UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Once is not enough : motivations driving initial and subsequent prosocial behaviour Kristofferson, Kirk Jason 2015

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

Item Metadata

Download

Media
24-ubc_2015_may_kristofferson_kirk.pdf [ 3.09MB ]
Metadata
JSON: 24-1.0102480.json
JSON-LD: 24-1.0102480-ld.json
RDF/XML (Pretty): 24-1.0102480-rdf.xml
RDF/JSON: 24-1.0102480-rdf.json
Turtle: 24-1.0102480-turtle.txt
N-Triples: 24-1.0102480-rdf-ntriples.txt
Original Record: 24-1.0102480-source.json
Full Text
24-1.0102480-fulltext.txt
Citation
24-1.0102480.ris

Full Text

ONCE IS NOT ENOUGH: MOTIVATIONS DRIVING INITIAL AND SUBSEQUENT PROSOCIAL BEHAVIOUR  by  Kirk Jason Kristofferson  B.COMM (Hons.), University of Manitoba, 2003  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)  April 2015  © Kirk Jason Kristofferson, 2015 ii  Abstract  Previous research has demonstrated that the cost or effort of an initial prosocial action is a key predictor of consumer responses to subsequent support requests. Specifically, consumers who perform costly (costless) prosocial behavior are more (less) likely to behave prosocially in the future. Investigating the prevalent issues of slacktivism and charitable support allocation, this dissertation extends this model by introducing additional factors that moderate previously documented findings. In essay 1, I show that social observability is a key moderator that predicts when and why token support for a social cause leads to more or less support for the cause. Importantly, I document the existence of slacktivism, uncover the motivations driving the behavior, and suggest strategies to mitigate its consequences. In essay 2, I investigate the impact that charitable support allocation, the proportion of consumer support passed directly on to cause recipients, has on consumers after consumers have provided effortful, or meaningful support for the organization. Specifically, I demonstrate that low charitable support allocation reduces consumer prosocial identity, leading to lower subsequent support not only for the originally supported cause, but also for other unrelated causes. Importantly, I reconcile this identity-based consequence from previously proposed theoretical models. Finally, I identify potential avenues for future research and discuss the theoretical and managerial implications of the work.    iii  Preface  I am the primary author of the work presented in this Ph.D. dissertation. I was responsible for conducting the literature review, developing the hypotheses, designing the experiments, collecting the data, analyzing the data, and preparing the manuscript. Additional contributions for each chapter are described below.   1 Introduction I am the primary author of this chapter. A portion of this chapter has been published. Kristofferson, Kirk, Katherine White and John Peloza, “The Nature of Slacktivism: How the Social Observability of an Initial Act of Token Support Affects Subsequent Prosocial Action.,” © 2014 by JOURNAL OF CONSUMER RESEARCH, Inc. I was the first author of this publication, with intellectual contributions from Katherine White and John Peloza.   2 Essay 1: The Nature of Slacktivism: How the Social Observability of an Initial Act of Token Support Affects Subsequent Prosocial Action A version of this chapter has been published. Kristofferson, Kirk, Katherine White and John Peloza, “The Nature of Slacktivism: How the Social Observability of an Initial Act of Token Support Affects Subsequent Prosocial Action.,” © 2014 by JOURNAL OF CONSUMER RESEARCH, Inc. I am the first author of this publication. I designed the experiments, supervised data collection, conducted the analyses and prepared the manuscript. Katherine White assisted in designing the experiments and provided intellectual contributions.  iv  3 Essay 2: Losing That Loving Feeling: Why Learning About Charitable Support Allocation Reduces Recurring Donation Behaviour I am the first author of this manuscript. I designed the experiments, supervised data collection, conducted the analyses and prepared the manuscript. Darren Dahl, Katherine White and Cait Lamberton assisted in designing the experiments and provided intellectual contributions.  4 Conclusion A portion of this chapter has been published. Kristofferson, Kirk, Katherine White and John Peloza, “The Nature of Slacktivism: How the Social Observability of an Initial Act of Token Support Affects Subsequent Prosocial Action.,” © 2014 by JOURNAL OF CONSUMER RESEARCH, Inc. I am the first author of this publication, with intellectual contributions from Katherine White, Darren Dahl and Cait Lamberton.   The research presented in this dissertation was supported by Standard Research Grants from the Social Sciences and Humanities Research Council to Dr. Darren Dahl and Dr. Katherine White.   Ethical approval for all experimental studies was obtained from UBC Office of Research Services Behavioural Review Board (Human Ethics) under the following certificates: Essay 1: H12-00566, Essay 2: H13-03221. Study participants were debriefed upon completion.     v  Table of Contents  Abstract ................................................................................................................................... ii	  Preface .................................................................................................................................... iii	  Table of Contents ..................................................................................................................... v	  List of Tables ........................................................................................................................... ix	  List of Figures .......................................................................................................................... x	  Acknowledgements ................................................................................................................. xi	  Dedication .............................................................................................................................. xiv	  Chapter 1: Introduction .......................................................................................................... 1	  1.1	   Overview of the Dissertation ................................................................................. 2	  Chapter 2: The Nature of Slacktivism: How the Social Observability of an Initial Act of Token Support Affects Subsequent Prosocial Action .......................................................... 5	  2.1	   Introduction ........................................................................................................... 5	  2.2	   Conceptual Background and Hypotheses Development ....................................... 9	  2.2.1	   The Consequences of Engaging in Token Support for a Cause ................ 9	  2.2.2	   The Moderating Role of Social Observability: The Public Versus Private Nature of Support ........................................................................................................ 12	  2.3	   Empirical Investigation ........................................................................................ 16	  2.3.1	   Study 1 ..................................................................................................... 16	  2.3.1.1	   Procedure ................................................................................................. 17	  2.3.1.2	   Results ..................................................................................................... 18	  2.3.1.3	   Discussion ................................................................................................ 19	  vi  2.3.2	   Study 2 ..................................................................................................... 20	  2.3.2.1	   Main Study Procedure ............................................................................. 21	  2.3.2.2	   Results ..................................................................................................... 23	  2.3.2.3	   Discussion ................................................................................................ 24	  2.3.3	   Study 3 ..................................................................................................... 25	  2.3.3.1	   Procedure ................................................................................................. 27	  2.3.3.2	   Results ..................................................................................................... 29	  2.3.3.3	   Discussion ................................................................................................ 31	  2.3.4	   Study 4 ..................................................................................................... 35	  2.3.4.1	   Procedure ................................................................................................. 37	  2.3.4.2	   Results ..................................................................................................... 39	  2.3.4.3	   Discussion ................................................................................................ 40	  2.3.5	   Study 5 ..................................................................................................... 42	  2.3.5.1	   Procedure ................................................................................................. 44	  2.3.5.2	   Results ..................................................................................................... 45	  2.3.5.3	   Discussion ................................................................................................ 46	  2.3.6	   General Discussion .................................................................................. 47	  2.3.6.1	   Theoretical Contributions ........................................................................ 48	  2.3.6.2	   Managerial Implications .......................................................................... 50	  2.3.6.3	   Directions for Future Research ................................................................ 53	  2.3.6.4	   Essay 1 Conclusion .................................................................................. 55	  Chapter 3: Losing That Loving Feeling: Why Learning About Charitable Support Allocation Reduces Recurring Donation Behaviour .......................................................... 56	  vii  3.1	   Introduction ......................................................................................................... 56	  3.2	   Conceptual Background and Hypotheses Development ..................................... 61	  3.2.1	   Why Supporters Bear the Brunt of Low Support Allocation .................. 64	  3.3	   Empirical Investigation ........................................................................................ 67	  3.3.1	   Pilot Study ............................................................................................... 67	  3.3.1.1	   Procedure ................................................................................................. 68	  3.3.1.2	   Results ..................................................................................................... 70	  3.3.1.3	   Discussion ................................................................................................ 71	  3.3.2	   Study 1: Internalizing Low Allocation .................................................... 72	  3.3.2.1	   Main Study Procedure ............................................................................. 73	  3.3.2.2	   Results ..................................................................................................... 77	  3.3.2.3	   Discussion ................................................................................................ 78	  3.3.3	   Study 2: Consequences of Reduced Prosocial Identity ........................... 80	  3.3.3.1	   Procedure ................................................................................................. 82	  3.3.3.2	   Results ..................................................................................................... 84	  3.3.3.3	   Discussion ................................................................................................ 88	  3.3.4	   Study 3: When Intentions Matter: The Ice Bucket Challenge ................. 89	  3.3.4.1	   Procedure ................................................................................................. 91	  3.3.4.2	   Results ..................................................................................................... 94	  3.3.4.3	   Discussion ................................................................................................ 99	  3.4	   General Discussion ............................................................................................ 100	  3.4.1	   Theoretical Contributions ...................................................................... 101	  3.4.2	   Managerial Implications ........................................................................ 103	  viii  3.4.3	   Directions for Future Research .............................................................. 105	  3.4.4	   Essay 2 Conclusion ................................................................................ 107	  Chapter 4: Limitations ........................................................................................................ 109	  Chapter 5: Concluding Remarks ....................................................................................... 112	  Chapter 6: Tables and Figures ........................................................................................... 113	  6.1	   Tables ................................................................................................................. 113	  6.2	   Figures ............................................................................................................... 117	  References ............................................................................................................................. 125	  Appendix: Experimental Materials ................................................................................... 141	  A.1	   Essay 1, Study 2 Instructions ............................................................................. 141	  A.2	   Essay 1, Study 2 Pre-Test .................................................................................. 142	  A.3	   Essay 1, Study 2 Sliding Scale Dependent Variable ......................................... 144	  A.4	   Essay 1, Study 3: Impression-Management Scale ............................................. 145	  A.5	   Essay 1, Study 3: Preference for Consistency Sub-Scale .................................. 146	  A.6	   Essay 1, Study 4: Charitable Organization Facebook Screen-Shots ................. 147	  A.7	   Essay 2, Study 4: Value Alignment Manipulation ............................................ 148	  A.8	   Essay 2, Study 4 Follow-Up Study: Online Petition Screen-Shot ..................... 149	  A.9	   Essay 2, Pilot Study Scenario ............................................................................ 150	  A.10	   Essay 2 Allocation of Support Stimuli .............................................................. 151	  A.11	   Moral Identity Scale (Aquino and Reed 2002) .................................................. 153	  A.12	   Essay 2, Study 3 Verbal Instructions: ................................................................ 154	  A.13	   Essay 2, Study 3 Charity Descriptions .............................................................. 155	   ix  List of Tables  Table 1. Summary of Results (Essay 1: Study 5) .................................................................. 113 Table 2. Correlation Matrix (Essay 1: Study 3) ..................................................................... 114	  Table 3 a. Summary of Results – Self-Only Ratings (Essay 2: Study 3) .............................. 115	  Table 3 b. Summary of Results – Self and Observer Ratings (Essay 2: Study 3) ................. 116	   x  List of Figures  Figure 1. Theoretical Framework of Dissertation ................................................................. 117	  Figure 2. Average Amount Donated in Response to a Second Request for Meaningful Support (Essay 1, Study 1) .................................................................................................... 118	  Figure 3. Number of Minutes Volunteered for Subsequent Request (To Stuff Envelopes) to Support Chosen Charity (Essay 1, Study 2) .......................................................................... 119	  Figure 4. Percentage of Participants Agreeing to Subsequent Volunteer Support Request (Essay 1, Study 3) .................................................................................................................. 120	  Figure 5. Percentage of Participants Agreeing to Subsequent Support Request (To Stuff Envelopes for Mail Campaign) as a Function of Value Alignment (Essay 1, Study 4) ........ 121	  Figure 6. Prosocial Identity as a Function of Allocation of Support and Prosocial Prime (Essay 2, Study 1) .................................................................................................................. 122	  Figure 7. Prosocial Identity as a Function of Allocation of Support and Support Success (Essay 2, Study 3) .................................................................................................................. 123	  Figure 8. Subsequent Support as a Function of Allocation of Support and Support Success (Essay 2, Study 3) .................................................................................................................. 124	   xi  Acknowledgements  I do not believe words can truly express the gratitude I feel for the many individuals who have helped me through this journey. However, given that some might say I have a love for words, and communication in general, I will endeavor to try. First, I am so grateful to my advisors Darren Dahl and Kate White, and committee member Cait Lamberton for everything they have done for me both professionally and personally during my PhD program. Darren, you are the reason I am here today, and for reasons other than the obvious incredible guidance, mentorship and support. Meeting, working with, and becoming friends with you many years ago exposed me to the world of academic research and lit a spark I am happy to say has turned into a passion. Your support and belief in me throughout these years has helped me more than you probably realize. Thank you for always making time for me, for pushing me to be my best and for walking the delicate balance between advisor and friend. You always understood how to get the best out of me, but more importantly you taught and showed me what it takes to succeed. I have always looked up to you and am so grateful for everything you have done for me. I can only hope to be as wonderful an advisor to my students as you have been to me.   Kate, I do not think I can say thank you enough for everything you have done for me. From almost the minute you came to UBC, I knew I wanted to work with you. Not only has working with you exposed me to a topic of research I love, but also taught me the style I love to conduct it in. Thank you for putting so much time, attention and energy into me, and for always being available to me professionally and personally. From reading (long) draft after xii  draft to watching (the same) talk after talk, I would not be here today were it not for your support and guidance. You knew when to let me solve things on my own, and when step in and help. You struck the perfect balance I needed to learn and grow. Thank you also for becoming such a good friend, and for going through the ups and downs of my first publication experience. You are absolutely awesome and I am forever grateful!  To the fabulous Cait Lamberton, thank you for everything you have done for me over the last five years and for being such an important part of my PhD experience! Thank you for taking the chance on me before I ever set foot in the PhD door, for devoting so much time to teach, guide and meet with me even from the other side of the continent and having students of your own, and for being such a good friend through all of the ups and downs this world brings. You always believed in me and for that I am so grateful! You are truly one of my favourite people and I am so happy to have you on this journey with me.  Thank you to everyone in the marketing division for all of their time, guidance, encouragement and support during my program. I am so lucky to have learned from such incredible minds and become friends and colleagues with such wonderful people. You all believed in and supported me during the challenging times, and “course-corrected” me when I needed it most. I am proud to be a part of the Sauder community. Special thank you to Joey Hoegg for all of her support and guidance both professionally and personally, especially during the first few years of my program.   xiii  Very special thanks to my fellow PhD students for all of their support and friendship during this journey. While we all come from different paths and have different goals, we are and always shall be a team, and I am honoured to have walked the path with you.  Finally, thank you to my wonderful family who have supported me during this adventure. To my incredible wife Melissa, were it not for you I would not be here. Thank you for taking a chance on me, and for walking beside me during this journey. Your unwavering love, support and belief in me are the reasons I am here today. Thank you for picking me up when I needed it most, and trusting in me in this uncertain world. You make me want to be better every day. I promised you would never be bored. To my parents, June and Ken Kristofferson, thank you for your unconditional love and belief in me. You supported me no matter what direction my life took me, trusted me to make the right decisions and believed I would succeed even when I did not. I can only aspire to be as wonderful a role model to my children as you both have been to me.   xiv  Dedication  This dedication is dedicated to my wife, Melissa, my son Kellan, and my parents, June and Ken Kristofferson. 1  Chapter 1: Introduction  Consumer researchers have long sought to understand the motivations that drive consumers to engage in recurring prosocial behavior for charitable or non-profit organizations. Although researchers have uncovered some conditions in which consumers will behave prosocially (versus selfishly), the reality is that charitable support is not a one-time event. Consumers do not volunteer or donate once and then cease support; in reality, ongoing support is vital to the success of non-profits. Research has shown that recurring donors donate significantly more money per year than one-time donors (Flannery and Harris 2009). My dissertation is focused on understanding the factors that drive consumers to provide on-going prosocial support to non-profit organizations.    Although multiple frameworks to understand prosocial behavior have been proposed, my dissertation adopts the view of economic models of prosocial behavior and contributes theoretically to that framework (Bénabou and Tirole 2004; Gneezy et al. 2011). These models suggest that the cost or effort of an initial prosocial action, and a supporter’s prosocial identity are key predictors of how supportive one will be in the future (Gneezy et al. 2011). Recent consumer research has shown that the cost or effort of an initial support behavior is a key moderator in predicting subsequent support. Gneezy et al. (2011) demonstrated that when consumers provide costly or effortful prosocial support, such as purchasing a product with a large donation component, they experience an increase in prosocial identity. This increase in prosocial identity, in turn, leads to higher subsequent support. Conversely, they find that costless, or low-effort support for a cause does not impact one’s prosocial identity, and leads to less 2  subsequent support. Although this is an excellent step in increasing understanding of how initial support impacts self-view and future behaviour, my dissertation proposes that conditions exist under which these findings do not hold. Investigating the prevalent issues of slacktivism and charitable support allocation, this dissertation extends this model of prosocial behaviour by introducing additional factors that moderate previously documented findings.   1.1 Overview of the Dissertation  This dissertation is comprised of nine experimental studies across two essays. Essay 1 investigates the prevalent consumer issue of slacktivism, defined as a willingness to perform a relatively costless, token display of support for a social cause, with an accompanying lack of willingness to devote significant effort to enact meaningful change. In five studies (2 field, 3 experimental), this essay seeks to demonstrate the existence of slacktivism, uncover the motivations driving subsequent meaningful prosocial support, reconcile competing predictions from established research streams by introducing a moderator not previously investigated and provide strategies to mitigate the behavior. In a field setting, study 1 demonstrates the existence of slacktivism and finds that providing private token support (versus public or no support) leads to more meaningful support. Study 2 replicates study 1 in a controlled environment using the signing of a petition as token support. Study 3 tests the full conceptual framework and provides mediational support for the proposed processes. Using participants’ personal Facebook accounts, study 4 manipulates the proposed process of value alignment and proposes a way to mitigate slacktivism. Finally, in a second field experiment, study 5 examines a boundary condition of the proposed effects: connection to the cause. 3   Essay 2 examines the complementary side of the prosocial support model – meaningful support – to understand when consumers are more or less likely to provide subsequent support for causes after providing meaningful, or effortful support (see figure 1 for overall theoretical framework). Presently, minimal research has examined this aspect of the economic model of prosocial behavior. Essay 2 addresses this gap by investigating a condition that affects the perceived impact of one’s meaningful support: charitable support allocation, defined as the proportion of consumer donations that a charity passes directly on to underlying cause recipients. In four studies (1 pilot, 3 experimental), this essay seeks to demonstrate that exposure to low support allocation information reduces consumer prosocial identity, and leads to lower subsequent support for both target and other unrelated causes. A pilot study shows that learning that one’s support is less impactful than originally expected reduces consumer prosocial identity and subsequent support for the organization. Using an experimental paradigm in which participants provide meaningful volunteer support to an organization, study 1 shows that upon learning the allocation of their support to cause recipients is low (versus high), consumers express a lower prosocial identity. Study 2 replicates the prosocial identity consequence from study 1 using a repeated measures design, and shows that this reduction in turn predicts lower subsequent support for both target, and unrelated social causes. Finally, by modifying the recent ice bucket challenge, study 3 examines a boundary condition to reverse the observed effects and reconcile the framework with previous research.    In a general sense, this work provides a more nuanced understanding of the factors that 4  motivate consumers to provide recurring prosocial support for charitable organizations. In doing so, it is able to address two prevalent social issues generating significant attention from consumers, mainstream media and consumer researchers. Moreover, it contributes two moderators to the economic model of prosocial behaviour theoretical framework that qualify previously held beliefs about how and why consumers choose to support. Given this updated understanding, this work not only has the potential to stimulate researchers to identify additional factors that impact prosocial support, but also to aid non-profit organizations in the strategic operations of their organizations to best fulfill their mandates.   The rest of the dissertation is structured as follows. I present essay 1 and essay 2 in turn. In each essay, I first position the relevant consumer issues to be addressed (Essay 1: Slacktivism; Essay 2: Charitable Support Allocation). Next, I provide the conceptual background and develop my hypotheses for the above-mentioned studies, with respect to the impact of providing initial support on subsequent support and the underlying processes. Then I report experimental studies that test the proposed hypotheses. Finally, I conclude each essay with a section that summarizes the theoretical and substantive contributions and highlight avenues for future research. I then discuss the general limitations of my dissertation and provide concluding remarks.  5  Chapter 2: The Nature of Slacktivism: How the Social Observability of an Initial Act of Token Support Affects Subsequent Prosocial Action  2.1 Introduction  "Look, if you make a Facebook page we will ‘like’ it—it’s the least we can do. But it’s also the most we can do."  ~Seth Meyers, Weekend Update, Saturday Night Live, September 22, 2012   Although intended as a satirical commentary, the above quote highlights what many assume is a prevalent behavior among consumers: slacktivism. I define slacktivism as a willingness to perform a relatively costless, token display of support for a social cause, with an accompanying lack of willingness to devote significant effort to enact meaningful change (Morozov 2009a; Davis 2011). A variety of factors, including the dramatic increase in social media presence among charitable organizations and advocacy groups, has made it increasingly easy for consumers to engage in small token acts of support for causes. However, one predominant criticism that has emerged is that this increased online presence has done little more than create a generation of “slacktivists” who will engage in token displays of support for a cause but are not likely to subsequently engage in more meaningful contributions to the cause (Morozov 2009b). This essay examines the conditions under which such “slacktivist” behavior occurs and proposes that the social observability of the initial act of token support can play an important moderating role. 6   Consumers have multiple avenues open to them to engage in small token forms of support for an issue or cause. This can include signing a petition, wearing a bracelet or pin in support of a cause, or engaging in various forms of online support such as liking or joining a page on Facebook. I refer to these types of behaviors as token support because they allow consumers to affiliate with a cause in ways that show their support to themselves or others, with little associated effort or cost. I contrast token support with meaningful support, which I define as consumer contributions that require a significant cost, effort, or behavior change in ways that make tangible contributions to the cause. Examples of meaningful support include donating money and volunteering time and skills.   