Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The similarities and differences of men’s and women’s personal work networks Stackman, Richard W. 1995

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

Item Metadata


831-ubc_1995-060675.pdf [ 3.16MB ]
JSON: 831-1.0088343.json
JSON-LD: 831-1.0088343-ld.json
RDF/XML (Pretty): 831-1.0088343-rdf.xml
RDF/JSON: 831-1.0088343-rdf.json
Turtle: 831-1.0088343-turtle.txt
N-Triples: 831-1.0088343-rdf-ntriples.txt
Original Record: 831-1.0088343-source.json
Full Text

Full Text

THE SIMILARITIES AND DIFFERENCES OFMEN’S AND WOMEN’S PERSONAL WORK NETWORKSbyRICHARD W. STACKMANB.S. (Honors), University of California, Berkeley, 1985A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESFaculty of Commerce and Business AdministrationIndustrial Relations Management ProgrammeWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJuly 1995© Richard W. Stackman, 1995ABSTRACTAlthough network analysis has proven a useful approach to the study oforganizations and organizational behavior, very little research has been done on the issue ofgender differences in personal work networks. While there is considerable conventionalwisdom about how men and women associate with their colleagues in the workplace, thematter still requires scientific scrutiny. The purpose of this thesis is to provide much-needed descriptive evidence about the ways and extent to which the personal worknetworks of men and women managers or supervisors differ, and to illustrate howpromotions impact such networks. Two separate studies comprise the thesis.The first study considered the personal work networks of men and womensupervisors or managers who had not undergone formal career transitions in the previous1 2 months. Network characteristic data were generated through a questionnaire returnedby 242 individuals, representing three large Canadian companies in the banking, forestry,and insurance industries. Multiple regression was used to test for gender differences andcompany moderator effects.Contrary to the common assumption currently found in the literature that there aregender differences in personal work networks, this study found that differences in men’sand women’s association patterns at work were more likely in their expressive, rather thaninstrumental, networks. Though managers and supervisors were more likely to formhomophilous ties, and men had networks of greater density, the results suggested that menand women had comparable instrumental work networks. The expressive networks,however, exhibited greater gender differences. Significant differences included the gender,location, the density, and the frequency of contact of these expressive ties.The second study considered the personal work networks of men and womensupervisors or managers who had recently been promoted. Network characteristic datawere generated through a questionnaire returned by 33 individuals working for a leadingCanadian bank; however, possibly because of sampling deficiencies, no significantinstrumental or expressive network differences were identified. Moreover, there were nodifferences in the turnover of individuals in men’s and women’s instrumental and expressivenetworks following promotions.A discussion regarding the status and future of personal work network researchconcludes the thesis.IIITABLE OF CONTENTSAbstractTable of ContentsList of TablesList of FiguresAcknowledgementsChapter One: Introduction and Literature ReviewNetworks DefinedRelationships and the Correlation Between Time and ProficiencyTypes of Work RelationshipsPersonal Work Networks and GenderTheoretical Explanations for Gender Personal Network Differences .Presumed Gender Differences in Personal Work NetworksNetwork Studies in Work OrganizationsPurpose of the Current StudyChapter Two: Research Questions and HypothesesThe Similarities and Differences of Men’s and Women’sPersonal Work NetworksInstrumental TiesExpressive TiesMen’s and Women’s Work Networks Following PromotionsInstrumental TiesExpressive TiesChapter Three: Research ModelsNon-Transition/Instrumental Network ModelPromotion/Instrumental Network ModelNon-Transition/Expressive Network ModelPromotion/Expressive Network ModelHypotheses 2 and 3Chapter Four: Populations and SamplesThe Data Sites .Non-Transition SamplePromotion SampleSummaryChapter Five: Data CollectionThe QuestionnairesThe Non-Transition QuestionnaireThe Promotion Questionnaires 49IIivvixixii12345681015161617171818192222262730303333343639404141ivPilot Testing of the Questionnaire 50Reliability of the Questionnaire 51Response Rate 54Chapter Six: Results 55The Non-Transition Study 56The Use of Multiple Regression 59The Regression Models 60Results of Company and Gender Effects 65The Promotion Study 77Chapter Seven: Discussion and Summary 83The Non-Transition Study 83Gender Effects 83Interpreting the Instrumental and Expressive Network Differences . . 87Company effects 88Time effects 90The Promotion Study 92Limitations 95The Status and Future of Personal Work Network Research 97Personal Work Networks 97Roles of Personal Work Networks 101Summary 104References 106Appendix A: Letters and Questionnaires 111Letter 1: Dean Goldberg’s Data Site Approach Letter 11 2Letter 2: Dr. Pinder’s Data Site Approach Letter 114Letter 3: First Cover Letter Included with Questionnaire 11 6Letter 4: Second Cover Letter Included with Questionnaire 11 7Letter 5: Follow-up Letter 11 8Exhibit 1: Non-Transition Sample Questionnaire 11 9Exhibit 2: First Promotion Sample Questionnaire 131Exhibit 3: Follow-up Promotion Sample Questionnaire 144Exhibit 4: Reliability Check Questionnaire 1 53Appendix B: Non-Transition Data: Variable Means, Standard Deviations,and Correlations; Promotion Data: Variable Means andStandard Deviations 161Appendix C: Respondent Set Demographics 189Appendix 0: Dependent Variable Beta Coefficients 199vLIST OF TABLESTable 1 .1: Summary of Relevant Personal Work Network Empirical Studies 1 3Table 3.1: Thesis Research Models 23Table 4.1: Male and Female Participation by Company in Non-Transition Sample 35Table 4.2: Instrumental Network and Expressive Network Site Locationof Individuals Listed by Company 37Table 5.1: Operationalizations of Dependent Variables: Instrumental Network . . 44Table 5.2: Operationalizations of Dependent Variables: Expressive Network . . . 47Table 5.3: Reliability of Respondent’s Demographic Information:Reliability Check Sample 53Table 5.4: Number of Individuals Remaining in, Added to, and Dropped fromRespondents’ Instrumental and Expressive Networks:Reliability Check 53Table 5.5: Non-Transition Study Questionnaire Response Rates 53Table 6.1: Mann-Whitney U Non-Parametric Results:Male vs. Female Instrumental Network Dependent Variables 57Table 6.2: Mann-Whitney U Non-Parametric Results:Male vs. Female Expressive Network Dependent Variables 58Table 6.3: Predictor Variable Operationalizations 62Table 6.4: Base, Shift, and Moderator Model Regression R-Squared:Instrumental Network Dependent Variables 67Table 6.5: Base, Shift, and Moderator Model Regression R-Squared:Expressive Network Dependent Variables . . 68Table 6.6: Comparison of Regression Models:Instrumental Network Dependent Variables 69Table 6.7: Comparison of Regression Models:Expressive Network Dependent Variables 70Table 6.8: Comparison of Respondent Sample vs. Individual CompanyComparison Results: Instrumental Network Dependent Variables . 71Table 6.9: Comparison of Respondent Sample vs. Individual CompanyComparison Results: Expressive Network Dependent Variables . . . 72Table 6.10: Gender Composition of Non-Parametric and Regression ModelsResults: Instrumental Network Dependent Variables 74Table 6.11: Gender Composition of Non-Parametric and Regression ModelsResults: Expressive Network Dependent Variables 75Table 6.12: Gender Non-Parametric and Regression Results:Promotion Sample Instrumental Network Dependent Variables . . . . 78viTable 6.13: Gender Non-Parametric and Regression Results:Promotion Sample Expressive Network Dependent Variables 79Table 6.14: Gender Composition of Turnover in Expressive andInstrumental Networks: Promotion Respondent Sample 81Table B. 1: Non-Transition Respondent Sample Variable Means andStandard Deviations: Instrumental Network 162Table B.2: Non-Transition Respondent Sample Variable Means andStandard Deviations: Expressive Network 163Table B.3a: Bank Respondent Sample Variable Means andStandard Deviations: Instrumental Network 164Table B.3b: Forestry Respondent Sample Variable Means andStandard Deviations: Instrumental Network 1 65Table B.3c: Insurance Respondent Sample Variable Means andStandard Deviations: Instrumental Network 166Table B.4a: Bank Respondent Sample Variable Means andStandard Deviations: Expressive Network 167Table B.4b: Forestry Respondent Sample Variable Means andStandard Deviations: Expressive Network 168Table B.4c: Insurance Respondent Sample Variable Means andStandard Deviations: Expressive Network 169Table B.5: Description of Variables Labels for Tables B.6 through B.1 1 170Table B.6: Two-Tailed Spearman Rank Correlations: Instrumental NetworkDependent Variables 172Table B.7: Two-Tailed Spearman Rank Correlations: Expressive NetworkDependent Variables 173Table B.8: Two-Tailed Spearman Rank Correlations: Expressive Network andInstrumental Network Dependent Variables 175Table B.9: Two-Tailed Spearman Rank Correlations: Predictor Variables 176Table B.10: Two-Tailed Spearman Rank Correlations: Instrumental NetworkDependent Variables and Predictor Variables 180Table B.1 1: Two-Tailed Spearman Rank Correlations: Expressive NetworkDependent Variables and Predictor Variables 182Table B.1 2: Promotion Study Variable Means and Standard Deviations:Instrumental Networks 186Table B.13: Promotion Study Variable Means and Standard Deviations:Expressive Networks 1 87Table B.14: Kilmogorov-Smirnov (Lilliefors) Test for Normality:Instrumental Network and Expressive Network Dependent Variables 188viiTable C.1: Non-Transition Study Cross-Tabulation Percentages:Categorial Demographic Data 1 90Table C.2: Between Gender and Among Company Non-Transition StudySignificant Differences: Categorial Demographic Data 191Table C.3: Non-Transition Study Male-Female Means and Standard Deviations:Numerical Demographic Data 193Table C.4: Non-Transition Study Company Means and Standard Deviations:Numerical Demographic Data 1 94Table C.5: Promotion Study Cross-Tabulation Percentages:Categorial Demographic Data 196Table C.6: Promotion Study Male-Female Means and Standard Deviations:Numerical Demographic Data 197Table D.1: Description of Predictor Variable Labels for Tables D.2 through D.30 . 200Table D.2: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed 201Table D.3: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Males Listed 202Table D.4: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Females Listed 203Table D.5: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Same Site 204Table D.6: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Different Site/Same City 205Table D.7: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Different Site/Different City 206Table D.8: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Same Function 207Table D.9: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Different Functions Listed 208VIIITable D.1O: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Hierarchical Rank of Individuals Listed 209Table D.1 1: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Hierarchical Range of Individuals Listed 210Table D.12: Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Density 211Table D. 1 3: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed 212Table D.14: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Males Listed 213Table D.15: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Females Listed 214Table D.16: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Overlapping Expressive and Instrumental Ties 21 5Table D. 1 7: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Overlapping Female Ties 216Table D. 1 8: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Seen Outside of Work 217Table D. 1 9: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Females Seen Outside of Work 218Table D.2O: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Same Site 219Table D.21: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Different Site/Same City 220Table D.22: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Different Site/Different City 221ixTable D.23: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Same Function 222Table D.24: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Different Functions Listed 223Table D.25: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Hierarchical Rank of Individuals Listed 224Table D.26: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Hierarchical Range of Individuals Listed 225Table D.27: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Supervisors Listed 226Table D.28: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Subordinates Listed 227Table D.29: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Density 228Table D.30: Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Frequency of Contact 229xLIST OF FIGURESFigure 3.1: Non-Transition/Instrumental Network Model of Variables . . 24Figure 3.2: Promotion/Instrumental Network Model of Variables 28Figure 3.3: Non-Transition/Expressive Network Model of Variables 29Figure 3.4: Promotion/Expressive Network Model of Variables 31Figure 3.5: Hypothesis 2 Model of Variables 32Figure 6.1: Non-Transition/Instrumental Network Model 66Figure 6.2: Non-Transition/Expressive Network Model 66xiACKNOWLEDGEMENTSOver two years ago, I envisioned a doctoral thesis that studied both personalnetworks and social support. In the end, this thesis focused solely on personal worknetworks, and ironically, I learned far more about social support by doing the thesis than Icould have ever learned from studying social support. What follows is my attempt to thankthose individuals who taught me so much about social support.My family. Though my parents (Wayne and Jackie), my sister (Deb), and mygrandparents (Ada and Bernard) provided a myriad of appreciated (and timely) tangiblesupport, I am indebted to them for all of the support that went unnoticed (andunacknowledged) while I was consumed with my studies.My supervisory committee. Craig Pinder, Nancy Langton, Bonita Long, and GordonWalter were always generous with their time, constructive in their criticism, and challengingwith their questions. I do not think I could have envisioned or assembled a bettercommittee. Period.I owe much to my advisor (“What is your name again?”) that mere words cannotbegin to convey. At various times, Craig supported me financially with his SSHRC grant(#5-57312) or UBC-HSS research grant (#5-71247). Craig allowed me to not only teachhim a new topic so that I could conduct this study, but he graciously suffered throughnumerous intrusions during his sabbatical. More importantly, though, Craig recognized howimportant teaching was to me (and ultimately to my motivation) and allowed me to continueteaching, knowing full well it would slow my progress on this thesis.A special “thanks” also goes to Nancy who first suggested the topic of networks forone of my comprehensive examination papers.The Dean and Data Site Contacts. This study would not have been possible withoutthe help of Dean Michael Goldberg, who lent his name as I tried to secure data sites, and ofmy data site contacts (Betti Clipsham, Barbara Hislop, Alexis Freidman, Brian King, BrentWhite, Jim Whyte, and Debbie Woodward), who tolerated my repeated requests for theirrespective study participants lists.Several Professors, Staff Members, and Fellow Ph.D. Students. Dr. Tom Knightgraciously provided me with summer employment and the use of the Centre for Labour andManagement Studies’ computer. Dr. Larry Pinfield of Simon Fraser University played apivotal role in helping me focus my thesis. Discussions with Dr. Keith Head improved mystatistical knowledge, and various conversations with fellow Ph.D. students Karen Harlos,Vivien Clark, JoAnn Carmin, and Dafna Eylon improved my thought processes. (Vivien wasalso my proofreader extraordinaire.) Irene Khoo, the IRM Divisional Secretary, and RubyVisser, the Ph.D. Program Administrator, were available day in and day out to handlevarious word processing and bureaucratic crises, respectively. Finally, Dr. Alan Kraus, theCommerce Faculty’s Ph.D. Advisor, convinced the Faculty of Graduate Studies not to pullthe plug on my candidacy in the summer of 1 994 because I wanted to teach again.My Friends. Last but definitely not least, I thank my friends. (Many have beenlisted above.) In completing this thesis, I am in some ways now an expert on personalnetworks, and I take this opportunity to state that I have, without question, the bestnetwork of friends. I cannot possibly begin to thank each and every one of them (and Iapologize to those that go unmentioned) -- time and space are working as restraints.However, there are several people who have been instrumental in helping me maintain thatever elusive balance in life for which we all strive.Cathy DuBois and I made a pact one day far too many years ago to cultivate andnourish the right sides of our brains while we pursued