you'd just go and break it,  Into a thousand pieces, just like you broke my heart When United breaks guitars.  United...       (United...) You broke my Taylor Guitar. United...       (United...) Some big help you are!  You broke it, you should fix it. You're liable, just admit it. I should've flown with someone else Or gone by car. 'Cause United breaks guitars.  Yeah, United breaks guitars.  Yeah, United breaks guitars.”   266 Appendix E Sample Source for Each Vested Group Considered in Study 6 Scenarios.  Vested Group in Scenario Considered Cancer HIV Aboriginal LGBT Canadian Cancer Survivors 25 8 9 8 Canadians with HIV/AIDS  9 25 8 8 Aboriginal Canadians 9 8 25 8 LGBT Canadians 8 9 8 25 Canadians at Large 24 24 26 26 n 75 74 76 75     267 Appendix F Focal Hypotheses, by Study Study 2  Hypothesis 1: The apprehension that a potential resister feels about acting in a given social setting is a negative function of the extent to which he/she perceives that setting as free space for the focal issue. Hypothesis 2:  The likelihood that a potential resister will act in a given social setting is a positive function of the extent to which he/she perceives that setting as free space for the focal issue. Study 3 Hypothesis 3:  A social setting perceived as comprising greater free space in a broader institutional context that is generally less receptive to the focal resistance issue (as reflected in, e.g., laws, regulations, social norms, social attitudes, cultural conventions) will be perceived as comprising lesser free space in a broader institutional context that is generally more receptive to it. Hypothesis 4: Self-perceived power attenuates free space perceptions such that those high in self-perceived power will view free space embedded within a hostile institutional context as less free, as compared to those low in self-perceived power.  Study 4 Hypothesis 5: Organizational decision makers will perceive greater risk to the organization from exploiting power when there is potential for a negative Internet-mediated response to the exploitation than when there is not.  Hypothesis 6: Organizational decision makers will perceive greater risk to themselves from the organization exploiting power when there is potential for a negative Internet-mediated response to the exploitation than when there is not. Hypothesis 7: The potential for Internet-mediated action against organizations who exploit power decreases power exploitation. Hypothesis 8: Organizations’ power relative to targets moderates the negative relationship between Internet-mediated response risk and pursuit of organizational goals to targets’ detriment.  The relationship will be less negative when the organization’s power advantage is smaller (vs. larger).   268  Study 4, cont. Hypothesis 9: Personal sense of power moderates the influence of Internet-mediated response risk in agents’ decisions to exploit power in pursuit of organizational goals.  Among agents with higher (vs. lower) personal sense of power, its influence will be stronger. Study 5 Hypothesis 10: People’s willingness to engage in social advocacy is greater in an online action context than in an offline action context. Hypothesis 11: People’s willingness to engage in a social advocacy is positively related to their perceptions of the utility of that action in accomplishing a desired social goal.  Hypothesis 12: People’s willingness to engage in a social advocacy is positively related to the positivity of their anticipated affect from doing so. Hypothesis 13: People’s willingness to engage in a social advocacy is positively related to their expectations of social support, should they do so. Study 6 Hypothesis 14: People’s willingness to engage in a social advocacy is positively related to the strength of social ties perceived with the intended audience. Hypothesis 15: Peoples’ willingness to engage in social advocacy is positively related to the extent to which they perceive that the general attitude toward the focal issue within their society is favorable. Hypothesis 16: People’s willingness to engage in a social advocacy is negatively related to the societal social stigma of the group vested in the social issue. Hypothesis 17: People’s willingness to engage in a social advocacy is greater when one has a family member(s) among the group vested in the social issue. Hypothesis 18: People’s willingness to engage in a social advocacy is positively related to their proactivity. Hypothesis 19: People’s willingness to engage in a social advocacy is positively related to their guilt-proneness.  269 Study 6, cont. Hypothesis 20: People’s willingness to engage in a social advocacy is negatively related to direct costs of participation (i.e., money, time, and effort that must be expended in order to undertake that action). Hypothesis 21: The influence of factors affecting people’s willingness to engage in social advocacy will differ depending on whether the act is to occur in an online or offline context Hypothesis 22: The moderating influences of online vs. offline context proposed in Hypothesis 21 will differ depending on whether potential actors are vested in the focal issue (i.e., three-way interactions among Online Context, Vested, and each of the other predictors).            

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0102482/manifest

Comment

Related Items