Anecdotally, marketing practitioners and social media pundits offer conflicting views of how token consumer support impacts subsequent meaningful support. On the one hand, critics suggest that participating in token acts of support may not necessarily lead to a higher likelihood of engaging in more substantial support for the cause in the future (Morozov 2009a). On the other hand, some observers laud these token endorsements as a positive stepping stone towards more meaningful forms of social engagement in the future (Center for Social Impact Communication 2011; Fox 2012). Charitable organizations also appear divided on the issue. The vast number of token-support campaigns created to engage consumers suggests that the causes that employ them believe that they are successful. However, other organizations have advocated against such token displays of support in an attempt to curb slacktivist behavior. For example, UNICEF Sweden launched its “Likes Don’t Save Lives” campaign, which communicates to consumers that meaningful financial contributions, rather than mere token displays of support for 7  the cause, are required to protect children in developing nations against disease (UNICEF Sweden 2013).  This essay explores the conditions under which an initial token act of support for a social cause increases subsequent meaningful contributions. I make the novel prediction that when the initial act of token support is high in social observability (i.e., it is public), people will be less likely to engage in subsequent meaningful contributions to the cause than when the initial act of token support is low in social observability (i.e., it is private). My conceptual framework proposes that when the initial support situation is high in social observability, impression-management motives become activated (e.g., Leary and Kowalsky 1990). Consumers can satisfy these impression-management needs by engaging in a publicly observable token act of support for a positively viewed, prosocial cause. As a result of these impression-management motives already being satisfied, consumers will not be particularly motivated to contribute to the cause when a subsequent request for more meaningful support is made. Conversely, I propose that when token support is low in observability, consumers will be focused on the private (versus public) self. Under private conditions, the desire to maintain consistency with one’s own values and behaviors will be most relevant for consumers when deciding to provide subsequent help for the cause (Bem 1972; Festinger 1957). Thus, after engaging in a private (versus public) initial act of support for a cause, consumers are predicted to be more likely to act consistently with their previous behavior and therefore be more inclined to help in response to a subsequent request.  This essay makes several notable contributions to the literature. First, I provide an initial empirical investigation of the conditions under which slacktivism will emerge—a topic primarily debated by the mass media. I outline the motives underlying a person’s likelihood to help after 8  engaging in an initial token display of support, and identify social observability as a key factor that determines when slacktivism may occur. In particular, I show that when an initial act of token support is made in a setting that is high in social observability, consumers will be less likely to help in response to a more meaningful follow-up request than when the initial act of token support is made in a setting that is low in social observability (studies 1–3). In doing so, I build on a body of work suggesting that people do tend to follow through with a second request after committing to a smaller act of support (e.g., Freedman and Fraser 1966). I also add to past work that examined moderators of this basic effect—the costliness of the initial behavior (Bénabou and Tirole 2004; Gneezy et al. 2011), the salience or importance of the initial behavior, and the degree of similarity between the first and second requests (Burger 1999)—by showing that social observability of the initial token act of support is an important moderator of these effects.  Second, I integrate theoretical perspectives from both impression-management and self-consistency theories to outline a conceptual framework highlighting the mechanisms behind my effects. I demonstrate that when the context of token support is highly observable to others, impression-management motives become active. Under such public conditions, the show of token support satisfies these motives, reducing the desire to engage in subsequent support (study 3). Moreover, I show that providing token support in less observable contexts activates consistency motives (study 3) and increases perceived value alignment between self and cause (study 4 follow-up), subsequently leading to a greater willingness to provide meaningful support.  Third, I highlight the conditions under which consumers who make an initial display of public token support for a cause can be impelled to help on a subsequent task (study 4), thus mitigating 9  the effects of slacktivism. In particular, I find that when consumers who have made an initial display of public support are able to focus on the alignment of values between the self and the cause (versus those who do not), they are subsequently more likely to contribute to the cause in more meaningful ways. This finding is of theoretical importance, showing the conditions under which the motive to be consistent with one’s own values and behaviors can override impression-management concerns. This result also makes a substantive contribution for those tasked with building consumer support for social causes, showing how marketing practitioners can minimize slacktivist tendencies.   Finally, I identify a boundary condition for the observed effects—connection to the cause—and find that for individuals who have a high level of affective involvement with the cause, displaying public token support can lead to higher subsequent support than when the token support is made in private (study 5). Thus, I find that token-support programs can be effective, but primarily among those who are highly connected to the cause or organization. I next turn to a review of the literature and outline my conceptual framework.  2.2 Conceptual Background and Hypotheses Development  2.2.1 The Consequences of Engaging in Token Support for a Cause  Both marketing practice and empirical research generate competing predictions for how an initial act of token support for a cause impacts more meaningful subsequent forms of support. One perspective is supported by anecdotal evidence from media commentary reporting that 10  consumers commonly behave in a “slacktivist’ fashion by committing only to small, token acts of support, but not subsequently engaging in more meaningful forms of helping behaviors. This perspective is supported by empirical work on moral licensing (Khan and Dhar 2006). Moral licensing occurs when prior prosocial behavior gives people “license” to subsequently engage in less moral or helpful actions (Khan and Dhar 2006, 2007; Mazar and Zhong 2011; Monin and Miller 2001). This phenomenon has been demonstrated in political, prosocial and consumer settings (i.e. Effron, Cameron and Monin 2009; Khan and Dhar 2006, 2007). In the consumer setting, for example, Khan and Dhar (2006) found that after consumers imagined engaging in community-service activities, they were more likely to choose luxury over necessity products, and less likely to donate part or all of their participation payment to charity. Sachdeva, Iliev, and Medin (2009) asked participants to write a short story about either themselves or someone else using positive trait words (fair, kind) or negative trait words (selfish, mean). At the end of the study, participants were given an opportunity to donate some or all of their participation payment to charity. Participants who wrote about themselves using positive trait words donated significantly less than those who wrote about themselves using negative trait words and those who wrote about someone else. In a particularly unique study, Mazar and Zhong (2010) investigated moral licensing through the usage of environmentally-friendly products. Participants selected products from an online store that displayed primarily environmentally-friendly or conventional items, then played a game in which they could earn extra money based on self-reported performance. Participants that chose from the environmentally-friendly store were significantly more likely to lie about their performance in order to receive extra money than participants who chose from the conventional store. In sum, the literature on moral licensing would predict that engaging in an initial act of token support for a cause will lead to a decreased 11  propensity to make future meaningful contributions to the cause. Recent consumer research on prosocial behavior also supports this prediction. In an investigation of cause marketing programs, Krishna (2011) found that consumers donated less to a charity after purchasing a cause-marketing product (versus the same product without donation).     An alternative perspective is supported by work on self-consistency, which predicts that once individuals have engaged in a particular behavior, they will be more likely to engage in congruent behaviors in the future (Aronson 1968; Bem 1972; Festinger 1957; Heider 1958). This is because consumers are motivated to maintain consistency between their past and future actions. For example, research on the foot-in-the-door (FITD) effect demonstrates that individuals are significantly more likely to comply with a large request after they have first agreed to a smaller request. The FITD compliance paradigm has generated extensive interest since Freedman and Fraser’s initial publication in 1966. Since that time, over 100 publications, as well as multiple reviews and meta-analyses, have been written investigating the effect size, boundaries and processes underlying this effect (i.e. Burger 1999; DeJong 1979; Dillard, Hunter and Burgoon 1984). The basic notion is that individuals are significantly more likely to comply with, or agree to, a large request after they have first agreed to a smaller request. In a classic investigation, Freedman and Fraser (1966) contacted housewives with a small initial request to answer a few questions regarding household cleaners. Three days later, researchers contacted the same housewives again, as well as others in a control condition who had not received the earlier request. This time, a larger request of a thorough search of their homes to catalogue household products was made. Women who initially agreed to the smaller request were significantly more likely to agree to the larger request than women who had not initially received the small request.  12   The FITD effect appears to be fairly robust (Burger 1999) and has been replicated using various manipulations of initial (subsequent) requests such as accepting small (displaying large) campaign signs (Seligman, Bush and Kirsh 1976); completing short (long) surveys (Wang, Brownstein and Katzev 1989); and agreeing to a short (long) meeting request (Harari, Mohr and Hosey 1980). Taken together, work on FITD effects, and self-consistency more generally, leads to the prediction that engaging in an initial act of token support for a cause will lead to an increased propensity to make future meaningful contributions to the cause. It is also noteworthy that many charitable organizations and advocacy groups appear to subscribe to the view that gaining initial token support from consumers leads to meaningful contributions to the cause. One investigation surveyed 53 advocacy groups in the United States and found a prevalent belief that the use of social media strengthens outreach efforts by finding and empowering new members towards furthering the missions of the respective causes (Obar, Zube and Lampe 2011).  2.2.2 The Moderating Role of Social Observability: The Public Versus Private Nature of Support  The extant literature, then, is unclear regarding the downstream consequences of providing token support for a cause. I propose a framework to outline when an initial display of support for a cause does or does not increase consumer willingness to engage in more meaningful subsequent contributions to the cause. In particular, my framework suggests that the consequence of engaging in an initial act of token support for a cause depends on the socially observable nature of the initial token behavior.  13   I make the novel prediction that providing token support will not subsequently lead to more meaningful support for a cause under conditions that activate impression-management motives. Impression management refers to the tendency for individuals to be motivated by a desire to present themselves in a positive light to others (Goffman 1959; Leary and Kowalski 1990; Schlenker 1980). Previous research has shown that impression-management motives prompt consumers to strategically alter their behaviors to present themselves positively (Ashworth, Darke and Schaller 2005; Sengupta, Dahl and Gorn 2002; White and Dahl 2006, 2007).  For example, sociology research shows that, in online contexts, people may use token displays rather than explicit statements of views to construct and communicate positive identities to others (Zhao, Grasmuck and Martin 2008). I first test for the role of impression-management concerns by varying the observability of the initial request. Following past research, impression-management concerns should become activated in more publicly observable settings (e.g., Ratner and Kahn 2002; White and Peloza 2009). Under public conditions, engaging in an act of token support for a cause satisfies these impression-management concerns because the act of support has been viewed by others. As a result, I anticipate that consumers who make an initial act of token support in public will be no more likely to provide meaningful support than those who engaged in no initial act of support.   Conversely, I propose that when the initial act of token support for a cause is done in private, this will to lead to consistency effects, wherein consumers will be more inclined to subsequently make meaningful contributions to the cause. I propose that this result is driven by a process of self-perception and consistency motivation. In particular, when behaviors are not 14  socially observable (i.e., private), the consumer is more likely to be focused on the private self (White and Peloza 2009), responses are less prone to social influence, and attitudes are more likely to be derived from individual values and reasoning (Lamberton, Naylor and Haws 2013; Simonson and Nowlis 2000). I believe that when social observability is low, this focus on the self makes one’s own values and behaviors salient, leading to the inference that one’s values are in line with previous actions. I make this prediction by drawing upon self-perception theory, which states that individuals infer their attitudes towards an object by observing their behavior towards it (Bem 1972). In my context, choosing to provide token support for a cause in a relatively private setting will lead individuals to infer that their attitudes and values must be in line with those of the cause when presented with a meaningful support request (i.e., “I supported this cause before, therefore I must agree with the values the cause stands for”). This focus on the private self will also motivate consumers to be consistent with these values. This is because when people are focused on the private self (versus the public self), their internally held attitudes and beliefs become salient, and inconsistencies among these beliefs become aversive (Gibbons 1990; Goukens, Dewitte and Warlop 2009). In other words, engaging in a token display of support in a private setting activates both a sense that one’s values align with the organization’s values and a motivation to behave in a consistent manner, namely by subsequently providing meaningful support to the cause in the future. As such, I anticipate that consumers will be more likely to provide meaningful subsequent support for a cause after an initial act of token support when social observability is low as opposed to high. Formally:  H1:  Consumers who engage in an initial act of token support for a cause in private will demonstrate increased willingness to provide more meaningful subsequent support than 15  consumers who provide initial public token support, or no initial token support (both of which will not differ from each other).  In addition, I test for the proposed underlying mechanisms. When token support is given in public (versus private), I propose that impression-management motives are satisfied, leading to lower meaningful support for the cause. Moreover, I suggest that after providing an initial act of support for the cause in private (versus public), consumers will show increased consistency motives, leading to higher meaningful contributions to the cause. Formally:   H2a:  Providing public initial token support for a cause satisfies impression-management motives, leading to a lower likelihood of engaging in subsequent support than when the initial token support is private.  H2b:  Providing private initial token support for a cause leads to increased consistency motives, leading to a higher likelihood of engaging in subsequent support than when the initial token support is public.   I test the proposed framework in five studies. In a field setting, study 1 provides a preliminary test of hypothesis 1, such that an initial show of token support leads to greater helping in response to a subsequent, more substantial request when the initial act of support is private as opposed to public (or no support). Study 2 replicates this effect in a controlled environment using the physical signing of a petition as token support. In study 3, I replicate the previous findings and provide additional evidence that impression management and desire for 16  consistency underlie the effects. Using the context of joining a Facebook group, study 4 employs a manipulation of value alignment between supporter and cause to demonstrate the conditions under which a public display of initial token support for the cause can lead to increased helping on a subsequent, more meaningful task. A final field study examines an additional moderator of the effects—connection to the cause.  2.3 Empirical Investigation  2.3.1 Study 1  I sought to provide a preliminary test of my predictions in a field study. This study was conducted at the University of British Columbia between the hours of 11am and 2pm on the last business day before Remembrance Day. On November 11th, Commonwealth countries recognize Remembrance Day to honor veterans who died in World War I. Wearing a poppy pin has come to symbolize support for veterans on Remembrance Day. I chose to test hypothesis 1 by offering a free poppy to individuals in a manner that would serve as either public or private token support for the cause. I compared three groups: one that publicly displayed their initial token support, one that privately engaged in initial token support, and one that was not requested to provide initial support. I then asked participants to engage in the more substantial act of donating money to the cause. I anticipated that individuals who accept a poppy privately will donate more than both individuals accepting and publicly displaying a poppy, and individuals not presented with the initial request, with no differences between the latter two conditions (H1).  17  2.3.1.1 Procedure  Ninety-two individuals (estimated age 19–65, mean age 25.5, 51% female) not wearing a poppy were intercepted by a research assistant as they entered the student union building and were randomly assigned to one of three conditions: Private Token Support, Public Token Support, and No Token Support. In the two token-support conditions, participants were asked if they would accept a poppy to show their support for Remembrance Day. According to previous FITD research (Burger 1999; Freedman and Fraser 1966), it is imperative for participants to choose to agree to the initial request, and not feel pressured or forced into compliance.   In the public token-support condition, participants were asked: “Good morning/afternoon. Would you accept a free poppy to wear right now to show your support for veterans and Remembrance Day?” Upon acceptance of this request, the research assistant ensured that the poppy was placed and visibly displayed on the participant’s coat or shirt. In the private token-support condition, the research assistant said: “Good morning/afternoon. Would you accept a free poppy to take with you to show your support for veterans and Remembrance Day?” Upon acceptance of this request, the research assistant gave the participant a small envelope containing a poppy to privately take with them.   After receiving the poppy, participants had to walk down a narrow concourse in order to enter the main cafeteria and shops. I positioned another research assistant at the end of this concourse who then made a second, more substantial request of participants. Specifically, this research assistant requested monetary donations on behalf of Canada’s war veterans, greeting 18  each participant with the following line: “Good morning/afternoon. Would you like to make a donation to support Canada’s veterans?” If the participant chose to donate, he/she placed the donation in a small envelope in a larger donation bin. I selected this procedure to keep the actual donation amount relatively private and to track donation amounts.   In the no token-support condition, individuals were approached using the same criteria as the experimental conditions (i.e., not wearing a poppy) and received the second request only. Finally, a third research assistant was covertly positioned between the two request locations to note the gender and approximate age of participants, and to subtly signal the second assistant which individuals to solicit with the subsequent request. Given that this investigation is interested in the subsequent behaviors of those who have freely chosen to engage in the initial act of token support, I note that the follow-up request was not made to two participants who did not initially accept the poppy.  2.3.1.2  Results  A one-way ANOVA with amount donated as the dependent variable was significant (F(2, 87) = 3.62, p < .05). Planned contrasts revealed that, as predicted in H1, participants in the private token-support condition donated significantly more money than did participants in both the public token-support (MPrivate-Support = .86, SD = .50 vs. MPublic-Support = .34, SD = .38; t(87) = 1.96, p = .05) and no-support conditions (MPrivate-Support = .86 vs. MNo Support = .15, SD = .36; t(87) = 2.65, p < .01), with these latter two conditions showing no difference from each other (MPublic-Support = .34 vs. MNo Support = .15; t(87) = .81, p > .40, NS, see figure 2). Similar results and 19  significance levels emerge if the percentage of participants agreeing to the donation request is used as the dependent variable, PPrivate = 45.0% (9/22), PPublic = 30.0% (6/35), PNo Support = 25.0% (5/33) .  In addition, I tested whether gender, race or ethnicity moderated the results given that previous research suggested differences in donation propensity are possible (e.g., Rooney et al. 2005). However, in this and all subsequent studies, gender, race and ethnicity had no impact on the results.   2.3.1.3 Discussion   In a real-world field study with a diverse age sample, I provide preliminary support for the hypothesis that the nature of an initial act of token prosocial support (i.e., public versus private) moderates the degree to which consumers will subsequently contribute to the cause. Specifically, individuals who perform a private initial act of token support for a cause donate more money to the cause when subsequently asked to do so than do those who engage in public token support or those who engage in no token support. Individuals providing public token support for a cause were no more likely to provide meaningful support than were those who did not engage in an initial token act of support. Interestingly, neither FITD nor moral-licensing frameworks predict differences in donation behavior between public and private token-support conditions. Specifically, a FITD framework would predict that after agreeing to the small request, individuals in both public and private support conditions would donate significantly more to the cause than individuals in the control condition, regardless of the public nature of support. Moreover, a moral-licensing framework would predict that after agreeing to the small 20  request, individuals in both support conditions would donate less than those in the control condition. The results of study 1 supports the proposition that the socially observable nature of token support is a key moderator of when initial support leads to meaningful support for the cause. As a final point of interest, in addition to the $100 donation made to acquire the poppies used in this study, participant donations amounted to more than $50, all of which was subsequently donated to the cause.  2.3.2 Study 2   Study 2 had two primary goals. First, I wished to replicate the findings from the field study in a more controlled laboratory setting. Second, I aimed to increase generalizability of the findings by examining another common form of initial token support, and by increasing the length of time between initial and subsequent requests. Although previous research has shown that the FITD effect holds when the time between requests is short (Chartrand, Pinckert and Burger 1999), replication with a longer period between requests would enhance the robustness of my findings.   Pre-Test. In order to determine the most appropriate charitable causes to use in this study, a pre-test was conducted. Consistent with previous research, I ensured that participants felt as though they chose to support a specific cause (Burger 1999), and sought to identify two different causes that were positively as well as similarly viewed among the population of interest. Thirty-two undergraduates were asked how likely they would be to donate to 18 different charitable causes on a scale of 1 (very unlikely) to 7 (very likely; e.g., White and Peloza 2009). Natural-21  disaster relief and developing-world poverty were selected as the two causes because both were rated significantly above the scale midpoint (MDisaster Relief = 5.66; t(31) = 6.46, p < .001; MPoverty = 5.41; t(31) = 4.68, p < .001), but not significantly different from each other (t(31) = 1.09, p > .28).  2.3.2.1 Main Study Procedure  Ninety-three English-speaking undergraduates from the University of British Columbia participated in the study in exchange for course credit or $10, and were randomly assigned to one of four conditions in a 2 (Nature of the Context: Public, Private) x 2 (Initial Token Support: Present, Absent) between-participants experimental design. Participants completed the study at individual computer terminals in groups of 10–15 per session.  Upon arrival to the lab, participants were told that they would be evaluating the effectiveness of communications used by local charitable organizations. Given the results of the pre-test, marketing materials were created to represent two charitable organizations ostensibly operating in the Vancouver area: “Combating Poverty in Developing Nations” and “International Disaster Support Relief.” Participants were told that both organizations were formed recently and were requesting feedback regarding their marketing materials. After reviewing the promotional materials for both charities, participants were presented with information to manipulate both token support and nature of the token-support context.  22  Initial Token Support. In the initial token-support conditions, participants were told that both organizations were lobbying the national government via a petition to increase funding for their causes, and that participants could choose to help by signing one of the two petitions (see Appendix A for specific descriptions). Which charity to support (and whether to support one) was the participant’s own decision. In the initial no-token-support conditions, participants were told that in order to ensure proper credit for participation, they should sign their name on one of the two attendance sheets.   Nature of Context. In the public conditions, participants walked to the front of the room and signed either one of the two organization petitions, or one of two attendance sheets. In the private conditions, participants were given a ballot to sign at their individual computer stations then were instructed to place their signed ballots in a box at the back of the room to either indicate their support for one of the two charitable petitions or to track attendance.   Participants then answered cover-story questions regarding the organization they chose to support. In the no-token-support conditions, participants were randomly assigned to evaluate one of the charities. In addition, participants answered the following manipulation check item on a scale of 1 (very private) to 7 (very public): “How public or private was the petition (attendance sign-in procedure) that you chose to sign?” In addition, participants completed a measure to assess the degree to which their own values were congruent with those of the cause (see study 4 follow-up for the specific items). After answering those questions, participants completed an unrelated study. At the end of the experimental session, participants were presented with the dependent variable: the larger request for support. 23   Dependent Variable. Approximately 45–60 minutes after agreeing to the initial request, participants were told that the charitable organization they chose to support through petition-signing (or evaluated in the no-support conditions) was requesting support from university students to help launch a mail campaign. Specifically, help was required to stuff envelopes for this campaign. The amount of time participants were willing to volunteer (on a sliding scale of 0 to 150 minutes) served as the dependent variable.   2.3.2.2 Results   Manipulation Check. The manipulation check revealed that efforts to vary the social observability of context were successful. Participants perceived signing the petition/attendance sheet at the front of the room to be relatively more public than placing the signed ballot in a box at the back of the room (MPublic = 5.40 vs. MPrivate = 4.07; t(87) = 3.42, p < .001). I note that although the mean in the private condition is at the scale midpoint, the interest is in the relative difference in perceived observability of the initial act of token support across conditions.    