our doctorates. I cannot speak forxiimyself, but Cathy has admirably fulfilled the intent of that pact. Professor David McPhillipsshared his insights on life and the golf swing at times when I needed advice the most.Finally, Andrea Dovichi, Susan Hargrave, Joshua Friedlander, Kevin Nelson, Randy Sidhu,Duane van Beek, Pat Whalen, and Ron Zayac willingly accepted my phone calls and/orrequests for lunch/dinner, a drink, a walk, a bike ride, etc., so that I could talk.In closing, I have chosen to dedicate this document to four friends who, in one wayor another, have been with me since this academic journey began. These four individualsare Larry Minner, Larry and Christie Schaffer, and the late Beverly Dahlgren. I would nothave a past, a present, or a future if they had not taken the time, in their own idiosyncraticand very special ways, to befriend me and enrich my life.XIIIChapter One:INTRODUCTION and LITERATURE REVIEWLittle is known about the similarities and differences between men’s and women’spersonal work networks, as there have been few in-depth explorations of the qualitativenature of work relationships, other than studies focusing on mentoring (Markiewicz &Devine, 1994). Ibarra (1992) argues that there is a need for empirical evidence and theorydevelopment in order to clarify the extent to which men’s and women’s personal worknetworks differ and the potential consequences of these differences. In particular,differences in work relationships and patterns can result in individuals’ experiencing adivergent amount of social support, which is inextricably linked to job proficiency (Pinder &Schroeder, 1987).A career is not a static endeavor. Individuals do not remain inert (i.e., in the sameposition, in the same location, and/or with the same employer) throughout their work lives.Instead, people experience career transitions that can lead to changes in their relationshipsat work. Understanding the similarities and differences between male and female personalwork networks is essential if we are going to appreciate fully the role that social support(e.g., work-related information, career-related information, and/or advice) plays in helpingindividuals achieve and maintain high levels of proficiency in their jobs, particularly aftercareer transitions. Work relationships may provide individuals with the social support that isnecessary to help them cope with and adapt to the various challenges that invariably ariseon the job.This thesis focuses on men’s and women’s personal work networks. There aretwo primary research questions of interest. The first is: To what extent are men’s andwomen’s personal work networks similar and/or different? The second query is: Whatimpact do promotions have on the similarities and differences between men’s and women’spersonal work networks?NETWORKS DEFINEDNetwork research focuses on either personal networks or social networks, and thedistinction between the two is important. In this thesis, the word “network” refers to thestudy of an individual’s personal network. Aldrich (1989) defines a personal network asthose persons with whom an individual has direct relations. A personal network isconstructed from the viewpoint of a particular individual and can involve relationships withone’s work, family, and/or friendship circles.A social network, on the other hand, is defined as a “specific set of linkages amonga defined set of persons, with the additional property that the characteristics of theselinkages as a whole may be used to interpret the social behavior of the persons involved”(Mitchell, 1969, p. 2). Social network analysis is the study of linkages (or ties) among adefined set of individuals. With social network analysis, one can uncover systematicdifferences in the ways in which men and women are located in an organizational context(Ibarra, 1992). When studying social networks, of interest are issues such as how thevarious ties cluster among themselves, how dense these clusters are, and which individualsare the most central in relation to the defined set of persons.The study of personal networks, unlike that of social networks, does not begin witha pre-defined set of individuals. Instead, a focal individual is identified and the direct ties ofthe focal individual are then identified for a given context, such as work, family, and/orfriends. The personal networks of various individuals are then compared according tocharacteristics, such as the size of the network, the gender mix of the network, and thefrequency of contact between the ties.1 The present study focuses on the similarities anddifferences between men’s and women’s personal work networks.1Personal networks are comprised of ties (or relationships) between the focal individualand those individuals with whom he/she associates. The terms “ties’ and “relationships”are used interchangeably in this thesis.2RELATIONSHIPS AND THE CORRELATION BETWEEN TIME AND PROFICIENCYPersonal networks are important to an individual’s daily existence because humanbeings are social animals (see Brett, 1984; Lynch, 1977; Spierer, 1981). Much of theresearch to date on relationships has shown that individuals need to associate with othersin order to cope with stress. Not surprisingly, the majority of research on social support isrooted in the stress literature (see Fenlason & Beehr, 1 994; Ganster, Fusiler, & Mayes,1986; House, 1981; Kaufman & Beehr, 1986; LaRocca, House, & French, 1980; Ullah,Banks, & Warr, 1985), where social support is hypothesized to either buffer or directlyimpact the individual’s level of stress.2Career transitions often impact personal work networks, and changes to work-related responsibilities and personal work networks have implications for how individualsadapt during the transition period (Sollie & Fischer, 1988). For example, consider the stressinvolved in learning a new job following a career transition.3 It is important that individualslearn the “ropes’ of their new jobs. The development of new relationships and themaintenance of old relationships in the workplace can have a major impact on the timerequired for an individual to become proficient in his or her job (Kaplan, 1984; Pinder &Schroeder, 1987).Becoming and remaining job proficient is a primary concern to both the individual2UIlah, Banks, and Warr (1985, p. 284) summarize the buffering and direct effectshypotheses:The ‘stress-buffering’ model of social support posits an interaction betweennegative life events and support, such that the beneficial effects of support areonly apparent during stressful life events; during periods when environmentalstress is absent, social support is assumed to have no impact... An alternative,the ‘direct effects’ or ‘independent effects’ model, states that social support canitself promote good health, both in the absence and in the presence of stressfullife events. This model predicts an overall effect of support on psychologicalhealth, rather than the interaction with life events which is predicted by thestress-buffering model.3The range of possible career transitions is best highlighted by Louis (1980) whoclassified career transitions into five different types: (1) re-entry or entry into the workforce, (2) taking a different role in the same organization, (3) moving to anotherorganization, (4) changing professions, and (5) leaving the labor pool.3and the organization. The organization wants individuals who learn to perform the job atthe highest possible level in the least amount of time, whereas individuals invariably want togain confidence in their ability to do their jobs -- in the least amount of time. This increasein confidence can have a far-reaching and positive effect on one’s levels of self-efficacy andself-esteem in the workplace. Workplace relationships are central to obtaining socialsupport (i.e., direct guidance, non-directive support, positive social interaction, and tangibleassistance; Barrera & Ainlay, 1983), and it is social support which, among other things,contributes to one’s ability to learn the new job and deal with difficulties as they arise(Pinder & Schroeder, 1987).TYPES OF WORK RELATIONSHIPSTwo distinct types of personal work relationships are defined in the existingliterature. Ibarra (1992, 1993a) most recently, for example, advances the distinctionbetween instrumental and expressive ties. She bases this distinction on the work of Lincolnand Miller (1979), Blau (1955), and Gouldner (1954) who wrote about instrumental andprimary ties. Instrumental ties are work contacts that aid and/or are necessary for theindividual to perform the tasks associated with his or her job. Instrumental ties involve theexchange of job-related resources, such as information, expertise, professional advice,political access, and material resources (Ibarra, 1 993a).Expressive ties are relationships with friends (Ibarra, 1 993a). Krackhardt (1992)writes that expressive relationships involve the exchange of friendship and social supportand are characterized by higher levels of closeness and trust than are those that areexclusively instrumental. Friendships can be seen as relationships that exist primarily forthe individuals’ personal satisfaction and enjoyment rather than for the fulfilment of aparticular task or goal (Sapadin, 1988; Wiseman, 1986). However, both expressive andinstrumental ties may enhance or impede the attainment of formal organizational goals(Lincoln & Miller, 1979). Moreover, ties can be instrumental and expressive at the sametime.4Nieva and Gutek (1981) emphasize that both instrumental and expressive ties areimportant to the individual in that a person can receive support -- crucial to proficiency --from anyone within or outside the work organization, regardless of whether the interactionsare prescribed or emergent in nature. It is easy to overlook the importance of friends atwork and the role they play in one’s ability to perform well. However, informal interactionsare ... critical to success at work because so much of the true requirements of the worksituation are not codified into formal rules and regulations” (Nieva & Gutek, 1981, p. 51).Also, these informal interactions within an expressive dyad are “systems for makingdecisions, mobilizing resources, concealing or transmitting information, and performingother functions closely allied with work behavior” (Lincoln & Miller, 1979, p. 179).Finally, instrumental relationships do not necessarily develop immediately after acareer transition, though many instrumental ties are prescribed within the organization bytask interdependence. The individual does not always step into his or her new role with fullknowledge about who will be instrumental to his or her performance. Such knowledgedevelops over time. Therefore, individuals who quickly develop these instrumental ties maybecome proficient at their respective jobs more quickly as well.Summary. Work relationships continually develop and change. Career transitionsdisrupt relationships, and the newly promoted or transferred individual is faced withdeveloping or maintaining relationships that require formal interaction (i.e., instrumental ties)as well as informal interaction (i.e., expressive ties) in order to become, and stay, proficientin a new job.PERSONAL WORK NETWORKS AND GENDERQuestions remain as to the similarities and differences between male and femalepersonal work networks and how men and women go about building and maintaining theirrespective work networks following career transitions. Although there is a substantialamount of research on gender differences, there is a dearth of theoretical and empiricalstudies on personal work networks and on gender differences particular to personal work5networks. A literature search produced an extensive list of studies on gender differences inrelation to work stress, leadership, depression, burnout, anger, adjustment to retirement,social support, mentoring, the effect of unemployment (e.g., psychological distress),perceived work competency, friendships in later life, job satisfaction, and creativity.However, there is scant research on gender network differences in the workplace, althoughmore is known about gender relationship differences in people’s overall personal networks(comprised of kin, friends, and co-workers). This research suggests that men and womenmove in different “relationship” worlds (Moore, 1990; Vaux, 1985). Studies of personalnetworks have found that women and men usually have networks of similar size (Fischer,1 982; Marsden, 1 987); however, when compared to men, women have fewer ties tononkin and more ties to kin, whereas men include more coworkers in their networks(Fischer & Oliker, 1983; Marsden, 1987; Moore, 1990; Wellman, 1985).Theoretical Explanations for Gender Personal Network DifferencesMoore (1990) argues that the differences between males and females are due tostructural, rather than dispositional, factors. The structural perspective explains genderrelationship differences by reference to opportunities and constraints arising out of women’sand men’s different locations in the social structure. The dispositional perspective reflectsgender differences resulting from, for example, differing traits or leadership and decision-making styles (see George, 1991, 1993; Judge, 1992). lbarra (1992, 1993a) and Ely(1994) address the importance of the structural and dispositional distinction in their workon workplace relationships.Ely (1994, pp. 227-228) suggests that her analysis “demonstrates how structuralfeatures of a firm may affect the nature and quality of interpersonal relationships at work,casting doubt on wholly person-centered explanations for the difficulties often observed inworkplace relationships among women.” Ibarra (1993a) argues for the integration andempirical testing of these two complementary perspectives, and her study of homophilousties exemplifies the dispositional perspective. Ibarra (1 993a, p. 423) states that“...explanations based on the notion of preference of homophily ...can be tested and6extended by explicitly taking into account structural constraints on preferences.”Without specifically addressing the dispositional and structural perspectives,Sapadin (1988, p. 388) summarizes the research on men’s and women’s friendships:studies have been remarkable in the similarity of their findings. Theyreport that female friendships involve more confiding, intimacy, personalconcern and emotional interactions than do male friendships. Thecommunication in women’s friendships is more empathetic andnurturing; interaction is more dyadic than group oriented. Malefriendships, in contrast, tend to be more group oriented. Males get closeby doing things together and showing enthusiasm for shared activities.Communication is more guarded and less self-disclosing about personalthoughts and emotions. Men’s interactions are more aggressive,competitive and oriented toward exchange of external information suchas sports and work interests (e.g., Dickens & Perlman, 1981; Fischer &Narus, 1981; Reisman, 1981).In short, Sapadin contends that women and men differ in the nature of their relationshipsand how their relationships develop. Women emphasize talking, emotional sharing, anddiscussing of personal problems with same-sex friends (Aukett, Ritchie, & Mill, 1988), andthese relationships depend on emotional closeness (Bell, 1991). Men tend to emphasizeshared activities, doing things with their male friends (Aukett et al., 1988) that correspondto an emphasis on group membership (Bell, 1991). Thus, men’s ties are more instrumental,whereas women’s are more expressive (Sollie & Fischer, 1988; Vaux, 1985). However, towhat extent these gender differences are due to structural and/or dispositional differencesremains unresolved.Moreover, differences in gender communication patterns are key to relationshipdevelopment and maintenance, and Tannen’s (1986, 1990) research highlights why womenand men seemingly move in different relationship worlds. Tannen indicates that, from birth,men and women are treated and spoken to differently.4 As a result, they end up talkingdifferently and moving in different worlds. For example, it has been only in the past fewdecades that questions have been raised about channelling girls into arts and boys intoscience (Hare-Mustin & Maracek, 1990a). Tannen notes that, from ages 5 to 15, children4lbarra (1993b) refers to this phenomenon as gender socialization.7play mostly with friends of their own sex (see Maltz & Borker, 1982), and in doing so,children learn how to have conversations and develop habits that continue into adulthood.For example, men’s conversations simulate negotiations and seem to be attempts toachieve and maintain status, whereas women’s conversations are more like negotiations forcloseness, confirmation, support, and/or consensus (Tannen, 1986, 1990).Summary. Women seem to have more intense, long-term, but fewer friends,whereas men tend to have more friends and less intimacy (Roberto & Kimboko, 1989). Theopportunity for men to remain intimate with their friends is reduced, on average, as thenumber of friends in their personal networks increases. This is not necessarily a problemfor men. Granovetter (1973, 1982) suggests that weak ties can extend a person’s accessto diverse social circles, thus increasing the social support (e.g., information) available