Dependent Variable. Four participants (1 public support, 2 private/no support, 1 public/no support) chose not to sign the petition or attendance sheet and were excluded from the analysis. As such, the analysis is conducted using eighty-nine valid participants. In order to test my hypotheses, I conducted a 2 x 2 ANOVA with the number of minutes volunteered as the dependent variable. The ANOVA revealed no effect of either initial token support (F(1,85) = .63, p > .40) or nature of context (F(1,85) = .80, p > .35) variables. However, the predicted 24  interaction was marginally significant (F(1,85) = 3.34, p = .07). I note that when I included value alignment as a covariate, it was significant (F(1,84) = 13.4, p < .001) and the interaction reached significance when including this covariate (F(1,84) = 4.53, p < .05). Planned contrasts revealed that, as anticipated, after an initial act of token support, participants were willing to volunteer significantly more time when the context was private as opposed to public (MPrivate-Support = 56.8, SD = 45.0 vs. MPublic-Support = 32.3, SD = 31.1; F(1, 43) = 4.64, p < .05). As would be predicted from the framework, no differences emerged in the support-absent conditions (MPrivate-Absent = 33.2, SD = 41.8 vs. MPublic-Absent = 41.6, SD = 50.9; F(1, 42) = .36, p > .55, NS, see figure 3). As a final note, mood was also measured to ensure that differences did not account for the observed effects. As expected, no main effects or interactions emerged when predicting mood in this or subsequent studies and will not be discussed further. As mentioned in study 1, neither gender nor cultural background affected the results.  2.3.2.3 Discussion  Study 2 replicated the field study findings in support of H1 such that when participants engaged in an initial act of token support towards the cause, they volunteered significantly more time in response to a subsequent request when the action had been undertaken in private as opposed to public. When no initial act of token support was made, I did not observe differences in willingness to contribute time as a function of the nature of the context.   I note that although planned contrasts supported my theorizing, the interaction only reached marginal significance. Because control participants were randomly assigned to evaluate 25  one of the charities (i.e., I did not let them choose a cause to support as in the other conditions), I believe it is appropriate to consider this analysis using perceived value alignment as a covariate. When included in the model the interaction reached significance, and contrast significance levels remained the same. Although I did not have the statistical power to conduct a moderation analysis, this organizational value alignment finding is very interesting, and is returned to in study 4.    2.3.3 Study 3  Taken together, the results of studies 1 and 2 provide support for the notion that an initial act of public token support is no more effective than no initial act of support, and less effective than a private act of token support, in motivating meaningful contributions to a cause. In study 3, I sought to provide evidence for the proposed processes underlying the effects. Although the manipulation of the public versus private nature of the support context (as in studies 1 and 2) is often used to infer that impression-management concerns are operating (e.g., Leary and Kowalski 1990), in the current study I wished to generate additional evidence for the proposition that a decreased willingness to provide meaningful support might be driven by the resolution of impression-management concerns among those whose initial support was made in public (H2a). This goal is achieved by employing a repeated-measures design and measuring impression-management motives before and after participants engaged in token support. If the logic underlying my conceptualization is accurate, and the act of engaging in a token display of support in public resolves impression-management motives, I should observe a decrease in 26  impression-management motives between time one and time two for participants providing public but not private token support.   I also wished to provide direct evidence that engaging in private (versus public) token support leads to an increased desire to be consistent with one’s values and behaviors. As noted above, my conceptualization proposes that when the setting is private in nature, individuals will focus on the individual, private self. This focus on the private self will lead to a) greater perceived value alignment between the self and the cause, as well as b) a desire to be consistent with those values. This is because when social observability is low, this focus on the individual self makes one’s own values and behaviors salient (Gibbons 1990). The act of providing token support for the cause in private then leads to the inference that one’s values are in line with previous actions (i.e., high value alignment; e.g., Bem 1972). In addition, people will become motivated to be consistent with these values because the focus on the private self makes internally held attitudes and beliefs salient, and inconsistencies among these beliefs become aversive (e.g., Gibbons 1990; Goukens et al. 2009). This desire for consistency will lead to increased meaningful support when token support is given in private as opposed to public (H2b). I test this prediction by measuring consistency motives after participants engage in token support, and predict that participants will show higher consistency scores after providing private (versus public) token support.   A third goal of study 3 was to employ a true control group. I wished to replicate the results of studies 1 and 2 while adhering more closely to previous FITD frameworks. One limitation of study 2 was that participants in the support-absent conditions were still exposed to 27  the charitable organization materials, and as such, they did not represent a true control group. Although the results of study 2 support my predictions, traditional FITD control conditions present participants only with the larger request. A separate control condition is employed in study 3 that presents only the second, more meaningful request. Moreover, to eliminate any misperceptions of time valuation (i.e., when selecting from a sliding scale of minutes to volunteer), in this study participants responded to the meaningful support request with a yes or no response, which holds the time-donation amount constant.  2.3.3.1 Procedure  One-hundred and thirty-two English-speaking undergraduates from the University of British Columbia (ages 18–56, Mage = 22.7, 56.3% female) participated in this study in exchange for $10 and were randomly assigned to one of three conditions in a between-participants design: Public Token Support, Private Token Support, No Support. Similar to study 2, participants arrived to the lab in groups of up to 15 people and were seated at individual computer terminals.   Impression-Management Motives, Time 1. I first sought to test the claim that providing public (versus private) token support satisfies impression-management motives (H2a). Before participants examined the marketing materials or engaged in an initial act of token support, they were directed to answer six questions regarding current impression-management motives on a 1 (strongly disagree) to 7 (strongly agree) scale (e.g., “I care about how positively others view me,” “I want to present myself in a positive way,” White and Peloza 2009).   28  Token Support. Participants then received the same charitable organization instructions, materials, and token-support manipulation used in study 2. As in study 2, the signing of the petition was their own choice. Evaluation of materials and signing of petitions took approximately five to ten minutes.  Impression-Management Motives, Time 2. Upon providing token support, participants were instructed to answer a series of follow-up questions before continuing with the study. These questions included the same six impression-management items assessed at the beginning of the study.   Consistency Motives. I next sought to test the consistency component of the framework (H2b). Consistency motives were operationalized by measuring participant scores using a six-item index from the Preference for Consistency (PFC) scale in a range from 1 (strongly disagree) to 7 (strongly agree) (Cialdini, Trost and Newsom 1995; e.g., “It is important to me that others view me as a stable person.”). Participants then completed an unrelated study for approximately 40 minutes before being presented with the dependent variable.   Dependent Variable. Participants were asked if they would be willing to help by stuffing envelopes for 60 minutes for an upcoming mail campaign. Responses were analyzed as a binary variable (not willing to volunteer  = 0, willing to volunteer = 1).   Control Condition. Following the traditional FITD paradigm, participants in the control condition completed only the unrelated study and were randomly presented with the support 29  request from one of the two charitable organizations at the end of the experimental session. They were not exposed to any of the charitable organization materials prior to receiving the meaningful request.   2.3.3.2 Results  Participants. Two participants in the private token-support condition did not sign the petition, and six participants (4 public, 2 private) failed attention or open-ended dependent variable checks and were excluded from the analyses. As such, the analysis is conducted using 124 participants.  Dependent Variable. I first created two dummy variables using the private-support condition as the reference condition and entered both into a logistic regression to predict agreement with the follow-up request. Replicating studies 1 and 2 and supporting H1, participants providing private token support were significantly more likely to provide meaningful support than were participants providing public token support (PPrivate = 77.8% vs. PPublic = 58.7%; b = -.90, Wald χ2(1) = 3.72, p = .05) and participants receiving only the larger request (PPrivate = 77.8% vs. PControl = 56.3%; b = -1.00, Wald χ2(1) = 3.92, p < .05), both of which did not differ from each other (p > .79, see figure 4).   Impression-Management Motives. In order to test the hypothesis that providing public token support satisfies impression-management motives, I contrast-coded token support and conducted a 2 (Token Support: Public, Private) x 2 (Impression-Management Measurement: Pre-30  Support, Post-Support) mixed-model ANOVA with impression-management measurement as a within-subjects factor. The impression-management measures showed acceptable reliability (αPre = .74, αPost = .83). Results revealed a marginal main effect of support type (MPrivate-Support = 5.73, SD = .57 vs. MPublic-Support = 5.50, SD = .83; F(1, 89) = 2.83, p < .10), a marginal main effect of impression-management measurement (MPre-Support = 5.66, SD = .66 vs. MPost-Support = 5.57, SD = .78; F(1, 89) = 3.32, p < .10), and a significant interaction between support type and impression management (F(1, 89) = 5.74, p < .05). Participants showed a significant decrease in impression-management motives after providing public token support (MPre-Support = 5.61 vs. MPost-Support = 5.39; F(1, 45) = 6.80, p < .05); however, participants showed no change in impression-management motives after providing private token support for a charitable organization (MPre-Support = 5.72 vs. MPost-Support = 5.75; F(1, 44) = .24, p > .60, see table 3). This result provides direct evidence for hypothesis 2a that engaging in public token support satisfies impression-management motives. To test whether satisfied impression-management motives led to a decrease in agreement with the follow-up request, I entered post-support impression-management scores into a logistic regression to predict volunteering. As predicted, lower impression-management scores led to a marginal trend indicating lower agreement to volunteer (b = .51, Wald χ(1) = 2.95, p < .09). I next tested the overall indirect effect following Preacher and Hayes (2007) bootstrapping procedure of 5,000 samples with replacement. As would be expected given the marginal result of impression-management scores on agreement with the volunteer request, the 95% CI included zero (b = -.14, CI95: -.49, .03), indicating that impression management satisfaction did not significantly mediate the effect of social observability on lower agreement with the volunteer request. I return to this result in the discussion section.  31  Consistency Motives. The six consistency items showed acceptable reliability and were averaged to form a consistency index (α = .78). In line with my conceptual framework and supporting H2b, participants exhibited higher consistency motives after engaging in private versus public token support (MPrivate-Support = 5.21, SD = .84 vs. MPublic-Support = 4.79, SD = .98; F(1, 89) = 4.90, p < .05). I next entered consistency scores into a logistic regression to predict agreement with the volunteer request. Consistent with the conceptual framework, higher consistency scores significantly predicted agreement to volunteer to stuff envelopes (b = .52, Wald χ2(1) = 4.08, p < .05). Following Preacher and Hayes (2007) bootstrapping procedure, I tested the indirect effect of token support type on volunteering through consistency. The 95% CI did not include zero (b = -.22, CI95: -.66 to -.01). This result supports H2b and suggests that providing an initial act of private token support leads to higher consistency motives, which, in turn, lead to higher subsequent support for the cause.    2.3.3.3 Discussion  Replicating the results of studies 1 and 2, study 3 shows that providing private token support leads to more meaningful subsequent support than does providing public token support or no token support (H1). Importantly, however, the results of study 3 also support the hypothesis that providing public (as opposed to private) token support for a cause leads to a resolution of impression-management motives, which, in turn, leads to a lower likelihood of agreeing to provide meaningful support for the cause (H2a). Moreover, providing private (versus public) token support leads to higher motivation to behave consistently, resulting in an increased willingness to provide more meaningful support in response to a subsequent request (H2b).  32   Although the findings from study 3 support the conceptual model, one limitation should be noted. The resolution of impression-management motives only marginally predicted the agreement to the volunteer request, resulting in a nonsignificant indirect effect for impression management at conventional significance levels. One possibility is that this is due to the nature of impression management, along with the procedure employed in this study. In this study, impression-management motives were satisfied after participants engaged in public token support for a charitable organization; an act that presumably communicated a positive impression, as well as support for the cause, to others. However, later in the experimental session, participants were presented with consistency motivation measures. Although the dependent variable was not viewed by others (i.e., it did not objectively allow for impression management), for a small subset of individuals, completing the consistency task might have activated a desire to continue to impression manage in consistent ways by helping the cause when asked to do so on a subsequent request. Those who were particularly attuned to the social context in which the study was held (i.e., Fenigstein, Scheier and Buss 1975), for example, might have been motivated to impression manage in this way. This may have led to a slight dilution of the impression management effect on the dependent variable. I do note, however, that the pattern of results is supportive of my framework and that the confidence interval was close to reaching significance (i.e., the indirect effect is significant at the 10% level of significance, b = -.14, CI90: -.46, -.001). I also note that this novel methodology of measuring impression-management motives both before and after participants engaged in token support provides strong support for the proposed process. Individuals who had the opportunity to provide token support in a public context reported lower impression-management motives at time 2 than did those who merely had 33  the opportunity to give token support in a private context. This provides compelling evidence that engaging in public token support resolves the desire to impression-manage, as my conceptual framework suggests.  In order to further examine the role of impression management and provide strong support for H2a, an additional study that took a different approach was conducted. Spencer, Zanna and Fong (2005) argue that “when a psychological process is difficult to measure, but easy to manipulate, a moderation-of-process experimental design is most appropriate” (Spencer et al. 2005, 848). Following this logic, I manipulated the opportunity to impression manage by varying whether the cause requesting support was a positively viewed nonprofit or neutrally viewed for-profit organization. An impression-management view would predict that an initial act of providing token support for a cause will allow the consumer to present a positive self-image to others only if the organization being supported is viewed in a positive light, and not if it is viewed as being neutral. Undergraduates (n = 238) took part in a 2 (Nature of Context: Public, Private) x 2 (Type of Organization: Positive, Neutral) between-participants experimental design in which participants provided either public or private token support for either positively-viewed (causes from studies 2 and 3) or neutrally-viewed (local automotive repair and financial planning) organizations, then received a volunteer request at the end of the session (online token support was provided using the Facebook paradigm presented in study 4). A pre-test providing descriptions of the organizations confirmed that support for the charitable causes provided significantly higher impression-management opportunities than did support for the for-profit organizations (e.g., “Others would view supporting this charity/firm in a positive light,” “I would view people that support this charity/firm in a positive light”; all positive versus neutral 34  contrast ps < .001). The results of the main study revealed that when an initial act of token support satisfied the need to impression-manage (i.e., the organization is positively viewed), participants were less likely to help on a subsequent task when the token support was publicly observable (as opposed to private; p < .05). However, when providing the act of token support did not offer any impression-management possibilities (i.e., the organization is viewed as being neutral), no differences in subsequent meaningful support emerged between public and private initial acts of support (p > .50). These findings suggest that the observed differences between the public- and private-support conditions emerge in contexts where an initial act of token support allows for impression management to occur, but not in contexts where this initial act of token support does not allow for impression management.   To summarize, study 3 provides multiple lines of evidence regarding the role of impression management. First, it shows that the social observability of the initial act of token support moderates the effect in predicted ways. Second, the employed repeated-measures design demonstrates that engaging in public token support satisfies impression-management motivation. Third, reduced impression-management motives marginally predict lower meaningful support. Finally, by manipulating the opportunity for impression management, the follow-up study shows that the decrease in meaningful support after providing public token support occurs only when the supported cause allows consumers to impression-manage.     35  2.3.4 Study 4   Taken together, the previous studies show that an initial act of token support for a cause can either have no impact or create an increase in subsequent support for a cause, depending on the social observability of the initial support. Importantly, consumers have multiple avenues open to them to engage in very public forms of token support (e.g., wearing a physical symbol in support of a cause, putting a symbol on one’s car, liking or joining a group on Facebook), in ways that are relatively easy and costless. One question that arises, then, is whether a way exists to overcome the tendency to withhold more meaningful contributions to the cause following an initial act of public support.    As noted earlier, I propose that consumers engaging in private token support (versus public) will be motivated to focus inward. This inward focus will subsequently lead them to perceive that the values of the cause are truly in line with their own values (e.g., Bem 1972), and this will result in a desire to be consistent with those values when subsequently asked to help the cause in a more meaningful way. In this study I manipulate the perceived value alignment between token supporter and cause. This methodology allows me to provide evidence that perceived value alignment between self and cause is the mechanism behind the tendency to be more likely to comply with a subsequent substantial request after an initial act of token support is made in private as opposed to public. I propose that if consumers who have made an initial public (versus private) display of token support are asked to focus on how their own values align with those of the cause, the tendency to show decreased support upon a subsequent, more meaningful request for help can be reversed. Conversely, I propose that if consumers who have 36  made an initial public (versus private) display of token support are asked to focus on how their own values are misaligned with those of the cause, no increase in value alignment would be experienced because their token support would be viewable by others, thus replicating the results of studies 1–3. As such, I propose that whether or not the individual is focused on value alignment at the time of token support will moderate the relationship between social observability and meaningful subsequent support.  H3:  Those who provide public token support while focusing on the alignment (misalignment) of values between self and cause will show an equivalent (lower) willingness to provide meaningful support compared to those who provide private token support.   In study 4, I extend the generalizability of my findings by investigating an additional medium as the initial act of token support: joining a Facebook group. The act of joining a group on Facebook allows me to examine if the observed effects emerge not only in public contexts where strangers will view the initial act of support (as in studies 1, 2 and 3), but also in contexts where the initial act is made public to one’s friends and acquaintances (Wilcox and Stephen 2013).     37  2.3.4.1 Procedure  One hundred and one English-speaking undergraduates (ages 17–26, Mage = 19.8, 58.7% female) from the University of British Columbia completed this study in exchange for course credit and were randomly assigned to one of four conditions in a 2 (Nature of Context: Public, Private) x 2 (Value Alignment Focus: Values Aligned, Values Misaligned) between-participants design.   Upon arrival to the lab, participants were told that two recently formed local charitable organizations were requesting feedback regarding the effectiveness of their online communication tools. Specifically, feedback was desired regarding their Facebook groups. The two charitable organizations were the same as those used in studies 2 and 3. I constructed two Facebook groups based on the written and graphical content previously used in studies 2 and 3.   After the introduction, participants were directed to log into their personal Facebook accounts to view the Facebook group pages of the two charitable organizations, and to then choose to join one of the groups. The joining of the Facebook group served as the initial token support for the organization. I manipulated the public versus private nature of the task through Facebook group structure. In the public (private) condition, the groups were set up as ‘Public’ (‘Closed’) Groups. Joining this type of group appears (does not appear) on the public personal feed of each user. Moreover, any posts or messages sent out from the group appear (do not appear) on the public personal feed of each user. In addition, I included a public (private) 38  statement within the Mission Statement. Participants in the public (private) condition read the following message in the Mission Statement section of the Facebook group description:  This Facebook Group is a public (private) group dedicated to our friends and supporters. Information posted here will (only) be viewable to you, and will (not) appear on your Facebook feed, so your actions here will (not) be viewed by your friends.  After joining one of the Facebook groups, participants were presented with the values manipulation. They were given a newspaper article ostensibly written by a local reporter about the recent activities undertaken by the organization that participants chose to support. The specific activities the article discussed were ambiguous in nature; that is, participants could interpret them as positive or negative (e.g., lobbying the federal government for special status to join a fund, publicly protesting the appointment of a federal official who had been linked to fraud accusations with previous charities). After reading the article, participants in the values-aligned (misaligned) condition were given the following instructions:  Before you move on to the specific questions, please take 1–2 minutes and reflect on how this cause's values are similar to (different than) your values. Once you've thought about this, please describe in the space provided this value (mis)alignment in as much detail as possible.   Participants then completed the same cover-story questions, nature-of-context-manipulation check questions from previous studies, and the following value-alignment 39  manipulation check questions on 1 (not at all) to 7 (very much) scales: This cause reflects my values; My personal values are aligned with this organization’s values; and I feel a personal connection to this cause. Then, participants moved on to an unrelated study for the rest of the session before being presented with the dependent variable at the end. Similar to previous studies, the subsequent request was programmed to be from the charitable organization whose group the participant chose to join at the beginning of the session.  Dependent Variable. Participants were asked if they would be willing to help by stuffing envelopes for an upcoming mail campaign (0 = No, 1 = Yes).   2.3.4.2 Results  Participants. One participant was not able to join the Facebook group due to experimenter error (did not see join group request), and 25 participants chose not to join either Facebook group (NS across groups, p > .20) and were excluded from this analysis. Therefore, the analysis is based on 74 cases.   Manipulation Checks. Participants in the public condition rated the context of the task as significantly more public than did participants in the private condition (MPublic = 5.77 vs. MPrivate = 4.23; t(72) = 3.59, p < .01). Moreover, participants in the values-aligned condition viewed their values as more in line with the cause than did participants in the values-misaligned condition (MValues Aligned = 5.37 vs. MValues Misaligned = 4.90; t(72) = 2.30, p < .05).  40  Dependent Variable. In order to perform the analysis on agreement to the second request, I contrast-coded both nature of context (Private = -1, Public = +1) and value alignment (Values Aligned = -1, Values Misaligned = +1) independent variables. I entered nature of context, value alignment, and their interaction term into a logistic regression to predict agreement to provide subsequent support. Results revealed no effect of nature of context (b = -.23, Wald χ2(1) = .63, p > .40), but a significant main effect of value alignment (b = -.78, Wald χ2(1) = 6.93, p < .01). Participants in the values-aligned condition were significantly more likely to agree to the subsequent request than were participants in the values-misaligned condition. Importantly, this main effect was qualified by a significant interaction with the nature of the context (b = -.59, Wald χ2(1) = 4.14, p < .05). When participants thought about how their values were misaligned with those of the charitable organization, the previous effects were replicated. In particular, participants providing an initial act of public token support were significantly less likely to agree to the subsequent support request than were participants providing initial private token support (MPublic = 31.6% (6/19) vs. MPrivate = 70.6% (12/17); χ2(1) = 5.46, p < .05). However, when participants thought about how their values were aligned with those of the charitable organization, participants providing public support were equally as likely to agree to the follow-up request as those providing private initial support (MPrivate = 77.3% (17/22) vs. MPublic = 87.5% (14/16); χ2(1) = .65, p > .40, NS, see figure 5).   2.3.4.3 Discussion  The results of study 4 highlight how charitable organizations can use value alignment to combat slacktivism, turning initial acts of public token support into more meaningful subsequent 41  support. I show that by focusing those who engage in an initial act of token support in public on the value alignment between self and cause, helping on a subsequent, more meaningful task can be increased. Indeed, when participants first thought about value alignment, those who made an initial public display of support became just as likely as private supporters to agree to a subsequent, more substantial contribution to the cause. Moreover, replicating the results of studies 1–3, participants who provided private token support but focused on the misalignment of values exhibited higher meaningful support for the cause did than participants providing public token support. One question that may arise, then, is why a focus on misalignment under private conditions did not lead to a decrease in subsequent support. One possibility is that this is due to individuals inferring that because they freely and privately chose to support the cause via an initial token act, their values are aligned with those of the cause, and they were more motivated to behave consistently with these values. In other words, the value alignment from freely choosing to support the cause in private appears to be more impactful than the misalignment manipulation.   By manipulating value alignment, I further support my proposed process of value alignment and consistency motives underlying the greater willingness to agree to subsequent support among private versus public token supporters. One additional follow-up study was conducted that conceptualized value alignment as a mediator through measurement. Using an online sample (n = 138), online petitions as the token support medium, and poaching of the African tiger as the cause, participants were randomly assigned to public or private token-support conditions and told that the study was investigating the effectiveness of online communication tools used by charitable organizations. Two identical petition websites were constructed under 42  the charitable organization name Save Our Tigers, which were used to manipulate the public (i.e., donor names would be publicly displayed) or private (i.e., names would be kept confidential) nature of the token support. After signing the petition and answering cover-story questions, participants completed the same value-alignment questions used in study 4. Participants were then presented with the dependent variable: willingness to donate their MTurk payment (0 = No, 1 = Yes). Ratings of value alignment significantly mediated the effect of social observability on monetary donations, supporting my attitudinal consistency process claim (H2b).   