to anindividual. Both the structural and dispositional perspectives (and how these perspectivesimpact communication patterns and gender socialization) provide plausible explanations forgender differences in personal networks (Ibarra, 1993a); however, there is currently notheoretical model (or models) to guide personal network research.5Presumed Gender Differences in Personal Work NetworksMoore (1990) has concluded that women’s personal networks will become more likemen’s (in regards to kin versus nonkin composition) when they move into paid employment.In line with Blau (1988) and Kanter (1977), Moore (1990, p. 734) indicates that:If men and women were in similar social structural positions their5Aldrich (1994) acknowledged the lack of a clear, definitive theoretical perspective toguide researchers in the study of how and why men’s and women’s personal worknetworks may differ. He started his talk by asking the audience two questions: (1) If webelieve that men’s and women’s personal work networks are similar, what has led us to thisbelief?, and (2) If we assume that men’s and women’s personal work networks differ, whathas led us to this assumption? Audience members listed socialization differences,differential access, genetics and personality differences, social comparison theory, anddiffering role explanations in response to why men’s and women’s personal work networkswould differ. Conversely, the same group noted that the lack of constraints, changingcorporate (and business school) cultures, similar work requirements, genetics (i.e.,distribution of male and female talent would be spread among companies), and externalcompetition (for the best employees) would lead to similar personal work networks.8behavior would differ little. As more women move into paidemployment, the genders’ network composition can be expected to bemore alike, with more close ties to non-kin, especially coworkers, andfewer ties to kin.However, questions remain as to whether men’s and women’s personal work networksdiffer on other network characteristics (e.g., gender composition), and if so, are networkcharacteristic differences also evident in men’s and women’s personal work networks?Researchers (e.g., Kanter, 1 977; Moore, 1 990) have intimated that the interaction(or relationship) patterns of men and women differ in the workplace, which may explainwhy women have not enjoyed the same level of success in career advancement anddevelopment as men. Starting with Kanter (1977) and continuing with the recent work ofDreher and Ash (1990), attention has focused on explaining women’s inability to break the“glass ceiling” and move up into the more senior positions in work organizations. BothKanter and Moore argue that one reason for women’s inability to break through to the upperlevels of management lies in their networks, which are supposedly different from men’snetworks.A logical conclusion from this assumption is that women need to modify theirinteraction patterns to mimic the interaction patterns of men in order to “succeed” (i.e.,advance) in organizations. Yet, little empirical research has been published that specificallyaddresses the similarities and differences between men’s and women’s relationships at theworkplace. Hare-Mustin and Marecek (1990a, p. 9) write that “... some socialpsychologists have pointed out that the perception of differences between men and womenhave been far greater than findings on the differences themselves.” Consequently, there isa need for additional empirical evidence and theoretical development to clarify the ways by,and extent to which, men’s and women’s work networks differ (Ibarra, 1 992) beforeattention shifts to the reasons and potential consequences of any observed differences.9NETWORK STUDIES IN WORK ORGANIZATIONSIn one of the few empirical studies on personal work networks, Ibarra (1993b)specifically considered gender differences in managerial networks. She concluded that,although gender and advancement opportunity did not account for observed differences inwork networks, men and women did use different approaches to derive similar networkbenefits at work (e.g., career-related support). High-potential women tended to have ahigher incidence of very close, relationship-focused ties in comparison to high-potential men(Ibarra, 1993b). In studying 63 middle managers, each working for one of four companies,Ibarra (1993b) also found that men’s and women’s workplace relationships did notseemingly differ in relation to their networks’ size, range, and multiplexity.6 However, inreporting that men’s and women’s workplace relationships did not necessarily differ, Ibarra(1993b) lumped the ties together into one large network and did not segregate the tiesbetween the support domains or the instrumental-expressive dichotomy. Not segregatingties, especially into the instrumental and expressive domains, may have masked the realsimilarities and differences between men’s and women’s workplace networks.With the exception of Ibarra’s (1993b) study, there is little research that looksspecifically at the differences in work relationships between men and women, except for apreponderance of same-sex relationships (or degree of homophily) in the work setting.Same-sex relationship research in organizations, which provides support for Tannen’sresearch on male and female communication patterns, dates back to the work of Kanter(1977). Kanter states that the majority of a manager’s time is spent communicating, andthe communication has to be rapid and accurate; therefore, there is a need for a common6The “size” of a person’s network is derived by the number of individuals listed by thefocal individual.“Range” refers to the degree of diversity of individuals listed in a personal network. Forexample, individuals could differ on the basis of function (i.e., are the individuals listed inthe same functional department as the focal individual?), position (i.e., are the individuals atthe same hierarchical level or a higher or lower hierarchical level as the focal individuals?),and/or location (do the individuals work at the same location as the focal individual?).“Multiplexity” is the degree to which a relationship is multi-dimensional (e.g., amultiplex tie would serve both an instrumental and expressive role).10language that is easily understood and predictable. In effect, there is a need forcommunication homogeneity in organizations; however, developing communicationhomogeneity is easier for men because it is men who have in the past tended to fill themajority of managerial and supervisory positions.Although women generally find it easier to associate and communicate with females(Kanter, 1 977; Tannen, 1 986, 1 990), the majority of their instrumental relationships mustinvolve men, making these relationships heterophilous. Ibarra (1992) studied workrelationship homophily in one organization and concluded that men have homophilousrelationships across their instrumental and expressive networks, whereas women exhibiteda different pattern. In an attempt to obtain greater access to both expressive andinstrumental ties, women must differentiate their relationship patterns by developingfriendships with other women and instrumental relationships with men. Additionally,Lincoln and Miller (1979) found that sex and race appear to have a greater influence on thedevelopment of expressive ties in the workplace than on instrumental ties.Ibarra (1993a, pp. 68-69) maintains that homophilous ties tend to be stronger thanheterophilous ties, and concludes:In sum, demography of the average American corporation is such thathomophilous ties are less available, have less instrumental value, andrequire more time and effort to maintain (due to dispersion and turnover)for women and minorities than for their white male counterparts.Consequently, women must work harder than men at developing and maintaininginstrumental ties.Finally, the findings of Aldrich, Reese, and Dubini (1989) and Brass (1985) aresimilar to those of lbarra, and Lincoln and Miller (1979). Aldrich et al. (1989) found thatthe personal networks of male and female entrepreneurs are, for the most part,homophilous. Brass (1985) found that women are not well-integrated into men’s networks-- especially in organizations’ dominant coalitions. Consequently, women are seemingly tiedto fewer influential men. One may conclude that even if women do develop workrelationships with influential men, these ties are inherently weaker because they are11heterophilous in nature (Ibarra, 1993).Because of women’s presumed exclusion from men’s personal work networks ingeneral, some people have concluded from Brass’ work that women are not good atbuilding informal networks. To the contrary, Brass found that women can be good atbuilding informal networks with other women and men can be good at building informalnetworks with other men (Brass, 1985). Yet, the fact remains that men retain the power inmost organizations.The relevant personal work network studies are summarized in Table 1 .1 regardingtheir samples, data collection methods, and methods for deriving the individuals’ networks.For example, Brass studied individuals in one organization (a newspaper), and though hisfindings relate to personal networks, he was more interested in the organization’s socialnetwork structure (i.e., the centrality and criticality7of individuals within the organization).On the other hand, Aldrich et al. (1989) studied entrepreneurs’ relationships outside of theirrespective organizations. These relationships included bankers, lawyers, accountants,consultants, and whomever else the respondents discussed plans for their business. Moore(1990) used the data gathered by the General Social Survey, which was comprised of1,534 English-speaking Americans. In the General Social Survey, respondents were askedto list up to five individuals with whom they discussed important matters. Theserelationships were designated as either kin or nonkin relationships.Summary. The studies by Ibarra (1992, 1993b) and Lincoln and Miller (1979) arethe only true, within-organization personal work network studies, and these studies providelittle cumulative information about differences in personal work networks. Ibarra’s 1 993b8study is the most informative in regards to men’s and women’s personal work7Centrality and criticality are social network characteristics, not personal work networkcharacteristics. Centrality represents the ease of access an individual has to others who arelinked to the focal individual either directly or indirectly. Criticality reflects the extent towhich a focal person controls the workflow (i.e., are there alternative routes through whichwork might flow if the focal individual is removed?).81 did not have access to this paper until after my thesis study was designed and datacollection had started.12- CA)TABLE1.1SummaryofRelevantPersonalWorkNetworkEmpiricalStudiesStudySampleSourceSampleDataCollectionMethodDerivationofNetworksDependentVariablesIbarra,FourFortune!63middleSociometricquestionnaire;NamepeopleintheorganizationRange1993bService500managerssemi-structuredinterviewsonthebasisoffivesupportConstraint1companiesdomains:information,advice,Multiplexityfriendship,career,cooperationHomophilyTiestrength2lbarra,Onepublic94full-timeSociometricquestionnaireNamepeopleintheorganizationCentrality1992relationsfirmemployeesonthebasisoffivesupportHomophilydomains:influence,advice,Multiplexitycommunication,influence,friendship.Therewasnolimitationofnominations;tenblanklineswereprovided.Moore,1985General1,534InterviewsNamepeoplewithwhomtheyRelationshipofindividuals1990SocialSurveyEnglish-havediscussed“importantlistedspeakingmatters”overthepastsixAmericans;months;85.6%namedbetween18yearsoldoneandfivepersonsandolderAldrich,Entrepreneurs264inNorthSurveyNamefivepersons(outsideofNetworkactivityReese,&inNorthCarolina,59thebusiness)withwhomtheyNetworkdiversity:CrossDubini,CarolinaandinItalyweremostlikelytoturnforsexties1989Italybusinessadvice;EstimationofNetworkdensity3numberofpeoplewithwhomtheydiscussedtheirbusinessplansBrass,Newspaper140non-Questionnaire;supplementedNameindividualscomprisingCentrality1985supervisorybydirectobservationandthreetypesofnetworks:Criticalityemployeesinterviewsconductedbeforeworkflow,communication,FrequencyofcontactwithquestionnaireadministrationfriendshipothersTABLE1.1(continued)SummaryofRelevantPersonalWorkNetworkEmpiricalStudiesStudySampleSourceSampleDataCollectionMethodDerivationofNetworksDependentVariablesLincoln&Threeresearch314QuestionnairesNamefivepersonsintheInstrumentalandprimaryMiller,organizations,respondentsorganizationwithwhomtheynetworkcomposition1979onehaveworkedcloselyinthepastrehabilitationmonthandfivepersonsinthecenter,andorganizationwithwhomtheyoneclinicwereclosefriends‘Constraintrepresentswhetherthefocalindividualhasmany,nonredundant,easilyreplaceablenetworkties(Ibarra,1 993b).2Tiestrengthisestablishedonthebasisof:(1)thefrequencyofcontact,(2)theemotionalintensityandlevelofmutualconfiding,(3)thelevelofreciprocitybuiltintotherelationship,and/or(4)thenumberofrolesplayedbyindividuals(e.g.,instrumentalandexpressive)(Granovetter,1973;Krackhardt,1 992).3Densityisdefinedastheextenttowhichindividualslistedknowand/ similarities and differences; however, her study falls short for one specific reason:Ibarra did not apply the instrumental-expressive dichotomy in testing for men’s andwomen’s personal work network differences. Though there were no significant genderdifferences in her respondents’ personal work networks, gender differences in theirexpressive and instrumental networks may have been masked.Purpose of the Current StudyWe do not know the extent of similarities and differences between men’s andwomen’s personal work networks in regards to the size, range, density, and frequency ofcontact, because researchers: (a) have not studied the various personal workcharacteristics at the same time, (b) have severely limited the number of individuals arespondent could list (e.g., Lincoln & Miller, 1979; Moore, 1990); (c) have studied oneorganization in isolation, thus limiting the generalizability of their findings (e.g., Brass, 1985;Ibarra, 1 992), and/or (d) have not utilized the instrumental-expressive dichotomy (e.g.,Ibarra, 1992, 1993b). Consequently, there remains the opportunity to study in multipleorganizations the personal work network characteristics of men and women. Researchshould first focus on how men’s and women’s personal work networks differ and shouldestablish the existence or non-existence of differences in men’s and women’s personalwork networks before researchers consider: (a) what conditions or dispositions create andreinforce gender network differences (if differences do, in fact, exist), and (b) the impact ofpersonal work networks on social support, job proficiency and promotion opportunities, forexample. The focus of this thesis is on the similarities and differences in men’s andwomen’s instrumental and expressive personal work networks.15Chapter Two:RESEARCH QUESTIONS and HYPOTHESESThis thesis entails two studies. The first study explores the similarities anddifferences between men’s and women’s personal work networks. The second studyexplores the fluidity of men’s and women’s work networks following promotions. Atpresent, there is scant network research on women in organizations (Ibarra, 1992) as wellas a lack of information concerning the way by which personal networks develop naturallyand change over time (Hays & Oxley, 1986). This limited network research coincides withthe paucity of studies that have specifically explored individual personal work networks(Ibarra, 1992, 1993a; Markiewicz & Devine, 1994).THE SIMILARITIES AND DIFFERENCES OFMEN’S AND WOMEN’S PERSONAL WORK NETWORKSThe research on male and female personal networks consistently finds that womenhave stronger expressive ties and seek closeness, men whereas have stronger instrumentalties and seek group membership through activities and shared experiences (Bell, 1991;Sollie & Fischer, 1988). In effect, women are socialized to develop emotionally-basedrelationships while social norms tend to inhibit male closeness (Roberto & Kimboko, 1989).A global stereotype has arisen that in modern society, including at work, the masculine roleis predominantly an instrumental one and men are seen as independent, competent, andrational, whereas the feminine role is predominantly expressive and women are seen assupportive, warm, and compassionate (Vaux, 1985).Moreover, research to date has provided us with a dichotomy, in that men’s andwomen’s work networks can be divided into two segments involving instrumental ties andexpressive ties. Ibarra (1993a, p. 79) states that instrumental and expressive networkrelationships affect ‘... the broader structural features of an individual’s personal networkand each is indicative of degrees of access to organization-wide instrumental and friendship16Instrumental TiesIn order to effectively and efficiently complete the tasks associated with their jobs,men and women ought to have instrumental networks that are relatively similar. When menand women occupy similar positions, their job responsibilities and requirements wouldrealistically result in their working with individuals with comparable hierarchical ranks andfunctional designations. If this does not occur, one sex could be at a disadvantage in thesocial support received. This could have an adverse impact on the level of proficiency anindividual achieves and maintains at his or her job because not knowing the “right” peoplecould ultimately hamper his or her performance. Unfortunately, except for the work onhomophilous ties by Ibarra (1992, 1993b) and Lincoln and Miller (1979), there is no extantresearch with which to offer formal hypotheses on dependent variables, such as size, range,and density. Therefore, this thesis will provide the first empirical test to the followingresearch question:RQ1 a. Do the instrumental work networks of men differ from theinstrumental work networks of women?Expressive TiesIf there are differences between men’s and women’s work networks, thesedifferences will more than likely be evident in their respective expressive networks.Because women are socialized to develop more dyadic, emotional relationships, womenwould seemingly be better at developing and maintaining strong expressive ties. On theother hand, men’s ties are assumed to develop through activities, and consequently, men’sexpressive ties may be the result of their work roles. Evidence of men’s developingexpressive ties from their instrumental ties would be exhibited by the extent to which theirinstrumental and expressive ties overlap, thus creating a multiplex tie. On the other hand,lbarra (1993a) writes that, until women comprise a higher percentage of management orsupervisory positions, they will invariably have to look outside their work activities in orderto develop expressive relationships with other women.17Where in the organization men and women look (or are able to look)1 to developworkplace friendships would seemingly impact the expressive relationships they develop.Consequently, the expressive networks of men and women may differ on a number of keydependent variables, including: number of expressive ties, gender of ties, overlap of tiesbetween the instrumental and expressive networks, range (both in terms of function andposition), and density. I hypothesize:Hia. The expressive work networks of men differ from the expressivework networks of women.MEN’S AND WOMEN’S WORK NETWORKS FOLLOWING PROMOTIONSFollowing a promotion, an individual has to cope with and adapt to a wide range ofpossible changes (i.e., a new work setting, new tasks, and/or new co-workers). One of themost pervasive changes involves the resulting transformations to an individual’srelationships. Kaplan (1984, pp. 50-51) writes:Every time managers change jobs... they must rebuild their networks... (and)general managers spend the first six months in a new job investing heavily informing new bonds. The more different the new job is from the manager’sprevious experience, the more overhauling the network will need.Although Kaplan does not specifically state this, one of the reasons for changes to anindividual’s relationships is that a new job presents individuals with new taskinterdependencies involving new people. Newly-promoted individuals must build workingrelationships with these co-workers if they are to succeed in their new positions. Bychanging jobs, a number of instrumental ties will no longer be useful or necessary and willmore than likely become latent -- regardless of whether one is male or female.Instrumental TiesA major task facing newly-promoted individuals is developing relationships with theindividuals with whom they share task interdependencies. Once individuals have had1The structure and demographics of an organization could hinder an individual’s ability todevelop friendships at the workplace.18sufficient time to rebuild their work relationships, the instrumental networks of men andwomen in comparable positions ought to be similar. Again, in order to get their jobs done,men and women would have relatively similar instrumental networks. If this is not thecase, then one sex may be at a disadvantage in achieving proficiency on the job. Theliterature provides little help in distinguishing the possible differences between men’s andwomen’s instrumental networks in relation to size, range, and density; consequently, thisthesis will provide the first empirical test to the following research question:RQ1 b. Do the instrumental work networks of men differ from theinstrumental work networks of women following a promotion?Expressive TiesKanter (1977) and Moore (1990) have intimated that women may need to mimic themale style of building work networks if they are to have the same career success as men.However, I have argued that men’s and women’s instrumental work networks should notnecessarily differ. That is, women need to create similar expressive work networks as men.The question is: Do women who enjoy career successes (i.e., promotions) have expressivework networks similar to those of their male counterparts? I argue no. For women to buildsimilar expressive work networks, women would need to develop networks that includemore males (i.e., less homophilous ties) that are less dyadic and more dense. Womenwould end up with expressive networks that included more superiors and subordinates andties with less functional range in their expressive networks.Therefore, I hypothesize:Hib. The expressive work networks of men differ from the expressivework networks of women following a promotion.Furthermore, two important changes to one’s expressive work networks can behypothesized to result following a promotion. First, men might experience more turnover22The term “turnover” was chosen to reflect the fluid nature of work networks.Following a career transition a personal work network undergoes change, as individualseither remain in the network or are dropped from it. At the same time, new individuals areadded to the personal network.19in their expressive work networks than women following promotions because men’sexpressive relationships may emanate from their instrumental relationships. One couldargue that men stand to lose a number of their expressive relationships because of theinstrumental-expressive network overlap. In effect, the activities associated with themaintenance of certain established expressive ties will most likely change. The expressivetie may become latent without the “instrumental activity providing the impetus for a givenrelationship. Women, who tend to separate their expressive and instrumental relationships(Ibarra, 1 993a), could maintain more of their expressive ties than could men.The second phenomenon involves women specifically. If there is any change inwomen’s expressive ties, it would be in terms of the positional range their expressive tiesexhibit.3 As women move up the organizational hierarchy, it is very likely that they couldand should develop more influential expressive relationships. Lincoln and Miller (1979) andMiller (1986) found that high-status individuals have more extensive network connections,thus linking these individuals to high-status people. I hypothesize:H2. Men experience more turnover in their expressive work networksthan do women following promotions.H3. The range (in terms of position) of women’s expressive networksprior to a promotion increases following a promotion.In summary, this thesis includes two research questions, two compositehypotheses, and two specific hypotheses. The research questions pertain to whetherdifferences exist between men’s and women’s instrumental work networks, whereas two ofthe four hypotheses specifically test for differences on multiple dependent variablesbetween men’s and women’s expressive work networks. The remaining two hypothesesfocus on two post-promotion outcomes -- (a) the turnover in men’s and women’sexpressive networks and (b) the increase in the positional range of women’s expressivework networks. The research models, outlining the variables associated with these3The same would not necessarily be true for women’s instrumental networks, given mycontention that men’s and women’s instrumental networks should be comparable.20research questions and hypotheses, are discussed in the next chapter.21Chapter Three:RESEARCH MODELSOn the basis of Research Questions la and lb and Hypotheses la and lb. Iexamined four research models. These models reflected the instrumental and expressivedichotomy and the gender differences within the non-transition and promotion samples.Table 3.1 outlines the four models. The research questions and hypotheses, with theexception of hypotheses 2 and 3, were composite in nature. The associated models forstudying the instrumental and expressive work networks of individuals, on the basis of thedependent variables, were intended to build evidence as to whether we could infer theexistence of substantial or minimal similarities or differences between the workplacenetworks of men and women.Non-Transition/Instrumental Network ModelTable 3.1 lists the variables associated with the non-transition/instrumental networkmodel. The dependent variables1 can be classified into three groups. The first grouprepresents the size of the network (i.e., the number of people listed) and the gender ofthese individuals. The second group involves the range of the people. Range is representedby the work location of the people, the number of people in the same function as the focalindividual, the number of different functions from which these individuals are drawn, andthe hierarchical rank and range of these people. The final dependent variable is the densityof the ties (i.e., the degree to which the individuals listed know and work with oneanother).Gender is the focal predictor variable in this thesis. However, gender is not the onlypredictor variable of interest. Other variables include a covariate, three plausible controlvariables, and a possible moderator variable.Initially, the number of people listed by each study participant is a dependent1The operationalizations of the dependent variables are discussed in Chapter 5: DataCollection.22TABLE 3.1Thesis Research ModelsModel Research Question or Hypothesis SampleNon-Transition! Do the instrumental work networks Managers or supervisorsInstrumental of men differ from the instrumental who have been in theirwork networks of women? positions at least oneyear.Non-Transition! The expressive work networks of Managers or supervisorsExpressive men differ from the expressive work who have been in theirnetworks of women. positions at least oneyear.Promotion! Do the instrumental work networks Individuals recentlyInstrumental of men differ from the instrumental networks of women followingpromotions?Promotion! The expressive work networks of Individuals recentlyExpressive men differ from the expressive work promoted.networks of women followingpromotions.23FIGURE 3.1Non-Transition/Instrumental Network Model of Variables1Predictor Variables Derendent VariablesCovariate: Instrumental Network Size andNumber of Individuals Listed Gender Mix:Number of Males ListedIndependent: Number of Females ListedGender of Study ParticipantRange:Control: Location of Individuals ListedJob Category Number of Individuals Listed in SameJob Level FunctionNumber of Different FunctionsModerator: Hierarchical Rank of IndividualsCompany ListedHierarchical Range of IndividualsListedDensity1The operationalizations of these variables and the variables listed in the figures that followthis one will be presented in Chapter Five.24variable; however, it should covary with every other dependent variable. The total numberof people listed by the focal person directly impacts, for example, the number of male andfemale individuals listed, the number of individuals at various locations, the number ofindividuals in the same function, and so on.Of the three possible control variables, the job category2and the job level3 of eachstudy participant’s position are the most salient. It is quite possible that males and femaleshave different job responsibilities at different hierarchical levels which could haveimplications for the networks they develop. This relates directly to the aforementionedstructural argument of Moore (1990). Possible differences include jobs involving expert oradministrative roles, as opposed to more managerial or decision-making roles (Dreher & Ash,1990; Kanter, 1977). Women often dominate (or are assumed to dominate) theadministrative roles, whereas men tend to dominate the decision-making or managerial roles(Dreher & Ash, 1990). Furthermore, fewer promotions occur within the expert oradministrative ranks, which Dreher and Ash (1990) argue would partially explain why menhave experienced more promotions than women. A priori, the existence of gendersegregation into jobs involving different roles and at different hierarchical levels in theorganization could result in different instrumental work networks; consequently, both thejob category and job level of the study participants were controlled, as shown in Figure 3.1.As for the other control variable, time in current position may impact how the worknetworks of individuals develop in the workplace. Moreover, it is conceivable that the timea person has been in his or her current position could impact “the cumulative knowledge”the person acquires about both the organization and the job. This knowledge could includehow his or her position fits in the established workflow of the organization, or who to go toto get help with a work task or work-related problem. Though intuitively important, time in2ln this thesis, respondents were asked whether their jobs were purely managerial,professional, technical, administrative in nature, or a mix of managerial and eitherprofessional, technical or administrative roles.3Job level relates to the position being designated as either executive, seniormanagement, middle management, first-line supervisor, or other.25current position is excluded from the model. To be included in the non-transition study,individuals had to have been in their positions at least one year, and consequently,individuals should have acquired the requisite knowledge about how their jobs fit into theworkflow patterns of the organization.Finally, there is the issue of the companies from which the sample is drawn.Intuitively, company would be cast as a moderator in that “company has no presumed apriori relationship to the dependent variables of interest.4 However, it is quite likely thatvarying company (or industry) characteristics could moderate the relationship betweengender and the work network characteristics of individuals, and therefore, the possibility ofa company moderator effect would need to be tested.5In summary, Figure 3.1 represents the non-transition/instrumental network model.The effect of gender, number of instrumental people listed, job category, job level, andpossible company moderator will be used to account for the variance associated with eachof the dependent variables.Promotion/Instrumental Network ModelThe promotion/instrumental network model is similar to the non-transition,instrumental network model -- with the exception of dropping company as a moderator4Lindley and Walker (1993) define a moderator effect as an interaction between apredictor variable and a moderator variable, such that the relationship between the predictor[Xl and an outcome variable [Yl differs depending upon the level of the moderator [ZI; theeffect of X on Y is conditional upon the level of Z. A moderator should not be correlatedwith either the outcome variable(s) or predictor variable(s).5lbarra (1993b, 1994) controlled for company-specific effects, arguing that the fourfirms included in the study differed on a variety of factors including industry, performance,and organizational culture. Three dummy variables were included to ensure that networkeffects could be observed net of company effects not measured directly.26variable because the sample for the promotion study will be drawn from one organization.A representation of the model can be found in Figure 3.2.6Non-Transition/Expressive Network ModelThe variables for the non-transition/expressive network model are listed in Figure3.3. The dependent variables include the same dependent variables as the non-transition,instrumental network model with a few additions. First, there are the total number ofpeople listed and the gender make-up of those people. In addition, there are the number ofexpressive ties that are also instrumental ties, which addresses the extent to which the tiesoverlap, along with the number of people who are seen socially outside of work and thenumber of women seen socially outside of the office. Second, the range variables remainthe same with the inclusion of the number of both subordinates and supervisors listed asexpressive ties. Density of the expressive ties is also considered, as is the frequency ofcontact of the focal individual with the people listed.Moreover, as with the instrumental network model, gender remains the focalindependent variable with number of expressive ties listed a covariate, job category and joblevel as control variables. Company serves as a moderator variable. Two additional controlvariables need to be considered. First, the individuals who were my data site contactsconsistently argued that child-rearing responsibilities play a pronounced role in influencingwho becomes and remains a friend. Child-rearing responsibilities can impact not only thenumber