Study 4 also casts doubt upon a potential alternative explanation for these findings. It is possible that participants in previous studies perceived the cost of the token support task to be higher in public versus private. In this case, the reduced subsequent support in the public-support conditions could be due to the perception that one has already given enough to the cause through previous time and effort. Study 4 rules out this explanation because both initial support tasks were equivalent in both required time and effort (one mouse click).   2.3.5 Study 5  Studies 1 through 4 provide a thorough test of my conceptual framework. The goal of study 5 is to examine one final boundary condition for the observed effects. In particular, I examine the role of connection to the cause as a moderator of the observed effects. I define connection to the cause as the level of affective involvement and identification an individual has with a social cause’s mission and goal (e.g., Bhattacharya and Sen 2003). This can be contrasted with the construct of value alignment, which I define as the cognitive perception of congruency 43  between one’s own values and an organization’s values (see Posner 1992 for a similar conceptualization in the organizational domain). It is reasonable to propose that those highly connected to a cause are more likely to engage in public forms of token support, and that this might also spur more meaningful contributions to the cause. Research examining citizenship behaviors in organizations supports this proposition, showing that when both prosocial and impression-management motives are high, individuals are more willing to engage in costly organizational citizenship behaviors such as verbally disagreeing with the opinion of the group (Grant and Mayer 2009). Displaying public token support may in fact validate the affect of those who are highly connected to the cause and motivate these individuals to support further. Conversely, displaying private support could actually generate dissonance within these individuals because it limits their ability to communicate this important cause’s need for support to others (Festinger 1957). Specifically, I predict that individuals highly connected to the cause will be more likely to agree to subsequent meaningful support after providing public (versus private) token support; and for individuals less connected to the cause, I expect to replicate the results of earlier studies such that providing private (versus public) token support will lead to higher consent to a subsequent request. Formally:  H4a: Consumers highly connected to a cause will show a higher willingness to provide meaningful subsequent support when the token support is public as opposed to private.  H4b: Consumers not connected to a cause will show a higher willingness to provide meaningful support when the token support is private as opposed to public. 44   I chose to test my hypotheses by conducting a second field study. Moreover, to extend the generalizability of my framework beyond charitable causes, I chose to investigate a cause that is a positively-viewed for-profit organization: the local National Hockey League team, the Vancouver Canucks.   2.3.5.1 Procedure  The field study was conducted on a busy sidewalk in downtown Vancouver in the early afternoon on two separate days. Eighty-eight individuals (estimated age 18–65, Mage = 30.9, 37.5% female) walking alone were intercepted by a research assistant and randomly assigned to one of three conditions: Private Token Support, Public Token Support and No Support. One possible critique of the study 1 manipulation is that although participants in the private-support condition received their poppy in an envelope, the poppy might eventually be displayed to others. In this second field study, I manipulated the public versus private nature of token support by varying the social observability of the product accepted as the token act. Specifically, in the public (private) condition, participants were asked to accept a pin (fridge magnet) of the Canucks team logo. The magnet and pin were of identical size and depicted the same logo. However, the pin is higher in public observability, while the magnet is lower in observability. Participants in the public (private) support condition were greeted with the following line: “Good afternoon. Would you accept this free pin to wear right now (magnet to take with you) to show your support for the Vancouver Canucks?” Upon acceptance of this request by the respondent in the public condition, the research assistant ensured that the pin was visibly displayed on the participant’s 45  jacket or shirt. After the participant’s acceptance of the pin or magnet, the research assistant asked the connection-to-cause measure. Specifically, participants were asked, “On a scale of 0 (low) to 10 (high), how big of a Canucks fan are you?” A second research assistant was positioned on the sidewalk at the end of the block who intercepted the participant approximately one minute later with the larger request: a five-minute survey on behalf of the Vancouver Canucks. Although this is a smaller subsequent task than those employed in the previous studies, this dependent variable is of significant inconvenience to these individuals who were predominantly in a hurry. Participants in the control condition were randomly selected to receive only the larger request, and the connection-to-cause measure was asked by the second research assistant. After responding to the larger request, participants were fully debriefed and informed that this study was conducted by UBC researchers and was in no way affiliated with the Vancouver Canucks. Finally, a third research assistant was unobtrusively positioned between the two others to record gender and approximate age. Thirty-one participants (16 private, 15 public) did not accept the magnet/pin and did not receive the larger request, and 6 participants did not provide a connection-to-cause score. As such, analyses were performed using 82 participants.  2.3.5.2 Results  Hypothesis Tests. To test H4, connection to the cause was mean-centered at 5.11 (SD = 3.23) and condition was contrast-coded. I entered condition variables, connection to cause, and their interactions in a logistic regression to predict agreement to the larger request. I entered day as a covariate because the Canucks lost a playoff game in between experimental days. The covariate was not significant (p > .60), and results and significance levels remain unchanged if it 46  is excluded. Results revealed no main effect of either condition (C1Pri-Pub: b = .91, Wald χ2(1) = 2.63, p > .10; C2Pri-Ctrl: b = .10, Wald χ2(1) = .06, p > .80) or connection-to-cause independent variables (b = .21, Wald χ2(1)  = 1.74, p > .15). However, the expected interactions between each condition variable and connection to the cause were significant (C1 interaction: b = -.67, Wald χ2(1) = 9.55, p < .01; C2 interaction: b = -.59, Wald χ2(1) = 12.1, p < .01). Each interaction was analyzed using the Johnson-Neyman floodlight analysis technique following Spiller et al. (2013), which allowed me to probe the interactions to identify the ranges of connection to the cause where the predicted simple effects of token support reach significance (Johnson and Neyman 1936). Replicating studies 1 through 4, providing private (versus public) token support led to higher agreement to the larger request among those less connected to the cause, or .21 SD below the mean value of connection to cause (BJN = 1.29, SE = .66, p = .05). However, providing public (versus private) token support led to higher agreement to the larger request among participants highly connected to the cause, or .75 SD above the mean value of connection to cause (BJN = -.78, SE = .40, p = .05). The interaction between private token support and no token support exhibited the identical pattern (low connection to cause: -.54 SD; high connection to cause: +.47 SD, see table 1).   2.3.5.3 Discussion  Using a real-world field setting, study 5 highlights a boundary condition of the previously documented effects. Among individuals less connected to the cause, I replicate the previous findings such that providing private token support leads to a higher likelihood of meaningful support than does providing public token support, or no support. Conversely, among those highly 47  connected to the cause, providing public token support leads to higher meaningful support than initially providing private support. This suggests that the reinforcement charitable organizations receive for their token-support campaigns may reside among the champion supporters of the respective cause, and not among average consumers. Finally, by examining a positively-viewed for-profit organization, study 5 also demonstrates that my framework extends beyond charitable organizations and may be applicable to for-profit brands or firms, a potentially provocative opportunity for future research to investigate.   2.3.6 General Discussion  Using multiple field and experimental settings, different operationalizations of initial acts of token support, and varied measures of more meaningful subsequent contributions, I provide an empirical investigation of slacktivism and when it is most likely to emerge. Across five studies, I demonstrate the existence of slacktivism, wherein an initial act of token support does not lead to an increased willingness to provide more substantial contributions to the cause. I propose a conceptual framework to outline when and why an initial act of token support may or may not lead to an increased contribution on a subsequent, more meaningful helping task.   In study 1, I show that individuals who engage in an initial act of private token support (privately accepting a poppy pin) for a cause donated more money to the cause than did those who engaged in an initial act of public support (publicly wearing a pin), or those who did not engage in an initial act of support. In study 2, I replicate this effect in a controlled environment using the signing of a petition as the token act of support for the cause. Study 3 provides a more 48  complete test of my full conceptual framework and demonstrates support for the proposed processes. I find that providing public token support satisfies impression-management motives, leading to a lower likelihood to provide meaningful support for a cause. In addition, study 3 shows that engaging in private token support leads to higher consistency motives, and a subsequent increase in meaningful support for the cause. In study 4, I make an important substantive contribution by identifying a tool that charitable organizations can use to combat slacktivism and garner meaningful support from public token-support campaigns: value alignment between supporter and cause. Finally, in a second field experiment, study 5 identifies an additional moderator of the observed effects; specifically, my findings are reversed among those who feel strongly connected to the organization.   2.3.6.1 Theoretical Contributions  This essay makes a number of theoretical contributions. First, I provide initial evidence for when slacktivism will emerge and identify the social observability of the token support as a key moderator in predicting when token support does or does not lead to meaningful support. I introduce a conceptual framework proposing that while in public settings the resolution of impression-management concerns becomes central, in private settings the consistency of one’s own privately held values becomes paramount. Importantly, I identify the social observability of the token support (operationalized as public versus private) as a key moderator in predicting which motive will be active; this moderator predicts when an initial act of token support will lead consumers to subsequently give more meaningful support to the cause.   49  This work adds to the current understanding of impression management by showing that when given the opportunity to provide token support for a social cause, consumers activate both consistency and impression-management motives. However, when given the opportunity to make larger contributions to the cause, satisfaction of impression-management concerns after providing public (versus private) support reduces inclinations to provide subsequent meaningful support. My findings add to previous impression-management research by identifying consistency as an additional motive that consumers are willing to trade off in favor of displaying a positive image to others (Ashworth et al. 2005; Sengupta et al. 2002). Interestingly, this essay also suggests that impression-management concerns can become relevant even in the context of friends. Ashworth et al. (2005) show that impression-management concerns are significantly lessened when in the presence of a close friend (versus on a first date). However, by utilizing participants’ personal Facebook accounts to provide support for a cause in study 4, my findings suggest that impression-management motives are active even among close friends. One explanation of this deviation from prior research could be that Facebook accounts are comprised of both strong ties (family, close friends) and weak ties (acquaintances, Gladwell 2010). Recently, Wilcox and Stephen (2013) showed that consumers do in fact seek to self-enhance even among close ties. My results lend support for their findings. Finally, this essay also adds to the impression-management literature methodologically by directly measuring changes in impression-management motives over time. My results are notable given that past research has suggested that it is often difficult to assess impression management motives through measurement in experimental settings (Bolino 1999). The repeated-measures impression-management task employed in study 3 suggests that the measurement of these motives over time might be one fruitful way to assess changes in the desire to present a positive image to others.   50   This essay also helps to elucidate the nature of opposing predictions from two established literatures by introducing a moderator not previously investigated: the observable nature of the initial task. In doing so, it contributes to both FITD and moral-licensing literatures by empirically demonstrating the importance of this variable to both research streams. I also add to economic models of prosocial-behavior literature by qualifying previous findings and conclusions regarding consumer repercussions of engaging in costless prosocial support (Bénabou and Tirole 2004; Gneezy et al. 2011). Previous research shows that costly prosocial behavior signals to individuals that they are ‘moral’, and as such, leads to increased subsequent prosocial behavior. Conversely, costless prosocial behavior has been shown to signal nothing regarding the prosocial identities of supporters, and leads to subsequent licensing behavior. I show that costless token support can in fact lead to subsequent prosocial behavior, but via a different route—perceived value alignment between self and cause and a desire for self-consistency. The findings reveal that by limiting the impression-management motive behind public token support, or by focusing token supporters on the value alignment between self and cause, meaningful prosocial behavior can result.   2.3.6.2 Managerial Implications  This essay makes a number of important substantive contributions for charitable and other non-profit organizations. First, I empirically demonstrate that public token support does not lead to increased meaningful support for social causes. I also show that this behavior is not limited to the online realms but occurs also with physical symbols of support. I acknowledge that 51  many of the symbols used by charitable organizations are utilized to generate awareness, and that awareness is a critical first step to the success of many charities and social causes. Awareness is extremely important and my goal is not to minimize this positive consequence of token support. However, it is also hoped by charitable and other organizations that promote these acts that such token displays will serve as stepping stones to more meaningful support for the respective causes. Intuitively, both consumers and charitable organizations believe that public displays of support may lead to greater future support than when initial support is not publicly displayed. Importantly, however, my findings demonstrate that this may not always be the case.   The inability of small public displays of token support to motivate subsequent helping behavior is a prevalent issue facing our society given that many charities and social causes have come to rely heavily on such tactics. For example, the ubiquity of “pink” paraphernalia displayed by those who wish to publicly support a cure for breast cancer has come under scrutiny. Many fear that ‘pink-washing’ has detracted the public focus from cancer research and support programs, instead creating a trendy and high-profile way to present a positive image for both individuals and firms (King 2006). My results suggest that under certain conditions, this concern is warranted. Specifically, I find that engaging in these forms of public support activates a desire to present the self in a positive light, and once this desire is satisfied the token act may not lead to increased support for the cause. Importantly, however, these findings offer insight to charitable organizations on how to counteract this behavior, and to better harness the power of small token acts to generate meaningful support for the cause. The findings from study 4 show that focusing consumers on the value alignment between self and cause at the time of public support leads to increased meaningful support for the cause. Charitable organizations can do this 52  in numerous ways. For example, social media content could draw attention to the values the organization stands for, and how these values are congruent with the consumer’s own values.  This value-alignment focus can also be directly communicated through face-to-face interaction when consumers accept physical tokens of support (e.g., ribbons, pins) from organizational volunteers. This essay can provide guidance to charitable organizations planning to undertake a public token-support campaign. I find that this form of support is most effective in garnering subsequent meaningful support among those already highly connected to the cause.    Finally, the results of study 5 potentially raise the question of who charitable organizations are attempting to target with token-support programs. As Zhou et al. (2008) find, top NGOs believe that token-support campaigns succeed in generating meaningful support from individuals not currently engaged in the cause. My results consistently find that public token support promotes slacktivism among all but those highly connected to the cause. If the goal of these programs is to generate new interest in and support for causes via a foot-in-the-door procedure, charitable organizations may be using their precious resources sub-optimally. The results of study 5 suggest that public token-support programs are effective only among those closely connected to their respective causes. While it is certainly vital to keep highly connected supporters motivated, charitable organizations must carefully consider if encouraging public token support is a successful strategy when trying to attract meaningful support from new donors.    53  2.3.6.3 Directions for Future Research  I believe that this essay provides consumer researchers with ample possibilities for future investigations. A natural extension of this work would be to directly examine the conceptual framework using brands and other for-profit organizations. The results from study 5 provide a glimpse into what could be a very useful tool for brands to engage and generate relationships with potential consumers. However, might token support for a brand operate differently than for a non-profit organization? A recent white paper by Lipsman, Aquino and Kemp (2012) suggests that aspects of my conceptual framework may differ, as a small, but significant effect of Facebook liking on sales was found for two high-profile brands (Target, Starbucks). Future research that identifies the conditions driving these differences would certainly add to current understanding.   Moreover, future research in the non-profit sector could examine the social observability of the meaningful request. In this investigation, I deliberately kept meaningful requests relatively private because I was interested in investigating the underlying motives of slacktivist behavior and the relationship between this behavior and token support. However, future research could vary the public nature of the meaningful request to examine whether it or its interaction with token support affects agreement rates. Similarly, future research could vary the cost of token and/or meaningful support. In this essay, meaningful support was more costly to participants relative to token support. My goal was to maximize the generalizability of the findings to important volunteer and donation-support behaviors, but future research could determine if an absolute cost ceiling of the effects exists.  54   Another interesting question for future research is to investigate the impact of communicating the positive behaviours performed by other people on our own subsequent support. Research has shown that direct self-presentation about the self can lead others to question one’s altruistic motives and actual motivation for the support behavior (Fein 1996; Reeder 2009, Berman et al. 2014). To mitigate this potential drawback of directly promoting the self, consumers can resort to other, less invasive, impression-management techniques, such as promoting the behaviours of those we are connected to. For example, a consumer posting about how proud she is of her mother for finishing a marathon, or her workgroup for volunteering for the United Way signals to others that by being associated with active and prosocial people, respectively, the consumer herself also has those traits. What is presently unknown is the effect this communication has on our own behaviour. Preliminary results from my lab suggest that publically communicating the support behaviours of close others may actually reduce our own subsequent support behaviour.    An additional factor outside the scope of this essay is the interaction of token prosocial behavior with moral identity (Aquino and Reed 2002). Aquino and Reed (2002) propose that an individual’s moral identity may be represented as an associative network of related moral traits, goals, and behaviors, and this identity is comprised of two components: internalization (private) and symbolization (public). Winterich, Mittal and Aquino (2013) show that recognition of support, such as a receiving a thank-you card in the mail after making a donation, increases subsequent charitable behavior among individuals high in moral identity symbolization but low in moral identity internalization. Moreover, individuals high in moral identity internalization 55  engaged in more efforts to uphold a moral self-image after being dishonest. An interesting direction for future research would be to examine the proposed relationship between token and subsequent support for a cause among consumers with different levels of the two moral-identity components.   2.3.6.4 Essay 1 Conclusion  In conclusion, this essay provides an empirical investigation into an important issue facing charitable organizations, consumers and consumer researchers: slacktivism. Not only do I provide a conceptual framework to understand when and why the behavior occurs and persists, I also provide tools that charitable organizations can use to combat this behavior and generate the meaningful support needed to achieve their missions.  56  Chapter 3: Losing That Loving Feeling: Why Learning About Charitable Support Allocation Reduces Recurring Donation Behaviour  3.1 Introduction  From five year old children to motorcycle gangs, news outlets are filled with touching stories of individuals cutting and donating their long flowing locks to be made into wigs for children and women suffering from diseases such as cancer and alopecia (Goldberg 2014). Although motivations to support range from losing a family member to cancer to minimizing bullying of suffering children, the goal underlying this generous and effortful behaviour is to make a positive impact in the lives of others.   However, a New York Times article called into the question the impact that these meaningful acts have by suggesting that the allocation of donated hair to wigs was significantly lower than donor expectation. Specifically, the article reported that Locks For Love - a non-profit organization that creates wigs out of human hair for children suffering from alopecia – sells or discards 80% of the hair it receives (New York Times 2007). This news report generated significant reaction among donors and the general public, and led to a subsequent investigation that revealed that the majority of hair received was not suitable for wig production (Nonprofit Investor 2013a). Most pertinent to the current investigation was the effect this information had on donors. Interestingly, donor responses centred on feelings of regret and remorse at not making the difference they thought they had made (Nonprofit Investor 2013b).  57   This example is but one of many of a prevalent social issue facing consumers and non-profit organizations: charitable support allocation. I define charitable support allocation as the proportion of consumer donations that a charity passes directly on to underlying cause recipients. Although registered charitable organizations are required to adhere to Income Tax Act standards, substantial variability exists regarding the proportion of funds that organizations allocate to recipients, fundraising costs, overhead costs and cash reserves (MoneySense 2014). Recent instances of low charitable support allocation have generated significant media attention for fundraisers conducted by both non-profit (e.g., Team in Training, Hays 2013) and for-profit organizations (e.g., Color Run, Leonard 2013) organizations, causing many donors to question the efficacy of their support and the positive benefits received through donating (Hays 2013). Although it has been suggested that overhead costs may affect selections of charities prior to giving (Gneezy, Keenan and Gneezy 2014; Oppenheimer and Olivola 2011), no empirical work has investigated the consequences this information has on consumers after support has been provided. To address this gap, this essay investigates the consequences that low support allocation information has on consumers who have provided meaningful support for a cause, and proposes that learning this information reduces consumer prosocial identity and subsequent support behaviour.   I define prosocial identity as the self-assessment of how helpful one’s voluntary actions have been in benefitting others or society in general. Prosocial support behaviours certainly include support given to charities, but are also comprised of any voluntary action intended to benefit others at a cost to the self (e.g., donating blood, helping a friend move). Although sharing 58  some conceptual similarities to altruism, defined as “a motivational state with the ultimate goal of increasing another’s welfare” (Batson 2014, 6), prosocial identity differs in that it focuses on how consumers perceive their identities along the dimension of their previous prosocial support behaviour. Altruism, on the other hand, focuses on the motivational state regarding a priori reasons for performing the support behaviour.   The debate about of non-profit fiscal practices continues to increase. Although consumer donations continue to increase each year, over 1.5 million registered non-profit organizations exist in the in the United States (National Center for Charitable Statistics 2014). Factor in difficult economic conditions and competition with for-profit companies vying for discretionary income, non-profits face a difficult task to survive, let alone fulfill their missions. To compete for these limited resources, non-profits have adapted to this competitive environment by implementing large-scale marketing programs and attracting, and paying for, highly skilled employees and leaders. The funds to support these business decisions must come from somewhere, and whether it is salient to consumers or not, these funds are drawn from consumer support. As a result, the overall proportion of support allocated to recipients of the underlying causes decreases. Although these managerial decisions are ethically and legally sound, and in fact lead to higher total amounts of support given to the cause recipients (Kristof 2008), research has shown that consumers view these allocation decisions as tainted and judge non-profit decisions much more harshly than those of for-profit firms (Newman and Cain 2014).   As such, the increased transparency of non-profit financial practices and higher-intensity spotlight often adds fuel to the debate among consumers and mass media. Importantly, the goal 59  of this essay is not to suggest that non-profit organizations are inefficient, deceptive or irresponsible, nor is it to minimize the invaluable support they provide to millions of people in need. It does, however, seek to examine the consequences that learning this information has on non-profit supporters, an important question previously unexamined by consumer researchers. Given that consumers experience an increase in prosocial identity after giving meaningful, or costly prosocial support to a cause (Gneezy et al. 2011; Bénabou and Tirole 2004, 2005), how might support allocation information affect this positive self-view? Do consumers internalize the lower perceived impact and experience a prosocial identity-based consequence, or protect their identities by focusing on their support intention (Kruger and Gilovich 2004)? Moreover, does learning that less of one’s support reached the expected target affect consumers’ future prosocial action?   This essay addresses these research questions and explores the consumer consequences of support allocation information. I make the novel prediction that after providing meaningful support for a charitable organization, consumers experience a prosocial identity-based consequence upon learning that a lower proportion of their support was allocated to cause recipients. My conceptual framework proposes that this novel consequence occurs because low support allocation information reduces the positive benefits received from the belief that one’s support made an impact in the lives of others (Aknin et al. 2014). Moreover, I propose that this identity-based consequence leads to lower subsequent support not only for the previously supported organization, but also to other unrelated organizations. Thus, managerial decisions made by one non-profit organization may have detrimental effects on the success of other organizations. This essay makes several notable contributions to the literature. First, I provide an 60  empirical investigation into the unintended consequences of charitable support allocation decisions–a topic receiving increased attention in the mass media. I outline how reducing the impact of consumer support through allocation decisions can actually alter the self-view of an organization’s supporters. In particular, I show that consumers who have provided costly, meaningful support for a charitable organization view themselves as less prosocial in nature when exposed to low (versus high) support allocation information (pilot study, studies 1-3). In doing so, I build on previous research that shows that consumers increase their prosocial identities after providing effortful support for a cause (Gneezy et al. 2011). I also show that support allocation information leads to this consequence unless consumer prosocial identity is affirmed prior to exposure (study 1).   