of expressive ties in one’s work network, but also the frequency of contact with6Two other control variables were considered for the promotion/instrumental ties model.These control variables included: (1) tenure with company and (2) number of previouspositions. The time an individual has been with an employer and the number of previouspromotions or positions with the employer are critical variables when we talk of rebuilding anetwork. This thesis is not concerned with how one goes about rebuilding, or how quicklythey go about rebuilding their network. The promotion sample’s instrumental networks willbe measured at the time of the promotion and at least eight months following thepromotion. It is conceivable that people with longer company tenures have moreknowledge about how the organization operates and could then apply this knowledge duringthe first few months in their new positions, thus helping them rebuild their instrumentalnetworks more quickly. However, the effect of company tenure and the number ofprevious positions should diminish greatly six months to a year after the promotion.27FIGURE 3.2Promotion/Instrumental Network Model of VariablesPredictor Variables Deiendent VariablesCovariate: Instrumental Network Size andNumber of Individuals Listed Gender Mix:Number of Males ListedIndependent: Number of Females ListedGender of Study ParticipantRange:Control: Location of Individuals ListedJob Category Number of Individuals Listed in SameJob Level FunctionNumber of Different FunctionsHierarchical Rank of IndividualsListedHierarchical Range of IndividualsListedDensity28FIGURE 3.3Non-Transition/Expressive Network Model of VariablesPredictor Variables Dependent VariablesCovariate: Expressive Network Size andNumber of Individuals Listed Gender Mix:Number of Males ListedIndependent: Number of Females ListedGender of Study Participant Number of Overlapping Instrumentaland Expressive TiesControl: Number of Overlapping InstrumentalJob Category and Expressive Female TiesJob Level Number of Individuals Listed SeenChild-rearing Responsibility Outside of WorkRace Number of Females Seen Outsideof WorkModerator:Company Range:Location of Individuals ListedNumber of Individuals in SameFunctionNumber of Different FunctionsHierarchical Rank of IndividualsListedHierarchical Range of IndividualsListedDensityFrequency of Contact29people--especially outside the workplace. lbarra (1993b, pp. 28-29) indicates thatsociologists.., have found that life-course factors such as child-bearing account forsignificant gender differences...” in personal networks. The second control variable is race.Like women, minorities have reduced opportunities to develop homophilous expressive tiesin comparison to their white male counterparts (Ibarra, 1993; Lincoln & Miller, 1979).In summary, the non-transition/expressive (see Figure 3.3) model is as follows: Theeffects of gender, number of individuals listed, job category, job level, child-rearingresponsibilities, race, and possible company moderator will be used to account for thevariance associated with each of the dependent variables.Promotion/Expressive Network ModelThe promotion/expressive network model (see Figure 3.4) is similar to the non-transition model -- with the exception of dropping company as a moderator variable. Thereason for this is that the sample for the promotion study was drawn from only oneorganization.Hypotheses 2 and 3The dependent variable for Hypothesis 2 is tie turnover, with gender theindependent variable. Company tenure must be added as a control variable to the model.Conceivably, those individuals with longer tenure have more time to develop and maintainexpressive ties that have weathered other career transitions within the company. If so,there is the chance that individuals with longer tenure have more stable expressive tienetworks than those with shorter tenure. Therefore, the turnover in expressive tienetworks may be reduced for those with longer company tenure. Hypothesis 2 isrepresented in Figure 3.5. Finally, Hypothesis 3 compares the hierarchical rank and rangeof the women’s expressive ties prior to the promotion and nine months following thepromotion.30FIGURE 3.4Promotion/Expressive Network Model of VariablesPredictor Variables DeDendent VariablesCovariate: Expressive Network Size andNumber of Individuals Listed Gender Mix:Number of Males ListedIndependent: Number of Females ListedGender of Study Participant Number of Overlapping Instrumentaland Expressive TiesControl: Number of Overlapping InstrumentalJob Category and Expressive Female TiesJob Level Number of Individuals Listed SeenChild-rearing Responsibility Outside of WorkRace Number of Females Seen Outsideof WorkRange:Location of Individuals ListedNumber of Individuals in SameFunctionNumber of Different FunctionsHierarchical Rank of IndividualsListedHierarchical Range of IndividualsListedDensityFrequency of Contact31FIGURE 3.5Hypothesis 2 Model of VariablesPredictor Variables Dependent VariablesIndependent: Expressive Tie Turnover:Gender of Study Participant Number of Individuals AddedNumber of Individuals DroppedControl: Number of Individuals RemainingTenure with Company32Chapter Four:POPULATIONS AND SAMPLESThere were two populations of interest for study in this thesis. Managers orsupervisors who had not undergone a formal career transition in the past year comprisedthe first population. This population was used to examine the similarities and differencesbetween men’s and women’s work networks. The second population was individuals whohad recently been promoted within a given organization; this population was used to studythe impact that promotions can have on men’s and women’s work networks.THE DATA SITESSixty major Canadian companies were contacted through letters written to thepresident/CEOs, human resource vice-presidents, or human resource managers. Theseletters1,outlining the thesis study, were signed by either Michael Goldberg, Dean of theFaculty of Commerce and Business Administration at the University of British Columbia, orby Dr. Craig Pinder, my thesis advisor. From these 60 letters, seven organizationsexpressed an interest in participating in the study. Four organizations were ultimatelyincluded. A fifth organization, a major utilities company, was dropped from the study fortwo reasons: (1) a major corporate reorganization was about to be announced and theemployees would have been quite suspicious of any company-approved questionnairerequesting information about their work relationships, and (2) several of the associations,which represented many of the company’s managers or supervisors, had instructed theirmembers not to fill out any company (or company-approved) questionnaires until furthernotice.For both the non-transition and promotion studies, the samples had to come fromorganizations with (a) a mix of men and women in managerial or supervisory positions and1Copies of these letters can be found in Appendix A.33(b) operating sites of at least 200 people. The reasoning behind this requirement was thatin smaller organizations, or in organizations with small operating sites, it is much easier forany particular employee to know practically “everyone.” The work networks of men andwomen in such organizations would be very similar, regardless of the sex of the individual.Furthermore, there would be little change to individuals’ work relationships following atransition in smaller organizations.Data collection consisted of administration of a questionnaire to the managers orsupervisors at three organizations2,representing the banking, forestry, and insuranceindustries. With the non-transition study, the original intent was to sample up to 100 or1 50 individuals from each of the four organizations while maintaining, if possible, a 50-50gender split. This was not the intention, however, with the promotion study. All thepromotion study participants came from the same organization to control for varyingdefinitions of what comprises a promotion. A review of the literature for a common,established definition of “promotion” was unsuccessful, and even the data site contactswere unable to provide their companies’ definitions of “promotion.” Consequently, onecompany was chosen that had enough promotions within a three-month period in order toprovide a sample of approximately 100 individuals. The banking company agreed to providea list of employees who had been promoted within the past two to three months. Again, a50-50 gender split was desired for the promotion study.Non-Transition SampleTable 4.1 provides the number of individuals receiving questionnaires in each of thethree organizations. All three organizations provided lists of managers or supervisors whohad not undergone formal career transitions in the past 1 2 months. When possible, nameswere randomly selected from the lists with the use of a random numbers table.All 203 people listed -- 1 33 from the insurance company’s headquarters and 702Data were also collected from a fourth organization in the telecommunications industry,but this company was dropped from the data analysis. Only 51 names were provided, andeven after a follow-up letter, only 14 questionnaires (from nine males and five females)were returned.34TABLE 4.1Male and Female Participation by Company in Non-Transition SampleCompany Males Females TotalBank Total 117 117 234Personal and Commercial1 74 76 1 50Investment and Corporate 43 41 84Forestry Total 107 38 145Headquarters 68 30 98Division A 14 3 17Division B2 25 5 30Insurance Total3 135 68 203Headquarters 85 48 133Claim Centers 50 20 7010f the 1 50 employees included in the study, 50 males and50 females were randomly selected from a list of 1 53 malesand 148 females, respectively. The remaining 50 names camefrom a separate list representing the bank’s largest branches.All 24 males and 26 of 51 randomly chosen females received aquestionnaire.2The 25 males include five men in similar positions as the fivefemales plus 20 randomly chosen males from a list of 148 men.3The male-female totals are crude estimates, as the listsprovided did not note the sex of the individual. The lists werescanned for possible female names.35from the company’s largest claim centers in the Lower Mainland of Vancouver, B.C. --received a questionnaire. Unfortunately, the lists provided did not note the sex of theindividuals. (The lists provided were scanned, and approximately 68 names appeared tobelong to females.)The forest company provided an original list with 521 names, representing 11different divisions. Only five of the 11 divisions had more than one female in a managerialor supervisory position, and only three divisions had more than two female managers orsupervisors meeting the established criteria. To ensure the highest participation rate offemales in relation to males, the company’s headquarters and the two divisions with morethan two females were selected to participate in the study.Finally, the study participant names from the bank were provided by two divisions --the Personal and Commercial Division and the Corporate and Investment Division.Originally, names were to come from the bank’s corporate and regional center sites with atleast 200 people; however, the two data site contacts argued that the bank was organizedinto regions and that the majority of banking managers and supervisors worked with othersthroughout the region and not necessarily with fellow employees at their actual work sites.In other words, bank employees developed instrumental and expressive ties within theseregions of well over 200 individuals.Table 4.2 provides the average number of ties listed by each study participant bycompany for the three site location designations and substantiates the claims of the twodata site contacts. The percentage of ties from different operating sites (in relation to thetotal number of ties) for jj instrumental and expressive ties was approximately doublethat for the banking study participants in comparison to the study participants from theforestry and insurance companies.Promotion SampleTo study the impact of formal career transitions on individual’s personal worknetworks, I chose to look at promotions for two reasons. The first was the “glass ceilingeffect,” and women’s alleged inability to break through to the upper levels of management.36TABLE 4.2Instrumental Network and Expressive Network Site Locationof Individuals Listed by CompanyBanking Forestry Insurance p-values1INSTRUMENTALSame Site 3.43 7.20 7.18 < .001Different Site/Same City 2.83 0.93 3.34 < .001Different Site/Different City 3.43 3.18 0.53 < .001Percentage at Different Site 65% 36% 35%EXPRESSIVESame Site/Same City 3.66 6.17 6.26 < .001Different Site/Same City 2.51 0.53 2.43 < .001Different Site/Different City 2.23 1 .30 0.67 < .001Percentage at Different Site 56% 23% 33%1The Kruskal-Wallis one-way analysis of variance was used to test forbetween-company differences.37Both Kanter (1977) and Moore (1990) have argued that one reason for women’s inability tobreak through to the upper levels of management lies in their networks, which aresupposedly different from men’s networks.The other reason involved my intent to study the most specific career transitionpossible, and with that in mind, this thesis was designed to study promotions in situ. It isobvious that a career transition, involving an individual’s transferring to another worklocation within the same company or moving to an entirely new organization, could result inmajor changes to one’s work network. However, the extent of changes to a networkfollowing a promotion in situ may not be as great. Though individuals may find themselvesworking with (and developing ties) with new individuals, they can still maintain contact withindividuals (especially expressive ties) because they are remaining at the same work site.The main reason for looking at promotions in situ was this: If there are extensive changesto one’s work network following a promotion in situ, then it would be quite likely there arechanges in one’s network following a promotion to other operating sites.Of the four participating organizations, the bank was the only organization thatcould provide a large enough number of employees who had been recently (within the pasttwo to three months) promoted; however, few of these promotions were in situ. Of the 33promotion respondents, only eight individuals remained at the same site following theirpromotions. Given the time constraints associated with completing a doctoral thesis andcollecting the promotion data3, it was not feasible to limit the study to only thoseindividuals who had received promotions in situ. Consequently, all 39 males and 61randomly chosen females (from a list of 123) in the personal and commercial divisionreceived a questionnaire. From the investment and corporate division list, all 31 males and31 females received a questionnaire. The lists provided by both divisions did not note3As will be discussed in Chapter 5, the promotion sample received two questionnaires.The first questionnaire was sent as closely to the time of their promotions as possible sothat a representation of their instrumental and expressive networks prior to the promotioncould be obtained. The second questionnaire followed approximately eight months later.The intent of the follow-up questionnaire was to obtain a representation of theirinstrumental and expressive networks following their promotions.38whether the promotions were or were not in situ.SummaryA total of 582 individuals comprised the non-transition sample and 162 individualsthe promotion sample. Though the names came from specified strata (e.g., managers orsupervisors who had not undergone career transitions in the past year, or newly promotedmanagers or supervisors), many were not randomly chosen as study participants.Consequently, the non-transition and promotion samples can be portrayed only asconvenience samples, and this places limits on the generalizability of the findings of boththe non-transition and promotion studies.39Chapter Five:DATA COLLECTIONData collection consisted of the administration of a sociometric questionnaire to thenon-transition and promotion samples.1 The non-transition sample received thequestionnaires through the respective companies’ internal mail systems, and each individualwas instructed to fill out the questionnaire and return it in a self-addressed stampedenvelope directly to the author. Participation was completely voluntary and confidentialitywas assured.2People in the promotion sample filled out two questionnaires. The firstquestionnaire was distributed through the company’s internal mail system to the recentlypromoted bank employees. The second questionnaire followed a minimum of eight monthsafter the promotion, and only those 43 individuals who had returned the first questionnairereceived the follow-up questionnaire. The participants had been told of the follow-upquestionnaire in the first questionnaire, and they filled out the follow-up questionnairewithout any access to information provided by them in the first questionnaire. Thequestionnaires were distributed through the company’s internal mail system and whencompleted were returned directly to the author. Participation was completely voluntary andconfidentiality was assured.The decision to use a questionnaire as the sole