Further, I propose that, in addition to reducing consumer prosocial identity, low allocation support information has detrimental effects on subsequent support. Specifically, my framework shows that this identity-based consequence actually leads to lower subsequent support not only for the previously supported organization (pilot study, study 3) but also to other unrelated causes (study 2). This suggests that legitimate managerial decisions that seek to maximize total donations can actually reduce support among an organization’s most valuable target market: recurring supporters (Flannery and Harris 2009). This finding adds a previously unexplored theoretical factor that qualifies previous economic prosocial behaviour models by showing that perceived impact moderates the documented relationship between prosocial identity and subsequent support behaviour (Gneezy et al. 2011). I also show that the reduced subsequent support is not driven by feelings of deception (e.g., Darke and Richie 2007), but by reduced 61  prosocial identity (studies 1-2), thus differentiating my framework from previous work on deceptive advertising (Pollay 1986; Darke and Richie 2007).   Finally, I identify a boundary condition for the observed effects–success of initial support–and find that the prosocial identity consequence does not occur for consumers who attempt, but fail to provide meaningful support for an organization (study 3). Specifically, consumers who try, but fail to fulfill their meaningful support goal assess their prosocial identities both through actual behaviour and intentions; however, successful supporters judge their prosocial identities solely on their support behaviour. Thus, I qualify previous research on action and intentions by identifying conditions in which consumers do not judge their prosocial identities by factoring in intentions (Kruger and Gilovich 2004).   This essay is organized as follows. First, I review previous literature to outline my theoretical framework and predictions. Second, I report four studies (one pilot, three experimental) that support this theorizing. Finally, I highlight the contributions of this work and explore avenues for future research.  3.2 Conceptual Background and Hypotheses Development  It is no secret that consumers choose to provide meaningful prosocial support to charitable organizations that support important causes and people in need. Consumers may support for a variety of reasons, such as being an altruistic person (Batson and Shaw 1991), receiving emotional benefits from helping others (Aknin et al. 2013; Dunn, Aknin and Norton 62  2008), sympathizing with others’ suffering (Small and Simonsohn 2008) or obtaining personal benefits (i.e., Bénabou and Tirole 2005; Cialdini et al. 1997). However, charitable giving is not a one-time behaviour. Although statistics vary, approximately 29% of the adult population volunteers on a monthly basis (Institute for Volunteering Research 2012). Moreover, recurring financial donors account for approximately 10% of total donation amounts, and research shows that recurring donors donate significantly more per year than one-time donors (Flannery and Harris 2009). As such, understanding the consumer consequences of the initial support experience and its impact on subsequent behaviour is paramount not only for researchers seeking to understand its theoretical implications, but also to charitable organizations that compete for these limited resources. What effect does learning that less of one’s support has reached the desired recipients have on consumer prosocial identity and subsequent behavior? Put another way, does lower support allocation threaten the prosocial identity benefits earned through one’s support? Interestingly, previous research offers conflicting perspectives.  On the one hand, previous research suggests that low support allocation information will not affect consumer prosocial identity or reduce future support due to the motivation to protect the self from ego or identity threats. For example, attribution research shows that individuals are more likely to attend to situational (versus individual) factors when judging their own actions versus those of others, and to attribute failure or other negative events to external circumstances (Carver, DeGregor and Gillis 1980; Greenberg, Pyszczynski and Solomon 1982; Mullen and Riordan 1988), especially when the failure is important to the self-concept (Klein and Kunda 1992). This suggests that donors exposed to low support allocation information will react defensively, attribute the low support allocation to other factors and show no reduction in 63  prosocial identity (e.g., “I did my part, but the fundraiser was expensive to run.”). Research on action and intention to support also suggests low support allocation will not impact consumer self-views. This research finds that when actual support fails to live up to altruistic intentions, consumers judge their actions based on the impact they intended to make rather than actual performance (Kruger and Gilovich 2004). In one study, Kruger and Gilovich had participants submerge their hands in ice water (cold pressor task) to earn money for charity, or had others observe them in the task. Cold pressor participants first predicted the length of time they would keep their hands submerged ($0.50 was donated per minute submerged), and then attempted the task. The water in this task is uncomfortably cold, making it very difficult to meet the expected support level; in fact, all the 18 participants included in the analysis failed to meet their predicted impact level. Participants then rated how altruistic they believed themselves to be. In the observer conditions, participants rated how altruistic the cold pressor participant was after watching them attempt the task. Participants who attempted the cold pressor task judged themselves to be significantly more altruistic than observers’ ratings of them. Taken together, these findings suggest that consumers exposed to low support allocation information will not experience a reduction in prosocial identity, nor will it affect future support behavior.   On the other hand, there is research that suggests consumers would internalize the reduced allocation of their support, resulting in a lower prosocial identity and less subsequent prosocial behaviour. Research has shown that although morality is central to the individual self-concept, situational factors may positively or negatively impact prosocial identity and subsequent behaviour (Aquino et al. 2009; Bandura 1991; Blasi 1980). Grouzet et al. (2005) proposed that self-interested goals such as financial success or self-image are diametrically opposed to the goal 64  of being moral, and the activation of such goals reduces attention to one’s morality, leading to lower moral behaviour (Aquino and Reed 2002). Building on this conceptualization, Vohs, Mead and Goode (2006) found that exposing individuals to money made them behave less prosocially on a subsequent task. More recently, Aquino et al. (2009) examined how situational factors can impact subsequent prosocial behaviour even among individuals who hold morality as central to their self-concepts. In one study, participants received a moral (versus control) prime and then completed a series of investment choices that represented a commons dilemma; specifically, participants either donated money to individual or joint accounts that accumulated varying interest rates. The feedback on group choices was structured such that investing in the personal account was more (less) beneficial to the self (group). Results showed that situational factors can indeed impact subsequent identity and behaviour in both positive and negative directions. Participants with high (low) baseline moral centrality displayed lower (higher) prosocial behaviour after receiving a self-interest (moral) prime. Taken together, these results suggest that consumer prosocial identity is malleable and that internalization of allocation information may not be shielded by established defense mechanisms.    3.2.1 Why Supporters Bear the Brunt of Low Support Allocation   The extant literature, then, is unclear regarding the consequences that exposure to low support allocation information has on consumer prosocial identity and subsequent support behaviour. I propose a framework to reconcile this inconsistency and to outline when and why consumers internalize this threatening information, or defend against it. Further, I investigate 65  how the resulting prosocial identity affects subsequent support actions. In particular, my framework suggests that the perceived impact of one’s support behaviour plays a critical role in understanding the consequences of low support allocation information.   The Cost of Behaving Prosocially on Perceived Impact. Research in economics has theorized the desire to make an impact to be a central motivation for engaging in charitable acts (Duncan 2004). Although not empirically tested, a close examination of recent work suggests the critical role this factor has in understanding the consequences of prosocial support. Economic models of prosocial behavior suggest that individuals use previous prosocial actions to infer their moral or prosocial identities, and reference this identity level when making subsequent prosocial decisions (Bénabou and Tirole 2005; Bem 1972). For example, Gneezy et al. (2011) find that engaging in costly (versus costless) prosocial behavior leads to an increase in prosocial identity, and this higher view of one’s morality drives subsequent positive behavior. Olivola and Shafir (2011) build on this notion by showing that individuals donate more money to a cause because they view their support as more impactful; specifically, consumers believe that difficult prosocial support makes more of a difference than when support is easy to provide. Further, Aknin et al. (2013) find that support that makes a significant (versus unknown) impact leads to higher emotional benefits for the supporter – they feel better about what they have given. Perceived impact has also been shown to drive the identifiable victim effect (Cryder, Lowenstein and Scheines 2013), and explain goal gradient fundraising success whereby the closer an organization is to achieving its fundraising goal, the more likely individuals are to support (Cryder, Loewenstein and Seltman 2013). Finally, in an investigation of overhead aversion before supporting a charity, Gneezy, Keenan and Gneezy (2014) found that participants were 66  more willing to donate to a charity when the overhead expenses were covered by someone else (e.g., their personal impact was higher).  Taken together, I suggest that the observed increase in prosocial identity following meaningful support occurs because supporters perceive to have made a significant impact to those in need. This increased prosocial identity, in turn, has been shown to lead to higher subsequent support behaviours. What effect then, would low support allocation information have on consumers? By definition, low support allocation means that a lower proportion of consumer support is directly passed on to the cause recipients. Put another way, the impact that each individual donation makes to those in need is reduced because a lower proportion of the donation is passed on. If perceived impact is responsible for the increase in prosocial identity after providing meaningful support (Gneezy et al. 2011), then exposure to low support allocation information would reduce prosocial identity because one’s support was less impactful than originally believed. Moreover, given that higher prosocial identity has been shown to predict higher levels of subsequent support (Bénabou and Tirole 2005; Gneezy et al. 2011), I propose this reduced prosocial identity will in turn lead to lower subsequent prosocial support. More formally,  H1: Consumers who engage in an initial act of meaningful support for a cause will experience a reduction in prosocial identity after learning the impact of their support was low (versus high).  67  H2: Reduced prosocial identity will mediate the relationship between support allocation information and lower subsequent support for related and unrelated charitable organizations.     I test the proposed framework in four experimental studies. A pilot study first demonstrates that consumers show a reduced prosocial identity after exposure to low (versus high) support allocation information. Using a volunteer support task, Study 1 provides further support for the framework by bolstering prosocial identity, and replicating the reduced prosocial identity consequence only when prosocial identity is not bolstered. Study 2 tests the full conceptual framework and provides direct evidence that exposure to low support allocation information reduces prosocial identity by employing a repeated measures design. It also shows that the reduced prosocial identity leads to lower subsequent support behaviour for both the same, and unrelated causes. Finally, Study 3 explores the moderator of support success to reconcile my framework with previous research on intention to support. In all studies, I operationalize impact of support as the percentage of consumer support passed directly on to underlying case recipients.  3.3 Empirical Investigation  3.3.1 Pilot Study  I sought to provide a preliminary test of the framework using an online sample from Amazon’s Mechanical Turk (n = 101, payment = $1). The goal of this pilot study was to provide 68  initial evidence for the claim that exposure to low support allocation information following effortful support for a cause reduces prosocial identity among consumers (H1) and leads to lower subsequent support (H2). In this study, participants were asked to imagine a scenario in which they provided meaningful support for a charitable organization that sought to combat child poverty, a cause shown to be perceived as important and positively-viewed among this population (Kristofferson, White and Peloza 2014). This meaningful support was achieved through donation, volunteering and promotion of the cause to others in their social networks. Participants were then shown one of two versions of a press release written by a third party that outlined how the supported non-profit organization allocated its donations between cause recipients, overhead costs, promotional costs, and retained funds. To manipulate support allocation, the proportion of funds directed to cause was altered to be either very high (97.2%) or low (31.3%). Prosocial identity was then assessed using an established measure from previous research (Kruger and Gilovich 2004). If consumer prosocial identity bears the brunt of reduced impact, then participants should perceive their prosocial identity to be lower after exposure to the low (versus high) support allocation information (H1). Participants then indicated how likely they would be to support the organization again in the future.   3.3.1.1 Procedure  Participants were randomly assigned to one of two conditions (Allocation of Support: High, Low) in a between-participants design. At the beginning of the study, participants were asked to imagine that they worked hard to support a charitable organization that sought to eliminate poverty in the developing world. Specifically, participants imagined that they worked 69  very hard to support the charity by volunteering a large  portion of free time, donating a large portion of income each month and frequently promoting information about the charity to their social network. Upon completion of the scenario, participants were then exposed to the support allocation manipulation in the form of a press release written by an ostensibly certified third party that described the support and cost allocations of the charity (see Appendix A for press release stimuli). Participants in the high (low) allocation read told that “because of the way the charity is run, 97.2% (31.2%1) of volunteer efforts actually supported the impoverished people who benefit from the charity”. The press release also included a pie chart visually depicting the breakdown of funds between support to recipients, administration costs, fundraising costs and retained funds. Importantly, to ensure that participants did not attribute the support allocation manipulation to their own poor decision-making, participants were explicitly told that they only learned of the allocation after their most recent donation. This information was included to reinforce that the allocation of support decisions were deliberate and under the control of the charitable organization (e.g., Weiner 2001).    Dependent Variables. To measure prosocial identity, participants compared themselves to the average person on a scale of 1 – much lower than the average person to 7 – much higher than the average person on the following attributes: helpful, generous, self-sacrificing, lazy (reverse-coded, RC), giving, selfish (RC), ashamed (RC), kind, self-serving (RC), ethical, inspiring and caring (Kruger and Gilovich 2004).                                                  1 Thirty-one percent was chosen as the low allocation amount based on Canada Revenue Agency’s Income Tax Act, which states that a registered charity’s ratio of costs to revenue over fiscal period above 70% would raise concerns (http://www.cra-arc.gc.ca/chrts-gvng/chrts/plcy/cgd/fndrsng-eng.html). 70  Finally, to measure subsequent support behaviour, participants predicted their likelihood of continued support for the charity on a scale from 1 (definitely will not) to 7 (definitely will).   3.3.1.2 Results  Participants. All participants completed the task as instructed and are retained. Prosocial Identity. The prosocial identity items showed acceptable reliability (α = .89) and were averaged to form a prosocial identity index. A one-way ANOVA with allocation of support (High Allocation, Low Allocation) as the between-participants factor and prosocial identity as the dependent variable revealed a significant main effect of allocation of support. Supporting H1, participants exposed to the low support allocation information showed a significantly lower prosocial identity than participants exposed to the high allocation information (MLow Allocation = 4.93 SD = .86 vs. MHigh Allocation = 5.31, SD = .91; F(1, 99) = 4.44, p < .05). This suggests that reducing the impact that consumers make leads to an identity-based consequence: they perceive themselves to be less prosocial.   Subsequent Support. To test the effect of allocation of support on subsequent prosocial behaviour, the same one-way ANOVA was run using the likelihood to support the organization again as the dependent variable. The ANOVA yielded a significant main effect of the allocation of support factor. Participants exposed to the low support allocation information were significantly less likely to support the organization in the future than participants exposed to the high allocation information (MLow Allocation = 2.80 SD = 1.11 vs. MHigh Allocation = 4.28, SD = .86; F(1, 99) = 55.5, p < .001).  71    The Impact of Reduced Prosocial Identity on Subsequent Support. Finally, to examine the impact of reduced prosocial identity on subsequent support, I followed Preacher and Hayes (2007) bootstrapping procedure. As noted above, participants exposed to the low support allocation information showed a significantly lower prosocial identity than participants exposed to the high support allocation information (MLow Allocation = 4.93 SD = .86 vs. MHigh Allocation = 5.31, SD = .91; F(1, 99) = 4.44, p < .05). Next, controlling for the allocation of support factor, prosocial identity significantly predicted likelihood to support the organization in the future (b = .51, SE = .10, t = 5.04, p < .001). I tested the conditional indirect effect of allocation of support on subsequent prosocial behavior through reduced prosocial identity following Preacher and Hayes (2007) bootstrapping procedure of 5,000 samples with replacement (PROCESS Model 7). The conditional indirect effect was significant (b = .09, CI95: .01 to .20). Supporting H2, the low support allocation information reduced consumer prosocial identity, and led to a lower likelihood of subsequent support for the cause.  3.3.1.3 Discussion  The pilot study provides preliminary support for the hypothesis that a charitable organization’s low allocation of support reduces the prosocial identity of consumers. It also shows that this prosocial identity consequence leads to lower subsequent prosocial behavioural intention. However, this pilot study is limited in that participants did not actually provide their own meaningful support for an organization, they simply imagined doing so. I address this limitation in three subsequent laboratory studies. In the next three studies, all participants 72  actually engaged in providing meaningful support for a charitable organization. Study 1 focuses on the prosocial identity consequence portion of the framework by having participants engage in meaningful support for an organization.  3.3.2 Study 1: Internalizing Low Allocation  Study 1 had two goals. The primary goal was to replicate the results of the pilot study using actual charitable support behaviour and show that after providing meaningful support for a charitable organization, consumer prosocial identity is reduced upon learning that the allocation of one’s support was low as opposed to high (H1). In this study, I introduced a task to bolster the prosocial identity of half of the participants. If participants exposed to low allocation information are experiencing a low sense of self-perceived prosocial identity, than a task aimed at bolstering self-perceptions of prosocial identity in another domain should mitigate the observed effects (Heine, Proulx and Vohs 2006).  A second goal of study 1 is to rule out a potential alternative explanation to the findings: generalized distrust (Darke and Richie 2007; Pollay 1986). Darke and Richie (2007, see also Darke, Ashworth and Richie 2008) find that consumers who are deceived by an advertising claim feel betrayed, and this feeling of betrayal leads to generalized distrust of advertising in general, and damages trust in non-offending organizations. It is possible that the low support allocation information causes consumers to show a generalized distrust of non-profit organizations. Study 1 seeks to rule out this explanation.  73  Pre-test. I first sought to identify consumer expectations of acceptable charitable support allocation. Although the pilot study operationalized high support at 97.2%, it is possible that consumers understand that charitable organizations incur costs and accept a slightly lower allocation. In fact, a recent ranking of the top Canadian non-profit organizations showed that the most efficient organizations allocated between 76% and 93% to cause recipients (MoneySense 2014). As such, identifying and replicating the effects using a more realistic low allocation manipulation would add to the generalizability of the findings. One hundred and thirty-eight MTurk participants read one of four version of the allocation press release manipulation from the pilot study that allocated the following percentages of support to starving children: 31.2%, 51.2%, 71.2%, 91.2% (all other content remained the same). After reading the press release, participants indicated their opinions about the percentage of donations the charity passes directly on to those in need on a scale from 1 (Very Low) to 7 (Very high). A one-way ANOVA with allocation percentage perceptions as the dependent variable was significant (F(3, 137) = 35.6, p < .001), with participants perceiving the allocation to be significantly better with each increasing percentage (M31% = 2.76, M51% = 3.65, M71% = 5.00, M97% = 6.06, all ps < .05). As such, I selected 91.2% as the high allocation proportion of support.  3.3.2.1 Main Study Procedure  Main Study. Ninety-three University of British Columbia marketing undergraduates participated in this study in exchange for course credit and were randomly assigned to one of four conditions in a 2 (Allocation of Support: Low, High) x 2 (Prosocial Identity: Bolster, 74  Neutral) between-participants design. Participants completed the studies in groups of 20-30 and were seated at individual computer terminals.  Prosocial Identity Bolster. In order to test the proposition that consumers internalize low support allocation and reduce prosocial identity, participants first completed the bolstering task. Participants were told that their initial study was an essay-writing task to learn more about their personal experiences. Participants in the bolster condition were asked to identify a time in which they performed a charitable or prosocial act (defined as a behaviour that benefitted others at some cost to the participant, such as volunteering or donating blood) and to write a short essay about it (McQueen and Klein 2006). They were asked to provide details of the behaviour, how it made them feel and their overall experience. Participants in the neutral condition were asked to write about what they ate and drank for lunch yesterday.   Meaningful Support Task. Upon completion of the affirmation task, participants were told that the Marketing Department had been working with a non-profit organization that works to combat hunger in the developing world, and that in lieu of completing the next study, they would provide support to this organization by completing an intellectual task. I operationalized meaningful support for a cause as volunteering one’s time and intellectual skills using an organization called Freerice (www.freerice.com). Freerice is an actual non-profit website that is owned by and supports the United Nations World Food Programme. Consumers wishing to provide support visit the website and answer multiple-choice English vocabulary questions; for each correct answer, Freerice donates 10 grains of rice to the UN World Food Programme. To control for participant experience (i.e., exposure to different questions, correct/incorrect answer 75  feedback or advertising), I manually programmed the website layout and eighty-five vocabulary questions into a survey. Further, I modified the instructions from the Freerice website such that participants were told that for every question answered correctly, approximately 10 grains of rice were donated to the World Food Program. Participants were presented with the following instructions:  Here is how you show your support and help to combat world hunger:  A new charity, called Free Rice, is trying to alleviate hunger in the developing world by donating rice to the poor. It works like this: supporters visit their website and take part in a basic vocabulary quiz. For every question answered correctly, approximately 10 grains of rice are donated to the World Food Program.    After answering all 85 questions, participants were thanked for their hard work and effort, and were told that they had earned around 8,500 grains of rice for those in need before being ‘re-directed’ to the next study. Average volunteer time ranged from 12 to 25 minutes.  Manipulation Check. Immediately after completing the support task, participants were presented with an open-ended manipulation check asking approximately how many grains of rice they earned for those in need. To measure perceived effort, participants indicated how much effort they gave in support of the organization on a scale of 1 (very little effort) to 9 (a lot of effort).   76  Allocation of Support Manipulation. After completing the manipulation check, participants were presented with more information about the charitable organization they just supported via a third-party press release. This press release was in fact the same press release stimuli used in the pilot study, with the content adapted to the Freerice.com cause. The press release included detailed information regarding the effortful support the participant contributed, information regarding the company's actual follow-through on promised contribution to those in need, and general financial management. Participants were randomly presented with either the high (91.2%) or low (31.2%) support allocation manipulation.   Prosocial Identity. After reviewing the allocation of support manipulation, participant prosocial identity was assessed. Specifically, participants completed the following four statements that indicated how prosocial they perceived themselves to be right now using a scale of 1 (strongly disagree) to 7 (strongly agree): “I am the type of person that helps others even at a cost to myself”, “I find it hard to help less fortunate people when they refuse to help themselves (RC)”, “I will help those less fortunate when I have more time and money (RC)”, “When faced with a decision to benefit me or the greater good, I choose the greater good every time”.   Organizational Credibility and Deception. Participants were then asked to provide general organizational opinions to assess credibility and perceived deception of non-profit organizations. Participants responded to the following four bi-polar measures to assess views of for-profit and non-profit organizations using a scale from 1 (not credible/unreliable/untrustworthy/inexpert) to 9 (credible/reliable/trustworthy/expert).   77  Finally, participants completed demographic measures and an open-response hypothesis guess measure regarding the nature of the study.  3.3.2.2 Results  Participants. One participant correctly identified the purpose of the study, two participants failed the manipulation check and two participants did not complete the Freerice support task as instructed. As such, the analysis is conducted using 88 participants.  Perceived Effort Manipulation Check. As expected, participants perceived their freerice.com charitable support to be effortful in nature (one-sample t-test versus scale midpoint, MEffort = 6.17, t(87) = 5.38, p < .001). No main effects or interaction with allocation of support or affirmation variables emerged, and I note that this result is consistent if each condition is analyzed separately.  Prosocial Identity. A factor analysis confirmed that the four items loaded onto a single factor and were combined to form a prosocial identity index. I contrast coded both prosocial identity (Bolster = -1, Neutral = +1) and allocation of support (High Allocation = -1, Low Allocation = +1) independent variables and entered them in a 2 x 2 ANOVA to predict prosocial identity. The ANOVA yielded a marginal main effect of prime (MBolster = 4.04, SD = .67 vs. MNeutral = 3.81, SD = .60; F(1, 84) = 2.83, p < .10), but no effect of allocation (MLow Allocation = 3.88, SD = .64 vs. MHigh Allocation = 3.99, SD = .65; F(1, 84) = .77, p > .35). Importantly, the predicted prosocial identity x allocation interaction emerged (F(1, 84) = 4.11, p < .05). 