data collection method was based onthe goal to sample from multiple organizations as many managers or supervisors in as timelya fashion as possible. With the exception of Lincoln and Miller (1979), past research on1Questionnaires were sent out on the following dates: Forestry, February 8, 1994;Insurance, March 14, 1994; Mobile Data, April 7, 1994; Banking (both non-transition andpromotion), April 25, 1994. The promotion follow-up questionnaires were distributed onNovember 21, 1994.2Copies of the two cover letters included with the questionnaires are in Appendix A.The first letter was signed by a participating company’s representative (e.g., humanresources manager or vice president). The second letter was signed by my advisor, Dr.Craig Pinder.40personal work networks had been conducted in single organizations (Brass, 1986; Ibarra,1992). Recently, Ibarra (1993b, 1994) focused on individuals in four Fortune 500companies; however, the total number of study participants was only 63.THE QUESTIONNAIRESThe initial questionnaire (see Appendix A), received by both the non-transition andpromotion samples, consisted of four parts and was modeled after the grid format used byTracy and Whittaker (1990). The grid format had individuals list the names of people intheir networks before providing such information (regarding each individual listed) as thelength of time they have known each individual, how often he or she saw each individual,and the support provided by each person.The Non-Transition QuestionnairePart One of the questionnaire consisted of some preliminary questions regarding theparticipants’ work relationships. Participants provided the names of their supervisors aswell as the names of anyone in the organization who had served them in a mentor ordevelopmental capacity.3 Participants also noted the number of individuals they directlysupervised and the percentage of their subordinates who were male or female.In writing the questionnaire, I considered Part One a warm-up to Parts Two andThree. These two sections, requesting the participants to list up to 1 5 instrumental and 1 5expressive ties, provided the data directly related to the research questions and hypothesesof this study. The primary reason for having the study participants list 1 5 individuals forboth their instrumental and expressive work networks was to keep the time required tocomplete the questionnaire manageable (Tracy & Whittaker, 1990). Moreover, a betterestimate of the size of personal work networks would be calculated by having studyparticipants list 15 individuals, instead of five as did Lincoln and Miller (1979). In Part Two,3These data were collected for possible future study.41participants were requested to list up to 1 5 names of instrumental ties4 on the basis of thefollowing, explicit instructions:Before breaking the seal on this part of the questionnaire, please list up to 1 5individuals within your company whom you consider to be useful in performingthe tasks required of your current job.Task-related relationships include those in your organization who aid you and/orare necessary for you to perform the tasks required in your job. In other words,you rely on these individuals to help you do your job. (Please note that you mayinteract with these individuals on a daily basis or as infrequently as once a weekor month.) Moreover, these individuals provide you with job-related resources(e.g., information, expertise, professional advice, and/or material resources).Please exclude your manager(s)/supervisor(s) as well as all of your subordinatesfrom this list.Please break the seal on this part of the questionnaire upon completing your listand provide the requested information in relation to each individual you havelisted.Two critical points must be made regarding the instructions. First, in listing thepeople who comprised their instrumental ties, respondents were asked to exclude theirsuperior(s) and subordinates. The reasoning behind this request was that the respondentswere to look beyond their immediate supervisors and subordinates in relation toinstrumentality because superiors and subordinates would always be instrumental to themanager’s or supervisor’s performance. The second point involved the participants’generating their complete lists of individuals prior to breaking a seal that kept the instrumentclosed. This seal was included to prevent the participants from “looking ahead” beforehaving read the instructions. There was also the fear that respondents might knowinglysuppress the number of individuals listed if they first saw the amount of additionalinformation required for each tie they listed. Once the list of instrumental ties wascomplete, the seal5 was to be broken and the following information regarding each tie was4Asking the study participants to first list their instrumental ties followed the orderapplied by Lincoln and Miller (1979). Furthermore, in the questionnaire, the term“instrumental” was replaced with “task-related.” Though “instrumental” is the establishedconstruct in the literature, the term “task-related” was better understood by thoseindividuals who participated in the pilot testing of the questionnaire.5The seal did not prevent respondents from adding names afterwards, however.42to be provided:(1) sex;(2) time known;(3) the work location of the individual;(4) the individual’s occupational title;(5) whether the individual worked in the same functional department;(6) the department name of the individual if he/she worked in adifferent functional department;(7) the hierarchial position of the individual in relation to the studyparticipant; and(8) who else on the list did the participant and each tie associate with inorder to complete tasks associated with the respondent’s job.Oerationalizations of instrumental network derendent variables. The tieinformation entered in Part Two provided the dependent variable data for the instrumentalnetwork size and the gender mix, range, and density of these instrumental ties for eachrespondent. The operationalizations of these dependent variables can be found in Table5.1. The majority of the variable operationalizations are self-explanatory. With theexception of hierarchical rank, hierarchical range, and density, the actual values for thevariables were derived by simply counting, for example, the number of people listed by therespondent, the number of people working at the same site as the respondent, and thenumber of different functions (excluding the function of the respondent) represented by thepeople listed.The hierarchical rank value, as calculated, represents the extent to which the peoplewere at the same hierarchical level6 as the respondent. A value greater than zero reflects anetwork with people who, on average, are one or more hierarchical levels above therespondent. A negative value depicts a network in which the majority of ties emanate fromhierarchical levels below the respondent. Hierarchical range complements the hierarchicalrank value in that it represents the extent to which a respondent’s ties come from two ormore levels above to two or more levels below the respondent. A value of zero reflects anetwork where all the ties are at the same hierarchical level as the respondent. A further6Respondents were asked to consider whether each tie was: (1) at least two levelsabove, (2) one level above, (3) at the same level, (4) one level below, or (5) at least twolevels below him or her in the organization’s hierarchy.43TABLE 5.1Operationalizations of Dependent Variables: Instrumental NetworkActualRange CalculationNetwork Size and Gender Mix:Number of Individuals Listed’ 0 to 1 8 CountNumber of Males Listed 0 to 1 4 CountNumber of Females Listed 0 to 1 2 CountRange:Number of Individuals perLocationSame Site 0 to 1 7 CountSame City/Different Site 0 to 1 5 CountDifferent City/Different Site 0 to 1 5 CountNumber of Individuals Listed 0 to 1 5 Countin Same FunctionNumber of Different 0 to 1 2 CountFunctionsHierarchical Rank -2.0 to 2.0 ((2 * of individuals two or more levels(Average rank of individuals above) + (1 * # of individuals one levellisted) above) + (0 * # of individuals at same level)+ (-1 * # of individuals one level below) +(-2 * # of individuals two or more levelsbelow)) / (# of individuals listed)Hierarchical Range 0 to 2.0 ((2 * # of individuals two or more levels(Average range of individuals above) + (1 * # of individuals one levellisted) above) + (0 * of individuals at same level)+ (1 * of individuals one level below) +(2 * # of individuals two or more levelsbelow)) I (# of individuals listed)Density 0.0 to 1 .0 (# of dyadic links listed) I ((number ofindividuals listed)(1 - number of individualslisted))1 Several respondents listed more than 1 5 individuals.44illustration could find a respondent with a hierarchical rank value of zero but a hierarchicalrange value of two. The interpretation of this example is a respondent who has an equalnumber of ties at two or more levels above and two or more levels below his or her ownhierarchical level.Density is defined as the degree to which a person’s network ties all have networkconnections with one another. A network where the people listed do not interact with anyother person would have a density of zero. A perfectly dense (1 .0) network is one in whichall people listed interact with one another.Part Three of the questionnaire was identical to Part Two, except it asked therespondents to list up to 1 5 expressive ties.7 The specific instructions for Part Three were:Before breaking the seal on this part of the questionnaire, please list up to1 5 individuals within your organization whom you consider to be “friends”of yours.Friends are defined as those individuals with whom you frequently or ofteninteract for personal satisfaction and enjoyment rather than just for thefulfilment of a particular task or goal. Those listed would include: (1)people you see socially outside of work, and (2) those people you spendtime with socially at work (e.g., at lunch and coffee breaks) but do not seeoutside of work.Therefore, you are to include anyone whom you consider to be a friend --even if you listed them as a task-related relationship on the previousquestion.Also, if you consider any supervisors and/or subordinates as friends, pleasebe sure to list them.Please break the seal upon completing your list of friends and provide therequested information in relation to every individual you have listed.Two key elements of the expressive instructions were that (1) they were to includenames from the instrumental ties list and (2) they could also list those supervisors orsubordinates they considered to be friends. Upon completing the list and breaking the seal,7ln the questionnaire, the term “expressive” was replaced with “friendship.” Though“expressive” is the established construct in the literature, the term “friendship” was betterunderstood by those individuals who participated in the pilot testing of the questionnaire.45the respondents were asked to provide, to the best of their ability, the following informationregarding each tie:(1) sex;(2) time known;(3) whether they saw this person socially outside of work;(4) the work location of the individual;(5) the individual’s occupational title;(6) whether the individual worked in the same functional department;(7) the department name of the individual if he/she worked in adifferent functional department;(8) the hierarchial position of the individual in relation to the studyparticipant;(9) whether this was an individual they managed/supervised;(10) how often they were in contact with the individual; and(11) who else on the list did the participant and each tie associate withsocially.Operationalizations of the expressive network dependent variables. Again, the tieinformation entered in Part Three provided the dependent variable data for the expressivenetwork size and the gender mix, range, frequency of contact, and density of the theseexpressive ties for each respondent. The operationalizations of these dependent variablesare listed in Table 5.2. With the exception of frequency of contact, the dependent variablesin Table 5.2 are either self-explanatory or are identical to the instrumental networkdependent variables discussed in Table 5.1.Frequency of contact8 represents, on average, the amount of contact a respondenthas with his or her expressive ties. A frequency of contact value of “five” would beinterpreted as respondents having daily contact with all of the individuals listed, while avalue of “two” would reflect respondents having contact with their set of tiesapproximately twice a month, on average.Part Four concluded the questionnaire and was comprised of demographic questionsrelating to the respondent and his or her current employment status.8Respondents were asked how often they had contact with each tie. The response setwas: (1) Daily, (2) Three times per week, (3) Once per week, (4) Twice a month, or (5)Less than once per month. In calculating the frequency of contact, the values were reversecoded.46Network Size and Gender Mix:Number of Individuals Listed’Number of MalesNumber of FemalesNumber of Overlapping Ties# of Female Overlapping Ties# of Individuals Seen Outsideof Work# of Females Seen Outsideof WorkRange:Number of Individuals perLocationSame Site-2.0 to 2.0 ((2 * of individuals two or more levelsabove) + (1 * # of individuals one levelabove) + (0 * # of individuals at same level)+ (-1 * # of individuals one level below) +(-2 * # of individuals two or more levelsbelow)) / (# of individuals listed)0 to 2.0 ((2 * of individuals two or more levelsabove) + (1 * # of individuals one levelabove) + (0 * # of individuals at same level)+ (1 * of individuals one level below) +(2 * # of individuals two or more levelsbelow)) / (# of individuals listed)Density 0.0 to 1 .0 (# of dyadic links listed) / ((number ofindividuals listed)(1 - number of individualslisted))1 Several individuals listed more than 1 5 individuals.TABLE 5.2Operationalizations of Dependent Variables: Expressive NetworkActualRange Calculation0 to 1 60 to 1 5Oto 140 to 100 to 70 to 1 5CountCountCountCountCountCountOto 14 Count0 to 1 5Oto 140 to 100 to 1 50 to 10CountCountCountCountCountSame City/Different SiteDifferent City/Different SiteNumber of Individuals Listedin Same FunctionNumber of DifferentFunctionsHierarchical Rank(Average rank of individualslisted)Hierarchical Range(Average range of individualslisted)47TABLE 5.2 (continued)Operationalizations of Dependent Variables: Expressive NetworkActualRange CalculationFrequency of Contact 0.0 to 5.0 ((5 * # of individuals seen daily) + (4 *# of individuals seen three times per week)+ (3 * # of individuals seen once a week)+ (2 * # of individuals seen twice a month)+ (1 * # of individuals seen less than onceper month)) / (# of individuals listed)48The Promotion QuestionnairesThe first promotion questionnaire was similar to the non-transition questionnaire,and the operationalizations of the dependent variables were identical. Except for severalquestions9added to the first and fourth sections, the only other change involved theinstructions to Part Two -- the section where the respondents listed their instrumental ties.In Part Two, the respondents were asked to ‘... list up to 15 individuals within yourcompany whom you consider to be useful in performing the tasks required of your jobPRIOR TO YOUR RECENT PROMOTION.’ This change to the instructions was criticalbecause many of the respondents had recently (within the past two to three months)assumed their new positions. To be able to test for changes to their instrumental networksfollowing the promotions, the respondents had first to provide a representation of theirnetworks pjj to assuming their new positions.1°The follow-up questionnaire was a shortened version of the first promotionquestionnaire, and consisted of three parts. Parts One and Two asked them to list up to 1 5instrumental and 1 5 expressive ties. When listing the 1 5 instrumental ties, the instructions9These questions were unrelated to the thesis’ research questions and hypotheses. InPart One, the questions asked the study participants to: (a) list the names of thoseindividuals who were instrumental in helping the study participants secure their promotions,and (b) explain how these individuals were helpful. In Part Four, the additional questionsasked about the study participants’ new salaries and new work locations following thepromotions.10Given the sample and data collection method, I had to place confidence in therespondents’ abilities to recall their instrumental networks prior to their promotions on thebasis of the questionnaires’ receipt coming within two months of the promotions. The non-transition bank respondent sample listed an average of 9.6 instrumental individuals,compared to 7.5 individuals listed by the promotion bank respondent sample. The banknon-transition and promotion respondent sets listed 8.1 and 8.8 expressive individuals,respectively.Informant accuracy in self-reports was debated in the 1 980s (see Bernard, Killworth,Kronenfeld, & Sailer, 1984). Freeman and Romney (1987), in an attempt to address theseconcerns, empirically tested informants’ ability to recall past interaction events. They foundthat individuals’ representations of their social structure (i.e., the relatively prolonged andstable pattern of interpersonal relations) were systematically biased toward the establishednorm of their interactions. In other words, informant accuracy is decidedly inaccurate whenindividuals try to reconstruct interaction patterns for a given event; however, accuracy ismuch improved when individuals try to reconstruct their stable pattern of interpersonalrelations.49specifically stated that they “... list up to 15 individuals within your company whom youconsider to be useful in performing the tasks required of your CURRENT JOB.” Part Threeconsisted of several demographic questions used as a check on the respondent. Examplesof both promotion questionnaires are in Appendix A.Eight months was deemed to be a long enough time period for changes inrelationships to stabilize following the promotions studied. Katz (1980) writes of threestages following an actual career transition: (1) a relatively brief socialization stage, (2) theinnovation stage, and (3) the adaptation stage. Social concerns dominate the socializationstage. The concerns are related to inclusion and becoming a contributing member of thework group (Katz, 1980). The innovation stage, with its emphasis on task performance,begins up to six months following a career transition with the adaptation stage occurring upto three years after the transition.Katz argues that following a career transition, such as a promotion, transitioners aremost concerned with meeting relatedness needs, and he posits that this concern is mostevident in the first six months following the transition. I assumed that in order for anindividual to eventually become proficient in his or her job, the majority of the changes toone’s work relationships should occur by the start of the innovation stage. Employees willwant to be proficient before their annual performance appraisals -- which are likely to occur1 2 months after assuming the new position. If so, stability in their work networks shouldoccur between the sixth and twelfth months. Since there was no empirical researchproviding a validated estimate as to when networks become stable following a careertransition, the post-promotion data were collected eight months following the actualpromotion.Pilot Testing of the QuestionnaireEarly versions of the non-transition questionnaire were critiqued by my thesissupervisory committee as well as by fellow doctoral students. During the summer of 1 993a questionnaire was given to three of my friends to complete under my observation. Thesethree friends worked at the same major, daily Northern Californian newspaper that I had50once been employed by, and these three individuals were chosen for two specific reasons.First, I had knowledge of their personal work networks and could check for the exclusion ofties (e.g., by oversight). Second, the friends would not have been bashful aboutquestioning indecipherable or confusing instructions or questions.The questionnaire was then revised and sent to the data site contacts for review.One manager at the telecommunications company filled out the questionnaire in theauthor’s presence. Otherwise, the final round of feedback came from the data sitecontacts. I met for over an hour with the insurance data site contact at which time eachand every question was reviewed. The author and one of the banking site contactsreviewed the questionnaire at length by phone. Ultimately, each contact at the four datasites approved the questionnaire.Reliability of the QuestionnaireTracy and Whittaker (1990) made no mention of the reliability of the grid format,and Ibarra (1993b) did not discuss the reliability of the sociometric questionnaire she usedto study manager networks. In that questionnaire, Ibarra (1993b) had respondents listnames of contacts representing each of five support domains: information, advice,friendship, career, and cooperation. Marsden (1990) states that network indices are largelyreliable when measures are taken to facilitate an individual’s capacity to recall and report hisor her network links. Reliability is increased when respondents are asked about fixed roles(e.g., friendships and task-related relationships) (van Groenou, van Sonderen, & Ormel,1990). Van Groenou et al. (1990) found that asking individuals to name network ties thatare based on exchange or affective feelings had a lower test-retest reliability in comparisonto the fixed-role approach. I used a fixed-role approach when asking respondents to listnames. In comparison, Ibarra’s (1 993b) approach was a mix of the exchange-, affective-,and fixed-role approaches.Furthermore, van Groenou et al. (1990) maintain that four weeks is a reasonabletime to test for the reliability of a questionnaire studying individuals’ personal networks.However, demonstrating the reliability of this sort of instrument is troublesome. Network51boundaries are flexible and within a short period of time changes can occur to the contentof relations (van Groenou et al., 1990). Within an organization, individuals are transferred,promoted, demoted, or leave the organization, which in turn impacts the make-up of theirown personal work networks as well as the work networks of others.In an attempt to test the stability of the instrumental and expressive lists, 20reliability check questionnaires’1were mailed to randomly selected non-transitionrespondents from the forestry and insurance companies one month after the originalquestionnaires were returned. Table 5.3 provides the correlation coefficients for thedemographic information. In all cases, the test-retest correlations were above 0.90.12However, as seen in Table 5.4, there was variation within a month’s time in theinstrumental and expressive lists provided by the respondents. Though there were nosignificant differences between men and women and the relative size of men’s andwomen’s networks did not change from Time 1 to Time 2, on average, 3.48 instrumentalties were added while 3.81 were dropped. With expressive ties, 1 .71 ties were addedwhile 2.00 ties were dropped.Given the high correlation coefficients for the demographic information, coupledwith the lack of male-female differences on the number of ties added or dropped in one’smonth time, there was no reason to assume that the respondents in this study had notreliably filled out the questionnaires to the best of their ability. However, the questionremains whether these changes were due to respondent oversight, to the natural evolutionof networks, or to both. If simple oversight is a problem, researchers will have to takesteps in the future to reduce this oversight, possibly by providing respondents with lists ofnames or with the chance to revisit and amend the lists at a later date.11An example of the questionnaire can be found in Appendix A.‘2Explanations for these “less than perfect” 1 .0 correlations include: (1) employeeswriting down having worked for their employer 1-1/2 years versus one year, and (2) thebirthdays of children occurring within the test-retest period.52TABLE 5.3Reliability of Respondent’s Demographic Information:Reliability Check Sample (n = 20)Demographic CorrelationVariables CoefficientsGender 1 .00Time with Employer 0.92Education Level 1 .00Number of Children 1 .00Age of Oldest Child 0.97Age of Youngest Child 0.92TABLE 5.4Number of Individuals Remaining in, Added to, and Dropped from Respondents’Instrumental and Expressive Networks1: Reliability Check (n = 20)Instrumental Ties Expressive TiesTi T2 Same Add Drop Ti T2 Same Add DropMale 9.7 9.8 6.1 3.7 3.6 8.3 8.2 6.6 1.6 1.6Female 11.3 10.5 7.3 3.2 4.0 9.6 9.1 7.2 1.9 2.4Overall 10.5 10.1 6.7 3.5 3.8 8.9 8.6 6.9 1.7 2.0‘The time between Time 1 and Time 2 was approximately one month.TABLE 5.5Non-Transition Study Questionnaire Response RatesReturned ResponseMailed Completed Rate Male’ Female’Banking 234 83 35.5% 39 (33.3%) 44 (37.6%)Forestry 145 60 41.4% 42(39.3%) 18(47.4%)Insurance 203 99 48.8% 67 (49.6%) 32 (47.1 %)Total 582 242 41.6% 148(41.2%) 94(42.3%)1 Listed are the number of returned questionnaires and the response rates for both sexes.53ResDonse RatesTable 5.5 provides specific information on the non-transition respondent samplewith the overall and company response rates. Nearly 42% of those receiving a non-transition questionnaire returned a completed form.13 From the first data collection in thepromotion study, 43 completed questionnaires were returned (for a 26.5% response rate).Of these 43 respondents, only 33 (20 females and 1 3 males) completed the follow-uppromotion questionnaire.Though not “high,” there are several plausible explanations for the non-transitionand promotion study response rates. First, the questionnaire did take 30 to 45 minutes tofill out, and this, coupled with the busy schedule of managers and supervisors, had tosuppress the response rate. Second, several individuals considered the request to providespecific names of individuals a sensitive issue, and returned the questionnaires withoutcompleting them.In an attempt to increase the response rate, reminder letters14 were distributed atthree of the data sites a minimum of two weeks after the initial receipt of thequestionnaires. The higher response rate at the insurance company could more than likelybe attributed to the use of electronic mail. All managers and supervisors at the insurancecompany received two e-mail messages from the individual serving as my data sitecontact.15 The first message informed them that they may receive a company-endorsedquestionnaire studying personal work networks. The second message reminded themanagers and supervisors of the questionnaires and followed two weeks after receipt of thequestionnaires.13Thirteen of the 582 questionnaires sent out were returned but were excluded from thestudy because they were either incomplete or incorrectly filled out.copy of this follow-up letter is in Appendix A.15This individual was employed by the human resources department.54Chapter Six:DATA ANALYSIS STRATEGY and RESULTSThe same format employed in the previous chapters is followed in this chapter withthe non-transition study results preceding those of the promotion study. The variablemeans and standard deviations are reported for the entire non-transition sample variablesand the intercorrelations1among these variables are in Appendix B. The presence ofmulticollinearity was tested with the variance inflation factor (VIF) (Neter, Wasserman &Kutner, 1985). The VIF values for all expressive and instrumental network predictorvariable sets, including the interaction terms2,were less than four. The majority of the VIFvalues were less than two, indicating no serious multicollinearity problems (Kutner et al.,1985).The normality of the dependent variables was also tested (see Appendix B, TableB.14): only 5 of 29 variables, the hierarchical rank and hierarchical range (in theinstrumental and expressive networks network) and the frequency of contact (in theexpressive network), were normally distributed. Finally, Appendix C provides an overviewand related tables regarding the demographic comparison of the non-transition study andpromotion study respondent samples. The male-female differences were the most salient tothis thesis, and there were a number of significant differences in the non-transition study(see Appendix C); however, the majority of the differences had no relevance to the researchquestions or hypotheses in Chapter Two. The lone exception was job category, which isbeing controlled in the research models, where 20.4% of the women were in administrativeor administrative/management roles, compared to only 8.8% for the men. Moreover,women were less likely to categorize their jobs as professional/management or1Correlations were calculated using the Spearman rank correlation coefficient method.2The interaction terms were created by multiplying the two company dummy variableswith the other predictors variables. Before creating the interaction terms, all first-ordervariables were first centered to attenuate for possible multicollinearity affects (Aiken &West, 1991).55technical/management in comparison to men, 20.4% to 38.5%, respectively.THE NON-TRANSITION STUDYDifferences in men’s and women’s instrumental and expressive networks were firsttested using the non-parametric Mann-Whitney U procedure because of the non-normallydistributed dependent variable data. The Mann-Whitney U is one of the most powerful non-parametric tests and is a useful alternative to the t-test when the researcher wishes toavoid the t-test’s assumptions of normally distributed data and homogeneity of variances(Siegel, 1956). With large samples, the Mann-Whitney U is also more powerful than theKolmogorov-Smirnov Test (Siegel, 1956). Table 6.1 provides the instrumental networkdependent variable means and standard deviations for men and women; Table 6.2 theexpressive network dependent variable means and standard deviations for men and women.Men and women differed significantly on only the gender composition and thedensity of their instrumental networks. Men’s instrumental networks were morehomophilous (7.24 males and 3.64 females, on average) and higher in density, whereas thegender composition of women’s instrumental networks (5.62 men and 4.61 women, onaverage) was more differentiated. Relating these findings to the research question inChapter Two, men’s and women’s instrumental networks do not seem to differ greatly.More gender significant differences were identified in the expressive networks,including:[11 gender composition -- men’s and women’s expressive networks tended to behomophilous;[21 the number of different functions represented by the individuals listed --women tended to derive more expressive ties from functional departmentsdifferent from their own;[31 the inclusion of supervisors in the expressive network -- men appeared toinclude more supervisors in their expressive networks than did women;[41 the frequency of contact with the people listed -- men invariably had morefrequent contact with their expressive ties than did the women;[5] the location of the expressive ties -- a larger proportion of the women’sexpressive ties tended to come from work sites different from their own;and[6] density of the ties -- the men’s expressive networks invariably exhibited higherdensity in comparison to the women’s expressive networks.56TABLE 6.1Mann-Whitney U Non-Parametric Results:Male versus Female Instrumental Network Dependent Variables (n = 242)Non-Male Female ParametricDependent Variables Means/(SD) Means/(SD) p-valuesNetwork Size and Gender Mix:Number of Individuals Listed 10.89 (3.89) 10.22 (4.48) .310Number of Males 7.24 (3.24) 5.62 (3.54) .007Number of Females 3.64 (2.29) 4.61 (2.65) .003Range:Number of Individuals Listedper LocationSame Site 6.20 (4.77) 5.43 (4.88) .180Same City/Different Site 2.50 (3.70) 2.66 (3.24) .350Different City/Different Site 2.18 (3.34) 2.18 (3.53) .980Number of Individuals in 4.51 (3.67) 3.88 (3.62) .140Same FunctionNumber of Different Functions 3.72 (2.79) 3.96 (2.77) .480Hierarchical Rank -0.08 (0.59) -0.05 (0.63) .920Hierarchical Range 0.86 (0.41) 0.89 (0.41) .550Density 0.29 (0.26) 0.21 (0.20) .03957TABLE 6.2Mann-Whitney U Non-Parametric Results:Male versus Female Expressive Network Dependent Variables (n = 242)Non-Male Female ParametricDependent Variables Means/(SD) Means/(SD) p-valuesNetwork Size and Gender Mix:Number of Individuals Listed 8.38 (4.62) 8.81 (4.05) .470Number of Males 5.97 (3.71) 3.50 (2.68) < .001Number of Females 2.46 (2.08) 5.31 (3.22) < .001Number of Overlapping Ties 2.61 (2.18) 2.34 (1.88) .510Number of Overlapping 0.69 (1 .07) 1 .20 (1 .27) .001Females# of Ties Seen Outside of Work 4.21 (3.69) 4.70 (3.81) .286# of Female Ties Seen Outside 1.00 (1.54) 3.12 (3.06) < .001of WorkRange:Number of Individuals Listedper LocationSame Site 5.79 (4.23) 4.70 (3.66) .062Same City/Different Site 1 .60 (2.62) 2.55 (3.22) .007Different City/Different Site 1 .21 (1 .95) 1 .58 (2.53) .380Number of Individuals in 5.28 (4.13) 4.64 (3.70) .280Same FunctionNumber of Different Functions 2.18 (2.03) 2.72 (2.19) .044Hierarchical Rank -0.31 (0.67) -0.41 (0.61) .140Hierarchical Range 0.91 (0.43) 1 .02 (0.36) .028Number of Supervisors 0.38 (0.51) 0.23 (0.43) .034Number of Subordinates 1 .37 (1 .78) 1 .30 (1 .57) .840Density 0.24 (0.22) 0.16 (0.19) .002Frequency of Contact 3.66 (0.84) 2.97 (0.93) < .00158These findings lend support to composite Hypothesis 1 a that men’s expressive networks dodiffer from those of women.Though interesting, the non-parametric results were not very informative given themodels outlined in Chapter Three. Boneau (1960, p. 49) writes that non-parametrictechniques:quite generally.., couple their freedom from restricting assumptions with adisdain for much of the information contained within the data... tests whichmake no assumptions about the distribution from which one is sampling willtend not to reject the null hypothesis when it is actually false as often aswill those tests which do make assumptions.Information contained within the data but not included in the non-parametric tests includethe effect of the company as a possible moderator variable and the other control variables(e.g., job category, job level, child-rearing responsibilities and/or race). At issue is: Dothese gender differences, identified in the non-parametric analyses, remain