78  Consistent with my framework and supporting H1, participants in the neutral condition exhibited a significantly lower prosocial identity when exposed to the low (versus high) allocation information (MLow Allocation = 3.63, SD = .59 vs. MHigh Allocation = 4.01, SD = .56; F(1, 84) = 4.12, p < .05); however, supporting H3, no differences in prosocial identity were present among bolstered participants (MLow Allocation = 4.12, SD = .61 vs. MHigh Allocation = 3.97, SD = .73; F(1, 84) = .68, p > .40, NS, see figure 6).   Ancillary Analysis: Organizational Credibility and Deception. To ensure that the allocation manipulation did not affect perceptions of organizational credibility or generate feelings of deception among participants, I conducted the same 2 x 2 ANOVA using opinions of both for-profit and non-profit organizations as dependent variables respectively. The four scale measures each loaded onto a single factor and were averaged to form for-profit and non-profit indices. The ANOVAs yielded no main effects or interaction on participant opinions of either for-profit (all ps > .17) or non-profit (all ps > .55) organizations, suggesting feelings of deceit did not result from the manipulations.   3.3.2.3 Discussion  Study 1 replicated the findings from the pilot study using a design in which participants provided their own meaningful support for a cause. Results provided support for the claim that consumers experience a decrease in prosocial identity upon learning that their meaningful support was less impactful than originally believed (H1). After supporting a non-profit organization through an intellectually challenging volunteer task, participants showed lower 79  prosocial identity when the allocation of their support to underlying recipients was low as opposed to high. However, this prosocial identity consequence occurred only when prosocial identity was not bolstered. This result supports the claim that low support allocation directly threatens prosocial identity. It also helps identify a potential tool organizations can use when presenting financial information to consumers: prosocial identity bolstering. Results showed that participants with bolstered prosocial identities did not show a reduction in prosocial identity upon exposure to the low allocation stimuli. As such, if non-profit organizations could bolster their supporters prior to exposure, the prosocial identity consequence may be reduced.  Study 1 also helped rule out an alternative explanation: generalized distrust (Darke and Richie 2007; Pollay 1986). As noted, Darke and Richie (2007, see also Darke, Ashworth and Richie 2008) find that consumers who are deceived by an advertising claim feel betrayed, and this feeling of betrayal leads to generalized distrust of advertising in general and damages trust in non-offending for-profit organizations. However, results from the ancillary analysis show that the allocation of support manipulation did not impact opinions of non-profit organizations in general, suggesting the prosocial identity consequence differs from that of deceptive advertising. Although somewhat surprising, this result aligns with my framework. Participants engaged in an effortful support task in order to make an impact in the lives of those less fortunate. In both high and low allocation conditions, the non-profit organization allocated a large amount directly to the cause recipients. Thus, in both conditions, an impact was made. However, my framework argues that prosocial identity is threatened because the impact made is reduced upon exposure to low support allocation information. As such, the non-profit organization is still making a difference in the lives of the less fortunate, thus mitigating feelings of distrust. For-profit organizations, on 80  the other hand, may foster feelings of generalized distrust because their primary purpose is financial (versus social) in nature.    3.3.3 Study 2: Consequences of Reduced Prosocial Identity   Study 2 has three primary goals. First, I wished to replicate the findings from study 1 by providing evidence that a charitable organization’s low allocation of funds directly reduces consumer prosocial identity after providing meaningful support (H1). I achieve this goal by employing a repeated-measures design, and measuring and comparing consumer prosocial identity at baseline and post-support time intervals. I also sought to show this outcome using a more established measure of prosocial identity. As such, in study 2 I operationalized consumer prosocial identity using the internalization dimension of Aquino and Reed (2002)’s moral identity scale (see Appendix A). The moral identity scale measures the degree to which a consumer’s self-concept centres on moral traits and is comprised of two distinct dimensions: internalization and symbolization. Internalization measures the degree to which moral traits are central to a one’s self-concept, whereas symbolization refers to the degree to which one’s observable actions imply moral traits to others. This operationalization of prosocial identity has been used successfully in previous prosocial behaviour research (i.e., Aquino et al. 2009; Kouchaki 2011). My framework predicts that learning one’s meaningful support is less impactful 81  reduces prosocial identity. As such, although I measure symbolization, low allocation information should only reduce levels on the internalization dimension. More formally,  H3: Low allocation information will reduce perceived moral identity on the internalization but not symbolization dimension.  Second, I wanted to examine whether perception of choice affected my framework. Although previous prosocial behavior research has shown consistent results in voluntary versus forced-support paradigms (Andreoni 1990, Dunn et al. 2008; Gneezy et al. 2011), I felt it necessary to examine whether choice interacts with allocation of support. It is possible that individuals forced to provide support, and then subsequently exposed to the low allocation stimuli, may attribute their negative reactions more externally (i.e., to being forced to support), and thus not experience a reduction in prosocial identity. However, if no interactions emerge, it would add significant robustness to my model. To rule out this explanation, participants in the choice condition were given the option to choose (yes, no) whether to engage in the volunteer task in this study.   Finally, the third goal of study 2 is to examine the effect that this reduced prosocial identity has on subsequent support behaviours; specifically, whether the reduced identity leads to lower generalized support behaviours (H2). If this is the case, it would imply that the low allocation decisions of one organization could have spillover effects, and reduce future support for other unrelated organizations. I utilize a behavioral dependent measure that has costly implications for participants to achieve this goal. Note that while this prediction of reduced 82  generalized support aligns with the findings in deceptive advertising (Darke and Richie 2007), I argue it results through a different process: reduced prosocial identity.  Data for study 2 was collected in two phases. Phase one took place at the beginning of the semester, and consisted of a survey capturing baseline participant moral identity scores (among other scales). Phase two occurred during the final week of classes of the same semester. One-hundred forty-one English-speaking undergraduates from the same participant pool as the pre-survey participated in the study in exchange for course credit and were randomly assigned to one of four conditions in a 2 (Support Task: Choice, No-Choice) x 2 (Allocation of Support: High, Low) x 2 (Prosocial Identity: Pre-Support, Post-Support) mixed-design with prosocial identity as a within-participants factor. Participants completed the study at individual computer terminals in groups of 20–30 per session.  3.3.3.1 Procedure  Upon arrival to the lab, participants were provided with the same cover story from study 1, and provided meaningful support using the same Freerice volunteer task.   Voluntary/Forced-Choice Manipulation. After reading about the meaningful support task, participants in the voluntary choice condition were given the choice manipulation: a yes/no decision regarding their willingness to support. Participants answering yes then received the identical instructions from study 1. Participants who answered no to the request were given the 83  following instructions, but were still directed to, and performed, the identical volunteer Freerice support task:  Your alternate task works like this: you will be redirected to a website through this survey and will take part in a basic vocabulary quiz. For every question answered correctly, approximately 10 grains of rice are donated to the World Food Program.  Participants in the no-choice conditions were given the instructions from study 1 and completed the identical 85-question task.   Manipulation Check. As in study 1, participants were then presented with an open-ended manipulation check asking approximately how many grains of rice they earned for those in need.   Allocation of Support Manipulation. After completing the manipulation check, participants were exposed to the same press release stimuli from study 1, showing that allocation of support was either high (91.2%) or low (31.2%).   Prosocial Identity. After reviewing the allocation of support press release manipulation, participants were presented with the 10-item moral identity scale (i.e., Aquino and Reed 2002), the internalization subscale serving as the prosocial identity measure (see Appendix for all scale measures).   84  Mood and Deception. To ensure that our manipulation was not impacting participant mood, participants completed the following emotion measures on a scale of 1 (low) to 7 (high): Positive: happy, inspired, giddy, determined, honest; Negative: angry, depressed, inadequate, sad, upset. In addition, measures of deception and betrayal were added to more directly rule this explanation out.   Dependent Variables. After completing demographic information and a neutral picture-rating task for approximately ten minutes, participants were presented with the dependent variable: a support request from a different charitable organization. Participants were told that a newly formed local non-profit organization needed help generating feedback on their marketing materials, and was asking marketing students to volunteer to stay immediately after this session to complete an extra survey (Essay 1). The study was conducted during the last week of classes when students were extremely busy with final reports and presentations; as such, agreeing to the request carried significant costs to participants. After answering the main dependent measure, I also asked participants to recall their Freerice support task from earlier in the session, and indicate the likelihood that they would use Freerice.com again in the future to provide support (1 = very unlikely, 7 = very likely).  3.3.3.2 Results  Participants. Five participants failed the manipulation check by more than 4,000 grains of rice and one participant skipped the meaningful support task; as such, all analyses are 85  conducted with 135 participants. Results and contrasts are consistent if all participants are included.    Prosocial Identity. The 10-item moral identity scale was factor analyzed for both pre- and post-support time periods using maximum likelihood with varimax rotation. Following previous research, the two-factor internalization and symbolization dimensions emerged for both time periods. Each dimension showed acceptable reliability (αPre-Int = .75, αPre-Sym = .78, αPost-Int = .79, αPost-Sym = .78) and the items were averaged.   To test predictions regarding prosocial identity (H1), I performed a 2 (Meaningful Support: Choice, No-Choice) x 2 (Allocation of Support: High, Low) x 2 (Prosocial Identity: Pre-Support, Post-Support) mixed-model ANOVA with prosocial identity (moral identity internalization) serving as a within-participants factor. Interestingly, the choice factor had no effect. Neither of the two-way choice x prosocial identity or choice x allocation interactions, nor the three-way prosocial identity x choice x allocation interaction approached significance (all ps > .45). This suggests the forced-support paradigm is appropriate to examine the effects.   Given these results, I performed a 2 (Allocation of Support: High, Low) x 2 (Prosocial Identity: Pre-Support, Post-Support) mixed-model ANOVA to examine my predictions. This analysis yielded a main effect of prosocial identity timing (MPre-Support = 6.17, SD = .86 vs. MPost-Support = 5.91, SD = .91; F(1, 133) = 18.4, p < .001), but no effect of allocation (MHigh Allocation = 5.94, SD = .99 vs. MLow Allocation = 6.14, SD = .75; F(1, 133) = 1.94, p > .15) variables. Importantly, the predicted prosocial identity x allocation interaction emerged (F(1, 133) = 4.54, p 86  < .052). Consistent with my framework, participants exposed to the low allocation information experienced a significant decrease in prosocial identity (MPre-Support = 6.33, SD = .68 vs. MPost-Support = 5.94, SD = .82, MDiff = .39, SD = .74; F(1, 66) = 18.3, p < .001); however, no differences emerge among participants exposed to the high allocation information (MPre-Support = 6.01, SD = .99 vs. MPost-Support = 5.88, SD = 1.0, MDiff = .11, SD = .66; F(1, 67) = 2.65, p > .10, see figure 7).   I also tested whether differences emerged in the symbolization dimension of participant moral identity scores. The same mixed-model ANOVA using symbolization scores as the dependent variable yield no significant effects of symbolization timing (MPre-Support = 4.12, SD = 1.03 vs. MPost-Support = 4.11, SD = 1.03; F(1, 127) = .02, p > .85, NS), allocation (MHigh Allocation = 4.03, SD = 1.17 vs. MLow Allocation = 4.20, SD = .99; F(1, 127) = .92, p > .30, NS) variables, nor was the two-way interaction significant (F(1, 127) = 2.16, p > .14, NS). This lack of effect on the symbolization dimension is very interesting and lends support for my framework. The result suggests that the consequence of exposure to low support allocation information is that it causes participants to lower the internal perception of their prosocial identities (i.e., their self-view), but not the identities they show to others. As such, H1 is further supported such that prosocial identity is reduced upon learning their support is less impactful than originally believed.   Dependent Variables. In order to examine the downstream effects of reduced prosocial identity on subsequent support behaviour (H2), I examined participant agreement to the unrelated charity’s volunteer request. I conducted a hierarchical logistic regression with                                                 2 The two-way prosocial identity x allocation interaction remains significant if choice is included in the model. 87  agreement to the volunteer request as the dependent variable (coded 1 = yes, 0 = no). I entered allocation of support and pre-support prosocial identity in step one, followed by post-support prosocial identity in step two to predict agreement with the volunteer request. As predicted, controlling for allocation of support and pre-support prosocial identity, post-support prosocial identity significantly predicted the likelihood to agree to help the unrelated charity (b = .72, Wald χ2(1) = 3.88, p < .05). If the same model is examined using likelihood of supporting the same charity (Freerice.com) as the dependent measure, similar results emerge (b = .38, t = 1.87, p = .06). These results support the notion that, upon learning of the lower impact one’s meaningful support made, reduced prosocial identity leads not only to less subsequent support for the supported cause, but also reduces future support behaviour in general.   Ancillary Analysis: Mood and Deception. The allocation manipulation had no effect on participant mood. One-way ANOVAs using allocation as the between-participants factor found no differences in either positive (MLow Allocation = 3.99, SD = .64 vs. MHigh Allocation = 3.89, SD = .71; F(1, 133) = .81, p > .35, NS) or negative (MLow Allocation = 2.73, SD = 1.21 vs. MHigh Allocation = 2.74, SD = 1.36; F(1, 133) = .01, p > .90, NS) mood. Analyzing each mood measure individually also yields non-significant results (all ps > .15, NS). Importantly, there was no impact on participant perceptions of feeling deceived (MLow Allocation = 3.03, SD = 1.45 vs. MHigh Allocation = 2.66, SD = 1.49; F(1, 133) = 2.12, p > .14, NS) or betrayed (MLow Allocation = 2.52, SD = 1.45 vs. MHigh Allocation = 2.32, SD = 1.30; F(1, 133) = .71, p > .40, NS) by the non-profit organization. I also note that if only the low allocation conditions are analyzed, scores on deception and betrayed fall significantly below the scale midpoint (Deceived: t(66) = -5.49, p < .001; Betrayed: 88  t(66) = -8.40, p < .001). Finally, neither feeling deceived nor betrayed predicted subsequent prosocial support (ps > .65, NS).    3.3.3.3 Discussion  Study 2 extends the findings from study 1 by showing that consumers experience a direct reduction in prosocial identity upon learning that their meaningful support was less impactful than originally believed (H1). Using a well-established measure of prosocial identity, I show this direct decrease within consumers by implementing a repeated measures design. Study 2 also extends the results of study 1 by showing that this reduced prosocial identity not only leads to reduced support for the previously supported charitable organization, but also reduces subsequent prosocial behavior in general (H2). Although previous research shows that individuals give themselves credit for their intentions to support even if support falls short of expectation, these results suggest that even the best of intentions cannot compensate for discovering one’s efforts were less impactful than originally perceived.   Finally, study 2 more directly rules out the alternative explanation of deception. The results suggest that learning of the low allocation details does not lead to feelings of deceit or distrust for non-profit organizations in general (Darke and Richie 2007), nor does it predict lower generalized prosocial support behaviour. Rather, consumers internalize the reduced impact of their initial meaningful support act, leading to lower generalized support behaviour.  89  3.3.4 Study 3: When Intentions Matter: The Ice Bucket Challenge  The goal of study 3 is to examine a novel theoretical moderator to reconcile the seemingly inconsistent predictions between my framework and that of Kruger and Gilovich (2004) on intention to support: success of initial support. Previous research finds that in order to assess prosocial identity after performing a support task for a charity, consumers refer to and award credit for their intention to support a cause in addition to their behaviour (Kruger and Gilovich 2004). In its present form, my framework appears to be in conflict with this literature. However, a closer examination of Kruger and Gilovich (2004)’s procedure and findings suggests that both frameworks may occur under certain conditions.   In this earlier work, female undergraduates completed the cold pressor task - an experimental set-up in which participants submerge their arms in ice water. Participants were told that for every minute they kept their arm submerged in the ice water, fifty cents would be donated to a charity of their choosing. Next, participants predicted how long they would keep their arm submerged. Finally participants attempted the task and subsequently rated their performance, prosocial identity (termed altruism) and intentions to support. All cold pressor participants were video recorded and these videos were shown to a separate group of observers who rated each participant using the same items. The results showed that cold pressor participants rated themselves to be significantly more altruistic than the observers rated them to be. The observers judged participant perceptions using their actual behaviour rather than intentions to support. This line of research runs contrary to my framework and findings from the first three studies. In fact, it predicts that consumers would not show the prosocial identity 90  reduction upon learning their actions were less impactful because they would assign themselves credit for their intention to support in addition to their support behaviour. How might the two frameworks be reconciled?  Closer inspection of the experimental paradigm employed by Kruger and Gilovich (2004) and subsequent results suggests that the success of initial meaningful support may play a moderating role. Participants included in the intentions analysis failed to meet their predicted support goal (length of time in the water), thus making salient the difference between actions and intentions. In fact, the authors note that nine participants met or exceeded their predicted support level, and if included in the analysis, the differences in perceived altruism between self and observer become non-significant (study 5, footnote 5, 338). This suggests a key difference from my model. I suggest that the internalization of prosocial identity occurs after engaging, and succeeding in meaningful support because consumers believe they have made an impact to those in need. Failing to provide the desired support, then, should lead to granting credit for one’s intentions regardless of allocation of support, and thus reconcile the two frameworks. More formally,  H4a: Consumers who attempt and succeed in providing meaningful support for a cause will show a lower prosocial identity when support allocation is low as opposed to high.  91  H4b: Consumers who attempt but fail to provide meaningful support for a cause will show no differences in prosocial identity when support allocation is low as opposed to high.  Given the goal to replicate, yet qualify previous research, I chose the same experimental paradigm and prosocial identity dependent measures as Kruger and Gilovich (2004); however I made minor modifications to successfully investigate my proposed moderator and take advantage of a recent successful support campaign: The ALS Ice Bucket Challenge.    3.3.4.1 Procedure  One hundred thirty-one female undergraduates from the University of British Columbia participated in this study in exchange for course credit and a chance to earn a $10 donation to a local charity, and were randomly assigned to one of five conditions in a 2 (Allocation: Low, High) x 2 (Support Success: Success, Failure) + 1 (Observer) between-participants design. Participants in the observer conditions viewed videos of participants in the other four conditions, but were not provided with any support allocation information. Following Kruger and Gilovich (2004), only female participants were recruited in order to control for differences in pain tolerance. Participants were run individually at thirty-minute intervals.   Upon arrival to the lab, participants were greeted by the experimenter and told that they would be completing multiple studies, the first of which was an opportunity to earn $10 for a local charity. The experimenter explained to participants that in addition to course credit, they 92  will be helping earn money for a charity and their support behaviour will determine donations from the Marketing Department. Participants were told that they must submerge their arm in cold water for ten minutes, and if successful, a $10 donation will then be made on their behalf to the charity. Participants were told that they could remove their arm at any time and still receive course credit, but not the $10 donation to the cause. The experimenter explained that although the task was harmless, it was painful as the water is between 0 and 10 degrees Celcius and is very cold to the touch. I selected ten minutes as the cut-off point to both make salient success or failure in providing support, and to maximize the effort required to support the organization3. Participants were then given descriptions of the two positively-viewed local charities from essay 1 and chose one to support. Finally, participants provided verbal and written consent to participate in the experiment, indicated whether they intended to keep their arm in the water for the full ten minutes, and provided consent to have the experiment video-recorded. I note that two participants did not intend to keep their arms submerged and are not included in the experiment.    Support Success Manipulation. Participants were brought into a different room to complete the ice bucket challenge, and success or failure was recorded by the experimenter. To manipulate the chance for success, I varied the temperature of the water. In the failure condition, the water temperature was set between 0o – 4o Celsius (32o – 35o Fahrenheit) as per previous research. However, in the success condition, the range of the ice water temperature was increased to be up to 12o Celsius (54o Fahrenheit). Although still cold and uncomfortable, the potential to complete the task was higher.                                                 3 Previous research has shown that the average amount of time undergraduate students kept their hand in the water was approximately one minute when they judged the discomfort as “too much to continue”. (Mitchell et al. 2004) 93    Support Allocation Manipulation. Upon completion of the ice bucket challenge, participants were given a towel, moved to individual rooms and presented with a questionnaire. Before beginning the questionnaire, however, participants were presented with the support allocation manipulation from previous studies (adapted to the charity they chose to support) under the guise of learning more about the charity they chose to support (Low Allocation = 31.2%, High Allocation = 91.2%).    Dependent Variables. To reconcile my framework with previous research, I presented participants with the measures used in Kruger and Gilovich (2004).  Participants were asked to compare themselves to other study participants using a scale of 0 (less than everyone else) to 100 (more than everyone else). First, participants assessed their actual performance in the ice bucket challenge by identifying “the time you kept your arm submerged”. Next, prosocial identity was measured using the following items: helpful, generous, self-sacrificing, lazy (reverse-coded, RC), giving, selfish (RC), ashamed (RC), kind, ethical, inspiring, caring and altruistic. Finally, intention to support was measured with the following two measures: “How much you wanted to help your selected charity”, “Your willingness to perform the cold water task for your charity”.   Subsequent Support. Upon completion of cover story questions, participants were asked to identify the likelihood that they would support the charity again using a scale from 1 (definitely will not) to 7 (definitely will).   94  Observers. After receiving the same written and verbal instructions as ice bucket participants, participants in the observer condition were told that instead of completing the support task themselves, they would watch videos of other participants. Each ice bucket participant was yoked to two observers, and each observer watched five participant videos. After each video, the observer rated the participant using the same performance, prosocial identity and intention measures described above. The presentation order of participant videos was randomized to account for any order effects in ratings.   3.3.4.2 Results  Participants. Five participants were not able to complete the ice bucket challenge as instructed (e.g., building fire alarm went off, experimenter set-up error). As such the analysis is conducted with 124 participants.  Prosocial Identity. I first wanted to test the prediction that support success moderates the relationship between allocation of support and reduced prosocial identity (H4a, H4b). Following Kruger and Gilovich (2004), the prosocial identity items were averaged to form a prosocial identity index (α = .89). A 2 (Allocation: Low, High) x 2 (Support Success: Success, Failure) ANOVA with prosocial identity serving as the dependent variable revealed no effect of allocation of support (MLow Allocation = 58.9, SD = 8.84 vs. MHigh Allocation = 63.3, SD = 14.5; F(1, 79) = 2.12, p = .15, NS), but a marginal main effect of support success (MSuccess = 63.1, SD = 10.9 vs. MFailure = 58.2, SD = 13.5; F(1, 79) = 3.80, p = .06) independent variables. Participants who succeeded in providing meaningful support showed a marginally higher prosocial identity 95  than participants who failed to provide meaningful support. However, this result was qualified by the predicted 2-way interaction (F(1, 79) = 4.62, p < .05). Replicating the results from studies 1 and 2 and supporting H4a, among participants who succeeded in providing meaningful support, participants exposed to the low allocation information reported a significantly lower prosocial identity then participants exposed to the high allocation information (MLow Allocation = 58.7, SD = 8.23 vs. MHigh Allocation = 68.1, SD = 11.6; F(1, 79) = 7.98, p < .01). However, among participants who failed to provide meaningful support, the allocation of support information had no impact on reported prosocial identity (MLow Allocation = 59.3, SD = 10.0 vs. MHigh Allocation = 57.4, SD = 15.9; F(1, 79) = .20, p > .65, NS, see figure 8).   Next, I sought to reconcile my results with previous research. For ease of comparison, statistical results and significance levels are presented in tables 3a and 3b.   First, I replicate the results of Kruger and Gilovich (2004) among participants that failed to provide meaningful support (see table 3a). As was the case with Kruger and Gilovich (2004)’s participants, even though ice bucket participants rated their actual performance in the support task as worse than other participants, they still rated their prosocial identity and intention to support the charity as significantly higher than other participants. As predicted, the credit that participants gave themselves for their intentions to support resulted in both high and low of allocation of support conditions.   Second, supportive of my framework, I qualify the results of Kruger and Gilovich (2004) among participants that succeeded in providing meaningful support to the charity (see table 3a). 96  As would be expected, participants who kept their arms submerged for the full ten minutes and read the high allocation information rated their overall performance in the task, prosocial identities and intentions to support the charity as higher than other participants. However, successful participants exposed to the low allocation information reacted much differently than high allocation participants. Even though they acknowledged that their performance in the support task was higher than that of other participants, they did not rate their prosocial identities as significantly higher than that of other participants. Thus, participants exposed to the low allocation information internalized their reduced impact and viewed themselves as less prosocial, even though their intentions to support the charity were high.  Finally, including the observer condition provides further support for my framework (see table 3b). Replicating Kruger and Gilovich (2004), observers judged participants solely on their overt support behaviour regardless of allocation of support condition4. When ice bucket participants failed to provide meaningful support, observers did not grant them the credit for their intentions they allocated to themselves. Observers rated participants as significantly less prosocial than they judged themselves to be. Conversely, when participants succeeded in providing support, observers judged them to be more prosocial–again regardless of allocation of support condition. Most interestingly, I note the contrast between self- and observer-ratings among low allocation participants. Supportive of the framework, results show that participants in                                                 4 I chose not to tell the observers about allocation of support condition as this information would not be readily available when viewing charitable support in more natural settings. 97  the success-low allocation condition viewed themselves to be less prosocial then observers judged them to be5.  Subsequent Support. To examine the influence of allocation and support success on subsequent support behaviour, I conducted a 2 (Allocation: Low, High) x 2 (Support Success: Success, Failure) ANOVA with likelihood to provide subsequent support as the dependent variable. Results revealed no effect of either allocation of support (MLow Allocation = 4.90, SD = 1.48 vs. MHigh Allocation = 5.41, SD = 1.53; F(1, 78) = 1.40, p > .24, NS) or support success independent variables (MSuccess = 5.15, SD = 1.49 vs. MFailure = 5.18, SD = 1.59; F(1, 78) = 0.00, p > .99, NS). However, the interaction was significant (F(1, 78) = 4.08, p < .05). Among participants who successfully supported the charity, participants exposed to the high allocation information were significantly more likely to subsequently support the organization than participants exposed to the low allocation information (MLow Allocation = 4.65, SD = 1.29 vs. MHigh Allocation = 5.73, SD = 1.52; F(1, 78) = 6.21, p < .05). However, allocation of support had no effect among participants who failed in the ice bucket challenge (MLow Allocation = 5.33, SD = 1.72 vs. MHigh Allocation = 5.05, SD = 1.51; F(1, 78) = .30, p > .55, see figure 9).  Mediation. To test whether reduced prosocial identity mediated the relationship between allocation of support and reduced subsequent support, I followed Preacher and Hayes (2007) bootstrapping procedure with 5,000 resamples (PROCESS Model 7). As discussed, successful ice bucket participants showed a significantly lower prosocial identity after exposure to the low                                                 5 I note that I do not replicate Kruger and Gilovich (2004) in the failure conditions regarding intention to support. Although not critical to my analysis, I believe this difference resulted from observer participants viewing five (versus one) participant videos.  98  (versus high) allocation information. Further, successful ice bucket participants also showed a lower likelihood to support the organization again after exposure to the low (versus high) allocation. No differences in either prosocial identity or subsequent support emerged as a function of allocation condition for ice bucket participant who failed the challenge. My framework would predict that the conditional indirect effect would be significant among participants who succeeded in providing support, but not for those who failed. Surprisingly, this result did not emerge (b = .06, CI95: -.09, .24). Further inspection of the data, however, shed additional light into the underlying process.   A 2 (Allocation: Low, High) x 2 (Support Success: Success, Failure) ANOVA with intention to support as the dependent variable yielded a marginally significant interaction (F(1, 79) = 3.60, p = .06), and displayed a similar pattern and significant contrast as prosocial identity results. After successfully supporting the charity, participants exposed to the low allocation information showed significantly lower intentions to support the charity than participants exposed to the high allocation information (MLow Allocation = 63.0, SD = 15.6 vs. MHigh Allocation = 76.3, SD = 19.9; F(1, 79) = 5.77, p < .05). I then conducted a regression and found that, controlling for independent variables allocation of support, support success and their interaction, prosocial identity significantly predicted intention to subsequently support the organization (b = .69, t = 4.00, p < .001, significant conditional indirect effect in success condition: b = 3.02, CI95 = 1.02, 6.40). Finally, controlling for independent variables allocation of support, support success, their interaction and prosocial identity, intention to support significantly predicted subsequent support for the charity (b = .02, t = 2.46, p < .05). This result sheds more light into the process underlying reduced subsequent support. The reduction in prosocial identity prompted 99  participants to alter their perceived original intention to help the organization. Although still high, participants lowered their initial perceived intention to support the organization, and this reduced intention led to a lower likelihood to subsequently support further. Although not predicted a priori, this result shows that ego defense mechanisms established in previous research are in fact being used (e.g. Carver, DeGregor and Gillis 1980), but operate in a novel manner in this paradigm.  3.3.4.3 Discussion  Adapting the classic cold pressor experimental paradigm, study 3 introduced a theoretical moderator to reconcile my framework with previous research on intentions: success of initial support. Replicating Kruger and Gilovich (2004), participants who failed to provide support gave themselves credit for their intentions, perceiving their prosocial identities to be higher than observers judged them to be. However, supporting my framework, participants who succeeded in providing meaningful support judged their prosocial identities to be lower upon learning that their impact was low as opposed to high. In fact, successful, yet not impactful participants actually perceived themselves to be less prosocial than observers judged them to be.   Study 3 also provided deeper insight into the spillover effects into generalized support behaviour. Results show that upon internalizing reduced impact of support, consumers adjust their original intention to support attitudes downward. This adjusted original intention, in turn, reduces consumer likelihood to help further. As a final note, as a result of participant support behaviour, the marketing department made a $500 donation to United Way. 100   3.4 General Discussion  Using multiple measures of prosocial identity and variations of initial and subsequent support behaviours, essay 2 examines the increasingly prevalent issue of charitable support allocation information on consumer prosocial identity and subsequent support. In four studies, this research shows that upon learning that one’s meaningful support is less impactful than originally perceived to be, consumers experience a reduction in prosocial identity. Importantly, results show that this identity-based consequence leads consumers to behave less prosocially in the future not only towards the previously supported cause, but also to unrelated charitable organizations. Importantly, I reconcile my framework with previous research on intention to support by introducing a novel theoretical moderator: success of initial support.   In a pilot study, I demonstrate that low support allocation information reduces consumer prosocial identity. This reduced prosocial identity led to significantly lower subsequent support for the organization. In study 1, I provide a more direct test of the conceptual framework by replicating the identity-based consequence from the pilot study using participants’ own support behaviour. Importantly, study 1 shows that this prosocial identity consequence emerges unless consumer prosocial identity is bolstered prior to exposure to this information. Participants primed with prosocial identity did not show a lower prosocial identity, suggesting bolstering may be a tool organizations can use to buffer consumers when presenting support allocation information. Study 2 tests the full conceptual framework and provides additional support for the proposed process by showing a direct within-subjects reduction in prosocial identity after 101  exposure to the low support allocation information. Study 2 also finds that this reduced prosocial identity reduced generalized future support behaviour for unrelated organizations, in addition to the previously supported one. Finally, study 3 reconciled my framework with the findings of intention to support researchers by introducing a novel moderator: success of initial support. It also shed additional light into the process driving reduced future support by showing that threatened consumers alter the belief in their original support intentions, leading to lower willingness to support later.   This essay makes a number of important contributions to the literature. First, I address an important social issue facing consumers and non-profit organizations: the downstream consequences of charitable support allocation. I find that upon learning that a lower allocation of one’s support for a cause goes to the underlying recipients, consumers experience a decrease in prosocial identity, and behave less prosocially in the future. I propose a conceptual framework that identifies why this consequence occurs, and identify the success of initial support as a key moderator in predicting when this threat will occur. This framework suggests that non-profit managerial decisions can have unintended and disastrous consequences to their most important target market: recurring supporters. These decisions, having the goal of increasing organizational competitiveness and maximizing the total support provided to cause recipients, can actually lead current supporters to alter their self-views and give less subsequent support not only to the originally supported cause, but also to other organizations in general.   3.4.1 Theoretical Contributions  102  This essay makes an important theoretical contribution to current understanding of economic models of prosocial behaviour by identifying the novel moderator of perceived impact. Previous research shows that performing a costly prosocial behaviour raises one’s prosocial identity and leads to increased consistent prosocial behaviour (Gneezy et al. 2011). By connecting this work to more recent empirical findings, I propose and find that this increased prosocial identity is driven by the benefits incurred from higher perceived impact of support. I operationalize perceived impact using the allocation of charitable support to underlying beneficiaries, and find that reducing impact eliminates the documented benefits of costly support on both consumer prosocial identity and subsequent support behaviour. To my knowledge, this is the first empirical investigation showing the consequences of perceived impact on identity and subsequent support. As such, I answer the call of recent research to investigate the effect perceived impact can have on consumer prosocial behaviour (Gneezy et al. 2014). By identifying this factor and novel unintended consequence, I contribute to theoretical understanding of the factors that lead to recurring prosocial behaviour following initial support. More importantly, I identify a unique identity-based consequence that leads to lower prosocial behaviour not only towards the previously supported charity, but also other unrelated organizations.   This essay also contributes to research on altruistic intentions to behave prosocially by providing a more nuanced understanding of when and why consumers assign credit for their intentions to support. Previous research found that consumers judge their own prosocial identities by assigning credit for both behavioural outcomes and intentions, but judge others only by their behaviours (Kruger and Gilovich 2004). I propose and find that this credit assignment occurs only when consumers fail to achieve their initial support goals. By clarifying that this 103  additional credit is assigned only when consumers fail to provide meaningful support, thus questioning the impact of their behaviour, I present a clearer understanding of this psychological mechanism. Further, I identity an additional condition under which consumers judge their prosocial identities using only their behavioural outcomes.   3.4.2 Managerial Implications   This essay also makes a number of important substantive contributions to charitable and other non-profit organizations by uncovering the unintended consumer consequences of low support allocation information has on subsequent support. Non-profit organizations operate in a very competitive landscape. In addition to competing with other non-profits for consumer donations, non-profit organizations also compete with for-profit firms for consumer discretionary income. This already low proportion of discretionary income allocated to donation is further reduced during economic downturns. In order to fulfill their mandates and maximize total support provided to cause recipients, non-profit organizations must adapt to this landscape by using a proportion of consumer support to implement integrated marketing campaigns and recruit the top executives and employees. My results show that, while these decisions may have benefits, there can be dangerous unintended consequences to the most important target market of current supporters and may actually be detrimental in the long-run. Non-profit decision-makers should factor in this potential risk when designing and executing these decisions.   104  In addition to suggesting caution, I also offer suggestions to help mitigate consumer consequences. Specifically, organizations can execute communications to highlight the impact current supporters have made through their previous support. Given that recurring donors donate significantly more money per year than one-time donors, ensuring that current supporters perceive their support to be impactful can help ensure long-term success. While Gneezy et al. (2014) show the effect that overhead information may have on deciding to support initially, this essay identifies how this information affects supporters after support is given. The results suggest charitable organizations could design communications or materials to highlight the impact their support has made (e.g., what programs result from supporting a campaign) in addition to outlining how their support is used.    Further, the results of study 1 may help to guide presentation of fiscal content communicated to consumers. In addition to communicating the direct impact consumer support has made, my results suggest that tailoring content to bolster consumer prosocial identity prior to presenting financial and allocation information may buffer the negative consequences to consumer prosocial identity. Although the experimental manipulation employed in this investigation involved a written bolstering task–an approach unavailable in one-way non-profit communications–research has shown that bolstering can occur in more subtle ways, such as nonconscious priming of affirming words (Sherman et al. 2009). Designing communication material to bolster prosocial identity through short written descriptions, or even pictorial content could help to mitigate the reduction in prosocial identity.   105  Finally, this essay suggests a potentially dangerous result for non-profit organizations in general from the increased media attention this issue is receiving. I suggest that consumers exposed to low allocation information become less prosocial in general. Given the important role that non-profits play in our society, losing donors is not an option. Although allocating supporter funds to marketing and other administrative programs is certainly legal and required to succeed, organizations should seek to identify and encourage potentially less-efficient organizations to change behaviours to avoid alienating current donors. Further, organizations should work with the media to minimize the negative attention currently being focused on the financial aspects of non-profit organizations.  3.4.3 Directions for Future Research  The findings from essay 2 offer researchers ample opportunities for future research. One such avenue could be to examine whether individual differences exist that amplify or mitigate the observed prosocial identity consequence? For example, could the connection the supporter feels to the organization or cause moderate the identity consequence? In essay 1, individuals highly emotionally connected to the cause responded more positively to public token support, suggesting a resistance to the impression management motivation and a reliance on the strong positive feelings one has towards the cause. However, might connection to the cause operate differently upon learning the impact one made to an important cause was lower than expected? It is conceivable that after making a significant contribution to a cause that one is highly emotionally connected to and then learning that less of one’s support actually goes to the recipients, a supporter may feel betrayed and exhibit reactions similar to Darke and Richie 106  (2007). On the other hand, it is also possible that consumers highly connected to the cause may grant the cause more leeway in making its financial decisions. Future research examining this factor could be most fruitful.   Another avenue future research could explore is whether additional factors exist that moderate the observed prosocial identity consequence. For example, one such moderator could be the type of meaningful support provided. Does the prosocial identity consequence operate differently when meaningful support is provided through volunteer versus financial channels? Previous research investigating the differences in consumer experiences between time and money finds that focusing on the experience, or time, augments the connection a consumer feels with a product as compared to focusing on money (Mogilner and Aaker 2009). As such, it is possible that the prosocial identity consequence may be amplified among consumers who provide meaningful support through extensive volunteering. Although experimental results on prosocial identity and subsequent support do hold across both volunteer (studies 1-2) and financial (pilot study, study 3) contexts, it is possible this factor may play an interesting moderating role if examined more directly.  Another interesting direction for future researchers would be to test for ceiling effects; more specifically, at what allocation proportion level does the prosocial identity consequence not hold? I deliberately selected a low, but legally allowable proportion of 31.2% as the low allocation condition to make salient the low proportion of support being directly passed on to cause recipients. However, results from the study 1 pre-test showed that preferences for allocation information significantly differed when allocation was raised in 20% increments. 107  Would my results hold if the support allocation was set to 50%? I call on future research to examine at what point the consequence no longer results. This information would be very helpful to non-profit organizations to better understand the range of allocation values they may use to strike a balance between effective promotion and current supporter maintenance.    Finally, future research could vary the type of organization or social cause to examine if the prosocial identity consequence holds across different causes. I deliberately selected positively-viewed causes in order to maximize both support in the experimental paradigm and participant awareness of the costs to non-profit organizations (e.g., providing food, disaster support). However, might reaction to support allocation information vary with complexity of the underlying cause? For example, consumers may understand that research into certain medical causes may be higher than others (i.e., complicated causes such as leukemia) and thus accept a lower allocation proportion.    3.4.4 Essay 2 Conclusion   In conclusion, this essay provides an empirical investigation into an important issue generating significant attention from and impacting charitable organizations, consumers and consumer researchers: charitable support allocation. I propose a conceptual framework suggesting a unique prosocial-identity consequence that predicts when and why allocation information can lead to more or less subsequent support for a charitable organization. 108  Importantly, I distinguish this novel consequence from previously documented competing explanations and introduce novel factors to reconcile my framework with previous research.  109  Chapter 4: Limitations  As is the case with any set of experimental studies, the current research does have limitations that must be noted.   One limitation of my dissertation research is the relatively short length of time between initial support behaviour and subsequent support requests. In the essay 1 paradigm, for example, although previous foot-in-the-door research shows that the effect is present even when time between first and second requests is very short, replicating the observed effects of social observability on subsequent support using a longer time frame would add confidence to my findings. For example, the original Freedman and Fraser (1966) investigation made the larger request three days after the initial request. Subsequent investigations have also used time periods of up to two weeks in nature. Future research could seek to replicate my findings using a longer time period. It would also be very interesting to see if and how the social observability effects change using an extended time period between token and meaningful support.  The essay 2 paradigm is also limited by short timeframes. In these studies, participants are exposed to the support allocation information immediately after providing meaningful support. As such, the costly support task is very salient in participant minds. Although this is useful in an experimental context to test my proposed framework, the impact that learning that less of one’s support actually goes to the cause recipients may be amplified because of the support salience. In reality, consumers may learn of the allocation information weeks or even months after providing support. It is possible that the prosocial identity consequence may 110  weaken as time passes. This would be a very interesting moderator to investigate, and I urge future researchers to test this effect. I note that timing is also an issue for the second meaningful request in essay 2. Rarely do consumers experience a second request for support shortly after providing costly support for a cause (through donation or volunteering). Replicating the effects using longer time periods between initial support, support allocation information and subsequent support requests would add confidence to the framework.  Another limitation of my dissertation research is the use of fictitious charitable organizations. Although I do replicate the predicted results of essay 1 using support for Remembrance Day and the Vancouver Canucks, the experimental studies focus on support for fictitious charitable organizations that participants were not familiar with. The benefit to this paradigm is that it allowed me to maximize the internal validity of the investigations such that I was able to test my theoretical frameworks holding constant cause knowledge and awareness among participants. Nevertheless, I acknowledge that replicating the results using actual organizations would add considerable generalizability to my findings.  A final limitation of the essay 2 results should be noted. Although I do operationalize meaningful support using both volunteer and financial support, participants did not contribute a significant amount of their own personal money to the cause. Many examples of meaningful support for non-profit organizations involve substantial financial contributions to organizations. In study 3, participants endured a painful task to earn $10 for charity, but I acknowledge that the contribution was coming from the Marketing Department and not from participant wallets. I call 111  on future researchers to extend my findings by testing the framework among participants who have made significant financial contributions to charitable organizations.   112  Chapter 5: Concluding Remarks   In sum, this dissertation provides insight into the factors that motivate consumers to provide or withhold meaningful and impactful support for charitable organizations. Through nine empirical studies, the current research contributes two additional moderators to economic models of prosocial behaviour, qualifying previously held beliefs regarding the cost of initial prosocial support, and adding to our theoretical understanding of recurring prosocial behaviour. Moreover, both essays make theoretical contributions to established literature streams by applying these moderators to reconcile competing predictions. In Essay 1, the social observability factor is able to reconcile competing predictions from literature streams nearly fifty years old. Further, Essay 2’s identification of support success helps to increase theoretical understanding of when and why consumers assign credits to their intention to perform a behaviour. Finally, and potentially most importantly, this dissertation empirically investigates two important and prevalent social issues facing consumers, non-profit organizations and academic researchers: slacktivism and charitable allocation support. I hope that applying a more empirical approach to these issues will help provide insight to consumers, non-profit organizations and policy makers regarding the optimal ways to work together to maximize the support to those most in need.    113  Chapter 6: Tables and Figures  6.1 Tables  Table 1. Summary of Results (Essay 1: Study 5)  Token-Support Condition Low Connection to Cause (CC) Effect High Connection to Cause (CC) Effect Low CC Significance Limit (SD) High CC Significance Limit (SD) Private vs. Public BJN = 1.29, SE = .66,  p = .05 BJN = -.78, SE = .40, p = .05 -.21 .75 Private vs. Control BJN = 1.07, SE = .54,  p = .05 BJN = -.85, SE = .44, p = .05 -.54 .47    114  Table 2. Correlation Matrix (Essay 1: Study 3)  Private Token Support     IM Pre-Support IM Post-Support Consistency IM Pre-Support   1 .78** 0.12 IM Post-Support .78** 1 .34* Consistency   0.12 .34* 1           Public Token Support     IM Pre-Support IM Post-Support Consistency IM Pre-Support   1 .79** 0.22 IM Post-Support .79** 1 0.2 Consistency   0.22 0.2 1           ** = p < .01         *  = p < .05                        115  Table 3 a. Summary of Results – Self-Only Ratings (Essay 2: Study 3)    Percentile Estimate of:  Success   Submersion Time Prosocial Identity Intention to Support High Allocation  80.7* 68.1* 76.3* Low Allocation   68.4* 58.7 63*      Failure     High Allocation  34.4 57.4* 69.6* Low Allocation   39.7 59.3* 72.7*      * = p < .05 Values significantly greater than average (50th percentile)            116  Table 3 b. Summary of Results – Self and Observer Ratings (Essay 2: Study 3)    Percentile Estimate of:    Submersion Time Altruism Altruistic Intent Success      High Allocation     Self  80.7* 68.1* 76.3* Observer  89.2* 73.4* 81.9*  Low Allocation     Self  68.4* 58.7 63* Observer   82.9* 72.2* 78.8*      Failure      High Allocation     Self  34.4 57.4* 69.6* Observer  32.1 50.3 45.8  Low Allocation     Self  39.7 59.3* 72.7* Observer   34.1 48.1 45.6 * = p < .05 Values significantly greater than average (50th percentile)     117  6.2 Figures Figure 1. Theoretical Framework of Dissertation        118  Figure 2. Average Amount Donated in Response to a Second Request for Meaningful Support (Essay 1, Study 1)               119  Figure 3. Number of Minutes Volunteered for Subsequent Request (To Stuff Envelopes) to Support Chosen Charity (Essay 1, Study 2)      120  Figure 4. Percentage of Participants Agreeing to Subsequent Volunteer Support Request (Essay 1, Study 3)     121  Figure 5. Percentage of Participants Agreeing to Subsequent Support Request (To Stuff Envelopes for Mail Campaign) as a Function of Value Alignment (Essay 1, Study 4)     122  Figure 6. Prosocial Identity as a Function of Allocation of Support and Prosocial Prime (Essay 2, Study 1)     123  Figure 7. Prosocial Identity as a Function of Allocation of Support and Support Success (Essay 2, Study 3)     124  Figure 8. Subsequent Support as a Function of Allocation of Support and Support Success (Essay 2, Study 3)     125  References  Aknin, Lara B., Elizabeth W. Dunn, Ashley V. Whillans, Adam M. Grant, and Michael I. Norton (2013), "Making a Difference Matters: Impact Unlocks the Emotional Benefits of Prosocial Spending," Journal of Economic Behavior & Organization, 88, 90-95.  Andreoni, James (1990), "Impure Altruism and Donations to Public Goods: A Theory of Warm-Glow Giving," The Economic Journal, 464-477.  Aronson, Elliot (1968), "Dissonance Theory: Progress and Problems." In Ableson et al. (Ed.), Theories of Cognitive Consistency: A Sourcebook (5–27). Chicago: Rand McNally.  Aquino, Karl, and Americus Reed (2002). "The Self-Importance of Moral Identity." Journal of Personality and Social Psychology, 83(6), 1423.  Aquino, Karl, Dan Freeman, Americus Reed II, Vivien KG Lim, and Will Felps (2009), "Testing a Social-Cognitive Model of Moral Behavior: The Interactive Influence of Situations and Moral Identity Centrality," Journal of Personality and Social Psychology, 97(1), 123.  Ashworth, Laurence, Peter R. Darke, and Mark Schaller (2005), “No One Wants to Look Cheap: Trade-offs Between Social Disincentives and the Economic and Psychological Incentives to Redeem Coupons,” Journal of Consumer Psychology, 15(4), 295–306.  126  Bandura, Albert (1991), "Social Cognitive Theory of Moral Thought and Action," Handbook of Moral Behavior and Development, 1, 45-103.  Blasi, Augusto (1980), "Bridging Moral Cognition and Moral Action: A Critical Review of the Literature," Psychological Bulletin, 88(1)  Batson, C. Daniel, and Laura L. Shaw (1991), "Evidence For Altruism: Toward a Pluralism of Prosocial Motives." Psychological Inquiry 2(2), 107-122.  ———, (2014), The Altruism Question: Toward a Social-Psychological Answer, Psychology Press.  Bem, Daryl J. (1972), "Self-Perception Theory," in Advances in Experimental Social Psychology, L. Berkowitz (Ed.), New York: Academic.  Bénabou, Roland, and Jean Tirole (2004), “Willpower and Personal Rules,” Journal of Political Economy, 112(4), 848–886.  ———  (2005), “Incentives and Prosocial Behavior,” National Bureau of Economic Research.  Berman, Jonathan Z., Emma E. Levine, Alixandra Barasch, and Deborah A. Small (2014), "The Braggart's Dilemma: On the Social Rewards and Penalties of Advertising Prosocial Behavior," Journal of Marketing Research. 127   Bhattacharya, Chitrabhan B., and Sankar Sen (2003), "Consumer-Company Identification: A Framework for Understanding Consumers' Relationships With Companies," Journal of Marketing, 76-88.  Bolino, Mark C (1999), "Citizenship and Impression Management: Good Soldiers or Good Actors?," Academy of Management Review, 24(1), 82-98.  Burger, Jerry. M (1999), “The Foot-in-the-Door Compliance Procedure: A Multiple-Process Analysis and Review,” Personality and Social Psychology Review, 3(4), 303–325.  Carver, Charles S., Eileen DeGregorio, and Rod Gillis (1980), "Field-Study Evidence of an Ego-Defensive Bias in Attribution Among Two Categories of Observers," Personality and Social Psychology Bulletin, 6(1), 44-50.  Center for Social Impact Communication, Georgetown University (2011), “Dynamics of Cause Engagement - Final Report,” Education.  Chartrand, Tanya, Shannon Pinckert, and Jerry. M. Burger (1999), “When Manipulation Backfires: The Effects of Time Delay and Requester on the Foot‐in‐the‐Door Technique,” Journal of Applied Social Psychology, 29(1), 211–221.  128  Cialdini, Robert B., Melanie R. Trost, and Jason T. Newsom (1995), "Preference For Consistency: The Development of a Valid Measure and the Discovery of Surprising Behavioral Implications," Journal of Personality and Social Psychology, 69(2), 318.  ———,  Stephanie L. Brown, Brian P. Lewis, Carol Luce, and Steven L. Neuberg (1997), "Reinterpreting the Empathy–Altruism Relationship: When One Into One Equals Oneness," Journal of Personality and Social Psychology, 73(3), 481.  Cryder, Cynthia E., George Loewenstein, and Richard Scheines (2013) "The Donor is in the Details," Organizational Behavior and Human Decision Processes, 120(1), 15-23.  ———, George Loewenstein, and Howard Seltman (2013), "Goal Gradient in Helping Hehavior," Journal of Experimental Social Psychology, 49(6), 1078-1083.  Darke, Peter R. and Robin J.B. Richie (2007), “The Defensive Consumer: Advertising Deception, Defensive Processing, and Distrust,” Journal of Marketing Research, 44, 114-27.  ———, Laurence Ashworth and Robin J.B. Richie (2008), “Damage From Corrective Advertising: Causes and Cures,” Journal of Marketing, 72, 81-97.  129  Davis, Jesse (2011), “Cause Marketing: Moving Beyond Corporate Slacktivism,” http://evidencebasedmarketing.net/cause-marketing-moving-beyond-corporate-slacktivism/. Accessed January 10, 2013.  DeJong, William (1979), "An Examination of Self-Perception Mediation of the Foot-in-the-Door Effect," Journal of Personality and Social Psychology, 37(12), 2221.   Dillard, James P., John E. Hunter, and Michael Burgoon (1984), "Sequential-Request Persuasive Strategies," Human Communication Research, 10(4), 461-488.   Duncan, Brian (2004), “A Theory of Impact Philosophy,” Journal of Public Economics 88, 2159. 2180.  Dunn, Elizabeth W., Lara B. Aknin and Michael Norton (2008), “Spending Money on Others Promotes Happiness,” Science, 319, 1687-1688.  Effron, Daniel A., Jessica S. Cameron, and Benoît Monin 2009), "Endorsing Obama Licenses Favoring Whites," Journal of Experimental Social Psychology, 45(3), 590-593.  Fein, Steven (1996), "Effects of Suspicion on Attributional Thinking and the Correspondence Bias," Journal of Personality and Social Psychology, 70(6), 1164. 130  Fenigstein, Allan, Michael F. Scheier, and Arnold H. Buss (1975), "Public and Private Self-Consciousness: Assessment and Theory," Journal of Consulting and Clinical Psychology 43(4), 522.  Festinger, Leon (1957), A Theory of Cognitive Dissonance. Vol. 2., Stanford University Press.  Flannery, Helen and Rob Harris (2009), “DonorCentrics U.S. Recurring Giving Benchmarking Analysis,” Target Analytics.  Fox, Lee (2012), “Stop Calling Them ‘Slacktivists’,” http://koodooz.wordpress.com/2012/08/09/stop-calling-them-slacktivists. Accessed June 15, 2013.   Freedman, Jonathan L., and Scott C. Fraser (1966), “Compliance Without Pressure: The Foot-in-the-Door Technique.,” Journal of Personality and Social Psychology, 4(2), 195.  Gibbons, Frederick X. (1990), “Self-Attention and Behavior: A Review and Theoretical Update,” Advances in Experimental Social Psychology, 23, 249–303.  Gladwell, Malcolm (2010), “Small Change,” The New Yorker, 4, 42-49.  131  Gneezy, Ayelet, Alex Imas, Amber Brown, Leif D. Nelson, and Michael I. Norton (2011), “Paying to be Nice: Consistency and Costly Prosocial Behavior,” Management Science, 58(1), 179–187.  Gneezy, Uri, Elizabeth A. Keenan, and Ayelet Gneezy (2014), “Avoiding Overhead Aversion in Charity,” Science, 346 (6209), 632-635.  Goffman, Erving (1959), The Presentation of Self in Everyday Life. Garden City, NY.  Goldberg, Eleanor (2013), “Locks Of Love Has More Than $6 Million Worth Of Donated Hair That Is Unaccounted For: Report,“ http://www.huffingtonpost.com/2013/05/14/locks-of-love-controversy_n_3269078.html. Accessed November 30, 2013.   Goukens, Caroline, Siegfried Dewitte, and Luk Warlop (2009), "Me, Myself, and My Choices: The Influence of Private Self-Awareness on Choice," Journal of Marketing Research, 46(5), 682-692.  Grant, Adam M., and David M. Mayer (2009), "Good Soldiers and Good Actors: Prosocial and Impression Management Motives as Interactive Predictors of Affiliative Citizenship Behaviors," Journal of Applied Psychology, 94(4), 900.  132  Greenberg, Jeff, Tom Pyszczynski, and Sheldon Solomon (1982), "The Self-Serving Attributional Bias: Beyond Self-Presentation," Journal of Experimental Social Psychology, 18(1), 56-67.  Grouzet, Frederick ME, Tim Kasser, Aaron Ahuvia, José Miguel Fernández Dols, Youngmee Kim, Sing Lau, Richard M. Ryan, Shaun Saunders, Peter Schmuck, and Kennon M. Sheldon (2005), "The Structure of Goal Contents Across 15 Cultures," Journal of Personality and Social Psychology, 89(5), 800.  Harari, Herbert, Deborah Mohr, and Karen Hosey (1980), "Faculty Helpfulness to Students, A Comparison of Compliance Techniques," Personality and Social Psychology Bulletin 6(3), 373-377.  Hays, David (2013), “Charity Runners or "The Purple Horde",” http://davidhays.net/running/tnt.html. Accessed November 30, 2013.  Heider, Fritz (1958), The Psychology of Interpersonal Relations, New York: Wiley.  Heine, Steven J., Travis Proulx, and Kathleen D. Vohs (2006), "The Meaning Maintenance Model: On the Coherence of Social Motivations," Personality and Social Psychology Review, 10(2), 88-110.  133  Institute for Volunteering Research (2012), Volunteering Stats, http://www.ivr.org.uk/ivr-volunteering-stats. Accessed March 10, 2014.   Johnson, Palmer Oliver, and Jerzy Neyman (1936), "Tests of Certain Linear Hypotheses and Their Application to Some Educational Problems," Statistical Research Memoirs.  Khan, Uzma, and Ravi Dhar (2006), “Licensing Effect in Consumer Choice,” Journal of Marketing Research, 43, 259–266.  ——— (2007), “Where There is a Way, is There a Will? The Effect of Future Choices on Self-Control,” Journal of Experimental Psychology: General, 136(2), 277.  King, Samantha (2006), Pink Ribbons, Inc.: Breast Cancer and the Politics of Philanthropy, University of Minnesota Press.  Klein, William M., and Ziva Kunda (1992), "Motivated Person Perception: Constructing Justifications for Desired Beliefs," Journal of Experimental Social Psychology, 28(2), 145-168.  Kouchaki, Maryam (2011), "Vicarious Moral Licensing: The Influence of Others' Past Moral Actions on Moral Behavior," Journal of Personality and Social Psychology, 101(4), 702.  134  Krishna, Aradhna (2011), "Can Supporting a Cause Decrease Donations and Happiness? The Cause Marketing Paradox," Journal of Consumer Psychology, 21(3), 338-345.  Kristof, N. (2008), "Geezers Doing Good," The New York Times, 17.  Kristofferson, Kirk, Katherine White and John Peloza (2014), “The Nature of Slacktivism: How the Social Observability of an Initial Act of Token Support Impacts Subsequent Prosocial Action,” Journal of Consumer Research, 40(6), 1149-1166.  Kruger, Justin, and Thomas Gilovich (2004), "Actions, Intentions, and Self-Assessment: The Road to Self-Enhancement is Paved with Good Intentions," Personality and Social Psychology Bulletin, 30(3), 328-339.  Lamberton, Cait Poynor, Rebecca Naylor and Kelly Haws (2013), "Same Destination, Different Paths: The Influence of Divergent Reasoning on the Persuasive Effects of Others," Journal of Consumer Psychology, 23(1), 74-89.  Leary, Mark. R., and Robin M. Kowalski (1990), “Impression Management: A Literature Review and Two-Component Model,” Psychological Bulletin, 107(1), 34.  Leonard, Allison (2013), “Colour Runs are Bullshit”, http://www.thealbatross.ca/27335/colour-runs-are-a-faux-charity-scam. Accessed November 5, 2013.  135  Lipsman, Andrew, Carmela Aquino and Patrick Kemp (2012), “The Power of Like 2: How Social Marketing Works,” comScore, http://www.comscore.com/Insights/Presentations-and-Whitepapers/2012/The-Power-of-Like-2-How-Social-Marketing-Works. Accessed January 15, 2015.  Mazar, Nina and Chen-Bo Zhong (2010), “Do Green Products Make Us Better People?,” Psychological Science, 21(4), 494–498.  Mogilner, Cassie, and Jennifer Aaker (2009), "“The Time vs. Money Effect”: Shifting Product Attitudes and Decisions Through Personal Connection," Journal of Consumer Research, 36(2), 277-291.   MoneySense (2014), “2015 Charity 100: A Smarter Way to Give”, http://www.moneysense.ca/planning/2015-charity-100-a-smarter-way-to-give. Accessed January 15, 2015.  Monin, Benoit and Dale T. Miller (2001), “Moral Credentials and the Expression of Prejudice,” Journal of Personality and Social Psychology, 81(1), 33.  Morozov, Evgeny (2009a), “The Brave New World of Slacktivism,” Foreign Policy, 19.  ——— (2009b), “From Slacktivism to Activism,” Foreign Policy Blogs.  136  Mullen, Brian, and Catherine A. Riordan (1988), "Self‐Serving Attributions for Performance in Naturalistic Settings: A Meta‐Analytic Review," Journal of Applied Social Psychology 18(1), 3-22.  National Centre for Charitable Statistics (2014), http://nccs.urban.org. Accessed March 15, 2014.  Newman, George E., and Daylian M. Cain (2014), "Tainted Altruism When Doing Some Good Is Evaluated as Worse Than Doing No Good at All," Psychological Science.  New York Times (2007), “Lather, Rinse, Donate.” http://www.nytimes.com/2007/09/06/fashion/06locks.html?oref=login&_r=0. Accessed November 30, 2013.  Nonprofit Investor (2013), http://www.nonprofitinvestor.org/locksoflove. Accessed November 30, 2013.  Nonprofit Investor Reddit (2013) http://www.reddit.com/r/Infographics/comments/1e91xx/6mm_of_hair_donations_unaccounted_for_by_locks_of/). Accessed November 30, 2013.  Obar, Jonathan A., Paul Zube, and Cliff Lampe (2011), “Advocacy 2.0: An Analysis of How Advocacy Groups in the United States Perceive and Use Social Media as Tools for Facilitating Civic Engagement and Collective Action,” SSRN eLibrary. 137   Olivola, Christopher Y., and Eldar Shafir (2013), "The Martyrdom Effect: When Pain and Effort Increase Prosocial Contributions," Journal of Behavioral Decision Making, 26(1), 91-105.  Oppenheimer, Daniel M., and Christopher Y. Olivola, (2011), “The Science of Giving: Experimental Approaches to the Study of Charity,” Psychology Press.  Preacher, Kristopher J., Derek D. Rucker, and Andrew F. Hayes (2007), “Addressing Moderated Mediation Hypotheses: Theory, Methods, and Prescriptions,” Multivariate Behavioral Research, 42(1), 185–227.  Pollay, Richard W. (1986), “The Distorted Mirror: Reflections on the Unintended Consequences of Advertising,” Journal of Marketing, 50, 18-36.  Posner, Barry Z. (1992), "Person-Organization Values Congruence: No Support for Individual Differences as a Moderating Influence," Human Relations, 45, 351-361.  Ratner, Rebecca K., and Barbara E. Kahn (2002), "The Impact of Private versus Public Consumption on Variety-Seeking Behavior," Journal of Consumer Research, 29(2), 246-257.  138  Reeder, Glenn D (2009), "Mindreading: Judgments About Intentionality and Motives in Dispositional Inference," Psychological Inquiry, 20(1), 1-18.  Sachdeva, Sonya, Rumen Iliev, and Douglas L. Medin (2009), “Sinning Saints and Saintly Sinners,” Psychological Science, 20(4), 523–528.  Schlenker, Barry R. (1980), Impression Management: The Self-Concept, Social Identity, and Interpersonal Relations. Monterey, CA: Brooks/Cole Publishing Company.  Seligman, Clive, Malcolm Bush, and Kenneth Kirsch (1976), “Relationship Between Compliance in the Foot-in-the-Door Paradigm and Size of First Request,” Journal of Personality and Social Psychology, 33(5), 517.  Sengupta, Jaideep, Darren W. Dahl, and Gerald J. Gorn (2002), "Misrepresentation in the Consumer Context," Journal of Consumer Psychology, 12(2), 69-79.  Sherman, David K., and Geoffrey L. Cohen (2006), "The Psychology of Self-Defense: Self-Affirmation Theory," Advances in Experimental Social Psychology, 38, 183.  ———, Geoffrey L. Cohen, Leif D. Nelson, A. David Nussbaum, Debra P. Bunyan, and Julio Garcia (2009), "Affirmed Yet Unaware: Exploring the Role of Awareness in the Process of Self-Affirmation," Journal of Personality and Social Psychology, 97(5), 745.  139  Simonson, Itamar, and Stephen M. Nowlis (2000), "The Role of Explanations and Need for Uniqueness in Consumer Decision Making: Unconventional Choices Based on Reasons," Journal of Consumer Research, 27(1), 49-68.  Small, Deborah A. and Uri Simonsohn (2008), “Friends of Victims: Personal Experience and Prosocial Behavior,” Journal of Consumer Research, 35, 532-542.  Spencer, Steven J., Mark P. Zanna, and Geoffrey T. Fong (2005), "Establishing a Causal Chain: Why Experiments are Often More Effective Than Mediational Analyses in Examining Psychological Processes," Journal of Personality and Social Psychology, 89(6), 845-851.  UNICEF Sweden (2013), "Likes Don't Save Lives Promotional Campaign," http://unicef.se/. Accessed June 3, 2013.  Vohs, Kathleen D., Nicole L. Mead, and Miranda R. Goode (2006), "The Psychological Consequences of Money," Science, 314(5802), 1154-1156.  Wang, Theodore, Richard Brownstein, and Richard Katzev (1989), "Promoting Charitable Behaviour with Compliance Techniques," Applied Psychology, 38(2), 165-183.  Weiner, Bernard (2001), "Responsibility for Social Transgressions: An Attributional Analysis," Intentions and Intentionality: Foundations of Social Cognition, 331-344.  140  Wilcox, Keith, and Andrew Stephen (2013), "Are Close Friends the Enemy? Online Social Networks, Self-Esteem, and Self-Control," Journal of Consumer Research, 40(3), 12-57.  Winterich, Karen Page, Vikas Mittal, and Karl Aquino (2013), “When Does Recognition Increase Charitable Behavior? Toward A Moral Identity Based Model,” Journal of Marketing, 77(3), 121-134.  White, Katherine, and Darren W. Dahl (2006), “To Be or Not to Be? The Influence of Dissociative Reference Groups on Consumer Preferences,” Journal of Consumer Psychology, 16(4), 404.  ——— (2007), “Are All Out-Groups Created Equal? Consumer Identity and Dissociative Influence,” Journal of Consumer Research, 34(4), 525–536.  White, Katherine and John Peloza (2009), “Self-Benefit Versus Other-Benefit Marketing Appeals: Their Effectiveness in Generating Charitable Support,” Journal of Marketing, 73(4), 109–124.  Zhao, Shanyang, Sherri Grasmuck, and Jason Martin (2008), “Identity Construction on Facebook: Digital Empowerment in Anchored Relationships,” Computers in Human Behavior, 24(5), 1816–1836.    141  Appendix: Experimental Materials A.1 Essay 1, Study 2 Instructions Token-Support Condition: Both charitable organizations, Combating Poverty in Developing Nations and International Disaster Relief Support, are currently running campaigns to lobby the Canadian government to increase support for their causes. We would sincerely appreciate it if you would choose to sign one of the petitions.   Please choose to sign one of the petitions now. To sign the petition, please go to the front of the room and sign your name on the petition taped to the white board (sign your name on the paper provided and place your paper in the appropriate charity's private boxes located at the back of the room). After the session, the Marketing Department will forward the signatures to the charitable organizations.  No-Support Condition: Before you continue with the survey and feedback, we would like to take attendance to ensure proper documentation for your participation. We would sincerely appreciate it if you would choose to sign your name.   To record your attendance in this study, please go to the front of the room and sign the attendance sheet taped to the white board (sign your name on the paper provided and place your paper in the private box located at the back of the room). 142  A.2 Essay 1, Study 2 Pre-Test  How likely is it that you would make a monetary donation (e.g., make a donation of money) to each of the following causes? (Please rate your response for each cause on the following scale, with 7 being “very likely” and 1 being “very unlikely.”)  Very                                          Very Unlikely                                     Likely          AIDS  1 2 3 4 5 6 7 Alzheimer’s Disease   1 2 3 4 5 6 7 Animal Rights  1 2 3 4 5 6 7 Breast Cancer  1 2 3 4 5 6 7 Colorectal Cancer  1 2 3 4 5 6 7 Depression  1 2 3 4 5 6 7 Drug Addiction  1 2 3 4 5 6 7 Heart Disease  1 2 3 4 5 6 7 Homelessness  1 2 3 4 5 6 7 Leprosy  1 2 3 4 5 6 7 Lung Cancer  1 2 3 4 5 6 7 Natural Disaster Relief  1 2 3 4 5 6 7 Pancreatic Cancer  1 2 3 4 5 6 7 Physical Disabilities  1 2 3 4 5 6 7 Schizophrenia  1 2 3 4 5 6 7 Skin Cancer  1 2 3 4 5 6 7 Stomach Cancer  1 2 3 4 5 6 7 Third World Hunger/Poverty  1 2 3 4 5 6 7 Violence Against Women  1 2 3 4 5 6 7 143   Page 2 1 2 3 4 5 6 7 Part II: Impressions Towards Causes  How likely is it that you would be publically associated (e.g., publically declare your monetary or volunteer support) with supporting each of the following causes? (Please rate your response for each cause on the following scale, with 7 being “very likely” and 1 being “very unlikely.”)  Very                                          Very Unlikely                                     Likely          AIDS  1 2 3 4 5 6 7 Alzheimer’s Disease   1 2 3 4 5 6 7 Animal Rights  1 2 3 4 5 6 7 Breast Cancer  1 2 3 4 5 6 7 Colorectal Cancer  1 2 3 4 5 6 7 Depression  1 2 3 4 5 6 7 Drug Addiction  1 2 3 4 5 6 7 Heart Disease  1 2 3 4 5 6 7 Homelessness   1 2 3 4 5 6 7 Leprosy  1 2 3 4 5 6 7 Lung Cancer  1 2 3 4 5 6 7 Natural Disaster Relief  1 2 3 4 5 6 7 Pancreatic Cancer  1 2 3 4 5 6 7 Physical Disabilities  1 2 3 4 5 6 7 Schizophrenia  1 2 3 4 5 6 7 Skin Cancer  1 2 3 4 5 6 7 Stomach Cancer  1 2 3 4 5 6 7 Third World Hunger/Poverty  1 2 3 4 5 6 7 Violence Against Women 1 2 3 4 5 6 7 144   A.3 Essay 1, Study 2 Sliding Scale Dependent Variable Earlier this session, you read about the charitable organization International Disaster Relief Support.   International Disaster Relief Support is requesting support from UBC students to help prepare for their campaign rollout to generate awareness of and support for their volunteering activities. This organization will be launching a mail campaign and needs help stuffing envelopes during the next two weeks. During this task, you will have an individual work station and will work alone.  Please use the slider below to indicate how much time you are willing to provide to this organization. If you are willing to participate in this session, you will be contacted to decide upon a convenient time to help us out.      145  A.4 Essay 1, Study 3: Impression-Management Scale Scale:  (1 – Strongly Disagree to 7 – Strongly Agree)  I care about how positively others view me. I want to present myself in a positive way. I want to make a positive impression on others. I want to make myself look good to others. I want to do what the norm is. I want to do what society believes is the right thing.     146  A.5 Essay 1, Study 3: Preference for Consistency Sub-Scale Scale:  (1 – Strongly Disagree to 7 – Strongly Agree)  I'm uncomfortable holding two beliefs that are inconsistent. I typically prefer to do things the same way. The appearance of consistency is an important part of the image I present to the world. It bothers me when someone I depend on is unpredictable. I want my close friends to be predictable. It is important that others view me as a stable person.     147  A.6 Essay 1, Study 4: Charitable Organization Facebook Screen-Shots    148  A.7 Essay 2, Study 4: Value Alignment Manipulation Instructions:   First, we would like you to read about some of the latest activities Combating Poverty in Developing Nations has been involved in. Here is a newspaper brief written about the organization.  Ambiguous News Article:  July 8, 2012. VANCOUVER. A newly formed charitable organization located in the Lower Mainland has been quite active in generating attention and support for their cause. Combating Poverty in Developing Nations is an organization that generates aid to those suffering from extreme poverty in areas such as Africa, India and parts of Asia. Their mission is to provide fair, equitable, and safe living conditions for those living in poverty. This past February, CPDN met with Federal officials to join the West Africa Children's Education Fund. This fund will provide effective, accountable, and timely financial support to trusted Canadian and international humanitarian and development partner organizations working in areas affected by extreme poverty. For every dollar Canadians donated between March 12 and September 12, 2012, to a registered charity supporting education programs in West Africa, CPDN has matched this by placing one dollar into the West Africa Children's Education Fund.  However, not all of the organization's activities have been positively received by the public. CPDN has recently very vocally protested the UK’s appointment of Alan Milburn as a member of parliament. Alan Milburn had recently been appointed as the Independent Reviewer of Social Mobility and Child Poverty, and these organizations demanded the UK Cabinet Office to provide evidence of Milburn’s work to justify the appointment. CPDN has accused Milburn of not being as transparent with the distribution and allocation of funds directed towards various charitable organizations. However, CPDN has not provided any evidence to back up such claims and other organizations have implied that CPDN is using these claims to gain additional media exposure for their cause.  While the end goal may be noble,  CPDN has publicly attacked Milburn in the various British media outlets and has essentially labelled him a crook before he has even begun to act. This behaviour could raise questions regarding the tactical strategy of the organization.  Values Aligned Manipulation:  Before you move on to the specific questions, please take 1-2 minutes and reflect on how this cause's values are similar to your values. Once you've thought about this, please describe this value alignment in as much detail as possible.  Values Misaligned Manipulation:  Before you move on to the specific questions, please take 1-2 minutes and reflect on how this cause's values are different to your values. Once you've thought about this, please describe this value mis-alignment in as much detail as possible. 149   A.8 Essay 2, Study 4 Follow-Up Study: Online Petition Screen-Shot    150  A.9 Essay 2, Pilot Study Scenario Please imagine the following:  You have a full-time job and work approximately 35 hours per week (140 hours per month).  Now, imagine that you work with a charity that addresses the issue of child poverty, and work very hard to support this charity. You frequently share information about the charity on social media, volunteer a large portion of your free time and donate a large portion of your income each month.    151  A.10 Essay 2 Allocation of Support Stimuli  High Allocation Condition Stimuli:  Now, because of the way the charity is run, 97.2% of volunteer efforts actually support the impoverished people who benefit from the charity. Please read the press release below to learn more about how the charity allocates its support. PRESS RELEASE: Where the Support Goes In developing countries, almost one out of every 15 children will die before they reach the age of five due to malnourishment. In response to problems in developing countries, an organization called Combating Poverty in Developing Nations was created. This organization provides individuals with the opportunity to support this cause through both volunteer and financial contributions. In 2012, individuals' support generated $845,000 USD of revenue. Recently, information regarding cost details has become available. Of the $845,000 raised by supporter efforts in 2012, 97.2% has been passed directly to those most in need as identified by the United Nations World Food Program.  Thus, giving support through Combating Poverty in Developing Nations is a step in the right direction. They are managing supporter efforts in ways they deem as appropriate.   ~Courtesy of the Centre for World Hunger Studies (CWH)   Retained	  funds	  Fundraising	  costs	  Administra?on	  costs	  Rice	  Allocated	  152  Low Allocation Condition Stimuli:  Now, because of the way the charity is run, 31.2% of volunteer efforts actually support the impoverished people who benefit from the charity. Please read the press release below to learn more about how the charity allocates its support. PRESS RELEASE: Where the Support Goes In developing countries, almost one out of every 15 children will die before they reach the age of five due to malnourishment. In response to problems in developing countries, an organization called Combating Poverty in Developing Nations was created. This organization provides individuals with the opportunity to support this cause through both volunteer and financial contributions. In 2012, individuals' support generated $845,000 USD of revenue. Recently, information regarding cost details has become available. Of the $845,000 raised by supporter efforts in 2012, only 31.2% has been passed directly to those most in need as identified by the United Nations World Food Program.  Thus, giving support through Combating Poverty in Developing Nations is a step in the right direction. They are managing supporter efforts in ways they deem as appropriate.   ~Courtesy of the Centre for World Hunger Studies (CWH)      Retained	  funds	  Fundraising	  costs	  Administra?on	  costs	  Rice	  Allocated	  153  A.11 Moral Identity Scale (Aquino and Reed 2002)  Here are some characteristics that might describe a person: Caring, Compassionate, Fair, Friendly, Generous, Helpful, Hardworking, Honest, Kind   The person with these characteristics could be you or it could be someone else. For a moment, visualize in your mind the kind of person who has these characteristics. Imagine how that person would think, feel, and act. When you have a clear image of what this person would be like, answer the following questions using the following scale provided (1 = Strongly Disagree, 7 = Strongly Agree).   _____ 1. It would make me feel good to be a person who has these characteristics. (I) _____ 2. Being someone who has these characteristics is an important part of who I am. (I)  _____ 3. I often wear clothes that identify me as having these characteristics. (S) _____ 4. I would be ashamed to be a person who had these characteristics. (I) _____ 5. The types of things I do in my spare time (e.g., hobbies) clearly identify me as having these characteristics. (S) _____ 6. The kinds of books and magazines that I read identify me as having these characteristics. (S) _____ 7. Having these characteristics is not really important to me. (I) _____ 8. The fact that I have these characteristics is communicated to others by my membership in certain organizations. (S) _____ 9. I am actively involved in activities that communicate to others that I have these characteristics. (S) _____ 10. I strongly desire to have these characteristics. (I)   154  A.12 Essay 2, Study 3 Verbal Instructions: You will be completing multiple studies today. The first study is an opportunity for you to raise money for a local charity.  In this study, you will hold your arm in a bucket of cold water for 10 minutes and in the marketing department will donate $10 to the charity. Although the task is harmless, it is painful as the water is between 0 and 5 degrees Celsius and is very cold to the touch. You can stop the experiment at any time and still receive your course credit, but not the $10 donation to the cause. Do you agree to participate?  Please read the descriptions of these two charities and let me know which one you would like to raise money for.   Do you intend to keep your hand submerged for the 10 minutes?  Finally, with your permission, we would like to video tape this portion of the experiment. Are you ok with that?   155  A.13 Essay 2, Study 3 Charity Descriptions Combating Poverty in Developing Nations   Almost half the world — over three billion people — live on less than $2.50 a day. According to UNICEF, 22,000 children die each day due to poverty. And they “die quietly in some of the poorest villages on earth, far removed from the scrutiny and the conscience of the world. Being meek and weak in life makes these dying multitudes even more invisible in death.” Around 27-28 percent of all children in developing countries are estimated to be underweight or stunted. The two regions that account for the bulk of the deficit are South Asia and sub-Saharan Africa.  Mission: We work to improve child well-being and to serve people around the world regardless of religion, race, gender or ethnicity. Poverty is a complex problem, and our non-profit organization seeks to address its root causes in a holistic way by focusing its efforts on three main areas.  Core Values:  We have identified certain values that lie at the centre of our work. These "Core Values" are the fundamental principles that determine our commitment to each other and to the poor.  1. We are committed to the poor.  We are called to relieve their suffering and to promote the transformation of their condition of life.  2. We value people.  We regard all people as equals. We give priority to people before money, structure, and systems.   3. We are stewards.  We are faithful to the purpose for which resources are given and manage them in a manner that brings maximum benefit to the poor.  4. We are partners.  We are members of an international non-profit partnership that transcends legal, structural, and cultural boundaries.  5. We are responsive.  We are responsive to the life-threatening emergencies where our involvement is needed and appropriate.  Why should people support? One word: Need. There is such a great need for children and families living in poverty to have their voices heard. Our volunteers advocate and fundraise on behalf of these children by encouraging people like you to partner with us to make a difference.      156  International Disaster Relief Support  Each year, disasters devastate millions of people around the world and are increasing in frequency and severity. Houses are destroyed, roads are washed away and livelihoods are ruined. But when disaster strikes, organizations must be there to help.  Mission: We are a non-profit organization headquartered in Vancouver, British Columbia, Canada. We are passionate about helping those affected by natural disasters, regardless of location in the world. Disasters strike each year and take years to recover - people need our help.  Our Beliefs: We believe in high ethics and helping people quickly. We also believe in transparency and stretching every dollar as far as possible. Have you ever given to a mega-charity and wondered when or how your money was used? We provide prompt tax receipts and communications about how and when your money is put to work to help others!  Our ability to respond immediately following a disaster with much needed supplies and personnel is essential for saving lives and minimizing the impact of disasters on vulnerable communities.  Disasters continue to strike and your support allows us to respond to emergencies, well before any money has been received from supporters. It also allows us to implement disaster preparedness programs to create stronger and more resilient communities. Help us help those in need. Join our group today and learn more about how you can help.   

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.24.1-0102480/manifest

Comment

Related Items