when onecontrols for these variables? To answer this question, more sophisticated statisticalanalyses were necessary.The Use of Multiple RegressionOn the basis of the arguments of Cohen (1968), I used multiple regression analysisinstead of analysis of variance. In choosing between multiple regression and analysis ofvariance, Cohen writes of their theoretical equivalence (as their null hypotheses aremathematically equivalent), but then argues for the practical advantages of multipleregression. Cohen (1968) states that if there are other independent variables of interest(e.g., main effects, interactions, covariates), they are more easily added to the model bymeans of multiple regression. Moreover, multiple regression is a general varianceaccounting procedure in the study of natural variation, which is at the heart of this thesis;analysis of variance is better for artificial or experimentally manipulated variation (Cohen,1968).The use of multiple regression requires the following assumptions: (1) normallydistributed dependent scores, and (2) equal variances for each dependent variable at each x59point (Kerlinger & Pedhazur, 1973). When these assumptions are not met, interpretingresults becomes problematic. For example, are the results significant due to differencesbetween the means, or are the results due to the violations of the assumptions (Boneau,1 960)? That question is a concern in this study. However, there is evidence that theordinary t and F tests are nearly immune to the violation of assumptions or can easily bemade so if precautions are taken (Boneau, 1960). Violation of the homogeneity of varianceassumption is “drastically disturbing” to the distribution of ts and Fs if the sample sizes arenot the same for all groups (Boneau, 1960). Boneau (1960, p. 56) states “... it wouldseem that the combination of unequal variances and unequal sample sizes might play havocwith F test probability statements.”Though increasing the sample size has the effect of off-setting the effects of theskew associated with the data, a combination of unequal sample sizes and unequalvariances “automatically produces inaccurate probability statements which can be quitedifferent from the nominal values” (Boneau, 1960, p. 56). Unequal sample sizes areespecially problematic when the larger sample has the larger variance (Boneau, 1960).Such a situation could result in a more conservative F test, whereas larger samples withsmaller variances could produce a higher percentage of “significant” Fs than expected(Boneau, 1960). In this thesis, there was no singular variance pattern between the maleand female respondent samples (see Tables 6.1 and 6.2).Kerlinger and Pedhazur (1973, p. 47) maintain that there is “... no need to assumeanything to calculate rs, bs, and so on.” It is only when we make inferences from a sampleto a population that we must pause and think of assumptions. Homogeneity of variance isimportant when regression results are used in statistical estimation procedures. Asdescribed in Chapter Four, the non-transition sample is a convenience sample, and I havealready acknowledged that my ability to make inferences to the population of managers andsupervisors is severely limited.The Regression ModelsCohen (1968) argues that one needs to specify the regression model (or models) to60be tested before conducting statistical analyses, and this requires an incisive priorconceptual analysis of the research problem. With both the instrumental and expressivework networks, this was done (see Chapter Three).Disregarding for the moment the possible company moderator effect, theinstrumental and expressive statistical models involve regressing the covariate (number ofindividuals listed for either the instrumental or expressive networks), gender of therespondent, and the control variables (e.g., the respondent’s job category, job level, race,and/or child-rearing responsibilities) on each dependent variable. This statistical model willbe referred to as the “Base Model.” Testing for company effects requires two additionalregression models. The first of these models will be known as the “Shift Model;” it testsfor company main effects. The second model, known as the “Moderator Model,” tests forcompany interaction effects.To reiterate, there are three regression models of interest when testing for genderand company effects.(1) The Base Model. No company effects tested. Predictor variables include:Covariate, independent variable, and control variables.(2) The Shift Model. Company main effects. Predictor variables include:Covariate, independent variable, and control variables (includingcompany as a control variable).(3) The Moderator Model. Company interaction effects. Predictor variablesinclude: Covariate, independent variable, control variables (includingcompany as a control variable), and interaction variables (created bymultiplying each covariate, independent variable, and other controlvariables by the company control variable(s)).Table 6.3 presents the operationalizations of each of the predictor variables.Each model provides the variance explained for each dependent variable.Comparisons can then be made between: (1) the Moderator Model and the Base Model, (2)the Moderator Model and the Shift Model, and (3) the Shift Model and the Base Model.These comparisons test for the increase in variance explained for each dependent variable inthe instrumental or expressive networks through the use of ordinary least squaresregression.61TABLE 6.3Predictor Variable OperationalizationsVariableTypes Variable ValuesCovariate: # of Ties Listed Count 0 to 1 8 [Instrumental]o to 1 6 [Expressive]Independent: Gender Dummy 0 if Male1 if FemaleControl2:Job Category/Manager Dummy 1 if Manager, Manager/Professional,Manager/Technicalo if Manager/Administrative, Administrative,Technical, Professional, or OtherJob Category/Administrative Dummy 1 if Manager/Administrative, AdministrativeO if Manager, Manager/Professional,Manager/Technical, Technical,Professional, or OtherJob Category/Technical, Dummy OmittedProfessional, or OtherJob Level/Sr. Management Dummy 1 if Executive or Senior Managero if Middle Manager, First-Line Supervisor,or OtherJob Level/Middle Manager Dummy 1 if Middle ManagerO if Executive, Senior Manager, First-LineSupervisor, or OtherJob Level/First-Line Supervisor Dummy 1 if First-Line SupervisorO if Executive, Senior Manager, MiddleManager, or OtherJob Level/Other Dummy OmittedChild-Rearing Responsibility Dummy 1 if Children 1 8 years old or youngero if No Children or Children over 1 8 years oldRace Dummy 1 if Non-Caucasiano if CaucasianModerator:Company/Forest Dummy 1 if Employed by Forestry Company0 if Employed by Bank or InsuranceCompaniesCompany/Bank Dummy 1 if Employed by Bank0 if Employed by Forestry or InsuranceCompanies62The first comparison is between the Moderator and the Base Models, and thiscomparison tests whether there are any significant main and/or interaction company effects.If there are company effects, additional comparisons can be made as to whether there arecompany moderator effects (Moderator Model versus Shift Model) and/or company maineffects (Shift Model versus Base Model).One of the equations for the influence of additional explanatory variables on themean of each dependent variable uses the calculated R-squared for each model (Kmenta,1971, p. 371):F = [(R2q- R2k)/(1 - R2q)] * [(fl - Q)/(Q - K)]where: q denotes the model with the greater number of predictorvariables;k denotes the model with the lesser number of predictorvariables;n = the number of respondents;Q = the number of predictor variables in the larger model; andK = the number of predictor variables in the smaller model.Degrees of freedom for tabulated value of F is Q-K, n-Q.The equation is equivalent to:F = [(SSRq- SSRk)/(Q - K)] / L(SSEq)/(fl - Q)] (Kmenta, 1971, p. 370); andF = [(SSEk - SSEq)/(K - Q)] / [(SSEq)/(n - Q)] (Neter et al., 1985, p. 91).As outlined in Chapter Three, there were seven predictor variables (excluding theconstant) in the instrumental network Base Model. The number of predictor variablesincreased to nine in the Shift Model with the addition of the two company dummy variablesand to 23 in the Moderator Model with the addition of the interaction terms. For theexpressive network models, the Base Model had nine predictor variables, the Shift Modelhad 11 predictor variables, and the Moderator Model had 29 predictor variables.A primary concern when conducting statistical analyses is to maintain power; 23and 29 predictor terms (for the instrumental and expressive Moderator Models,respectively), given a sample of 242 individuals, diminishes greatly the statistical power.Cohen (1968) writes that with a few factors one can generate a very large number of63distinct independent variables, and such features of data in an analysis “must be resisted,”and Cohen (1968, p. 442) states that “... each esoteric issue posed to the data costs adegree of freedom which is lost from the error estimate... enfeebling the statistical power ofthe analysis.” Cronbach (1975, p. 119) adds: “... once we attend to interactions, weenter a hall of mirrors that extends to infinity.” Thus, Cohen (1968) argues aggressivelyagainst studies with “prodigious numbers” of independent variables as well as dependentvariables. Unfortunately, the large number of dependent variables in this thesis could notbe avoided.With the addition of the interaction terms, the instrumental and expressive personalwork network models were no longer conceptually parsimonious. There was also theconcern of statistical power. Therefore, I reduced the number of predictor variables forboth the instrumental and expressive network models after first creating a hierarchy for theset of predictor variables. The covariates, gender and job category were the mostimportant variables in the instrumental model (see Chapter Three). By excluding job level,the Moderator Model had 14 predictor variables (including interaction terms) instead of 23.For the expressive network, the covariate and gender were essential to test the hypothesis,and I decided that job category and child-rearing responsibilities were more important thanthe individual’s race and job level on the basis of the discussion in Chapter Three. Previousresearch has indicated that most gender differences in network properties disappear whenstructural variables such as hierarchical rank are controlled (Ibarra, 1993b, 1992; Moore,1990). The expressive network Moderator Model had 17 predictor variables (including64interaction terms) instead of 29. Figures 6.1 and 6.2 present the “revised models.3Results of Company and Gender EffectsThe R-squared for the three regression models are listed in Table 6.4 (instrumentalnetworks) and Table 6.5 (expressive networks). Tables 6.6 (instrumental networks) andTable 6.7 (expressive networks) present the calculated F-statistics for the modelcomparisons. The Moderator versus Base Model comparison results highlight the effect therespondent’s company had on the variance explained for each instrumental network andeach expressive network dependent variable. Significant (p < .05) company main and/orinteraction effects were found in nine of the 11 instrumental network dependent variablesand 1 2 of 1 8 expressive network dependent variables. The respondent’s company had asignificant main effect (Shift versus Base Model comparison) on 9 of 11 instrumentalnetwork dependent variables and 10 of 1 8 expressive network dependent variables.Significant company interaction effects were found for three instrumental variables and sixexpressive variables. The “company” effect is further illustrated in Tables 6.8 and 6.9. Thepattern of gender differences for the instrumental and expressive dependent variables wasnot consistent among the three companies.These model comparison results reveal that company effects need to be controlledwhen comparing men’s and women’s work personal networks. Minimally, company as amain effect must be controlled through the use of dummy variables. Beta coefficients (andtheir standard errors) for the three models on all the dependent variables are presented inAppendix D. Considering the shift model, which tested for company main effects, a3There were no significant differences between men and women in regards to race(95.2% of the men and 90.3% of the women listed themselves as caucasian) and job level(19.6% of the men and 13.8% of the women listed themselves as an executive or seniormanager; 34.5% of the men and 41 .5% of the women listed themselves as middlemanagement; and 36.5% of the men and 40.4% of the women listed themselves as firstline supervisors).There were significant differences when one considered the job category and childrearing responsibilities for the respondents. Cross-tabulations revealed that 20.8% of thewomen and 8.8% of the men worked in administrative/administrative management roles.Only 39.4% of the women had children 1 8 years or younger compared to 60.8% of themen.65FIGURE 6.1Non-Transition/Instrumental Network ModelPredictor Variables Derendent VariablesCovariate: Instrumental Network Size andNumber of Individuals Listed Gender Mix:Number of MalesIndependent: Number of FemalesGender of Study ParticipantRange:Control: Location of Individuals ListedJob Category Number of Individuals in Same FunctionNumber of Different FunctionsModerator: Hierarchical Rank of TiesCompany Hierarchical Range of TiesDensityFIGURE 6.2Non-Transition/Expressive Network ModelPredictor Variables DeDendent VariablesCovariate: Expressive Network Size andNumber of Individuals Listed Gender Mix:Number of MalesIndependent: Number of FemalesGender of Study Participant Number of Overlapping Instrumental andExpressive TiesControl: Number of Overlapping InstrumentalJob Category and Expressive Female TiesChild-rearing Responsibility Number of Individuals Listed Seen Outsideof WorkModerator: Number of Females Seen Outside of WorkCompanyRange:Location of TiesNumber of Individuals in Same FunctionNumber of Different FunctionsHierarchical Rank of TiesHierarchical Range of TiesDensityFrequency of Contact66TABLE 6.4Base, Shift, and Moderator Model Regression R-Squared:Instrumental Network Dependent Variables (n = 242)ModeratorBase Model: Shift Model: Model:Dependent Variables n R-Squared1 R-Squared2 R-Squared3# of Individuals Listed 242 .008 .038 .095# of Males 242 .695 .720 .730# of Females 242 .409 .456 .476Location of Individuals Listed:Same Site 240 .319 .391 .414Different Site/Same City 240 .064 .125 .159Different Site/Different City 240 .058 .242 .266# of Individuals in Same Function 241 .135 .185 .239# of Different Functions 241 .317 .331 .355Hierarchical Rank 241 .030 .064 .128Hierarchical Range 241 .041 .096 .129Density 227 .080 .084 .1 541Base Model: There are five explanatory variables (including the constant) in the instrumentalties model.2Shift Model: There are seven explanatory variables (including the constant) in the instrumentalties model.3Moderator Model: There are 1 5 explanatory variables (including the constant) in theinstrumental ties model.67TABLE 6.5Base, Shift, and Moderator Model Regression R-Squared:Expressive Network Dependent Variables (n = 242)ModeratorBase Model: Shift Model: Model:Dependent Variables n R-Squared1 R-Squared2 R-Squared3# of Individuals Listed 242 .01 5 .033 .044# of Males 242 .71 5 .726 .766# of Females 242 .580 .596 .655# of Overlapping Ties 242 .299 .339 .378# of Overlapping Females 242 .168 .172 .219# of Individuals Listed Seen 238 .435 .440 .474Outside of Work# of Female Seen Outside of Work 238 .338 .349 .400Location of Individuals Listed:Same Site 238 .482 .522 .555Same City/Different Site 238 .177 .223 .273Different City/Different Site 238 .119 .226 .265# of Individuals in Same Function 238 .477 .514 .563# of Different Functions 236 .196 .248 .324Hierarchical Rank 234 .078 .091 .131Hierarchical Range 234 .026 .110 .174# of Supervisors 238 .073 .076 .160# of Subordinates 238 .228 .251 .313Density 224 .054 .060 .090Frequency of Contact 224 .166 .186 .2201Base Model: There are six explanatory variables (including the constant) in the expressive tiesmodel.2Shift Model: There are eight explanatory variables (including the constant) in the expressive tiesmodel.3Moderator Model: There are 1 8 explanatory variables (including the constant) in the expressiveties model.68TABLE 6.6Comparison of Regression Models:Instrumental Network Dependent Variables (n = 242)Moderator vs. Moderator vs. Shift vs. BaseBase Model: Shift Model: Model:Dependent Variables F-Statistic1 F-Statistic2 F-Statistic3# of Individuals Listed 2.793** 2.448* 3.693*#of Males 2.932** 1.093 10.256***# of Females 2.933** 1.093 10.259***Location of Individuals Listed:SameSite 3.631*** 1.075 13.818***Same City/Different Site 2.528*** 1.144 8.023***Different Site/Different City 6.368*** 0.917 28.251***# of Individuals in Same Function 3.078** 1.977 7.240*# of Different Functions 1 .341 1 .077 2.388Hierarchical Rank 2.542** 2.055* 4333*Hierarchical Range 2.299* 1.080 7.156**Density 1.874 2.210* 0.505* p < .05; ** p < .01; *** p < .0011Moderator-Base Model Level of Significance: Instrumental (F10>20 at the .05% level = 1 .91,at the .01 level = 2.47, at the .001 level = 3.24).2Moderator-Shift Model Level of Significance: Instrumental (F8>120 at the .05% level = 2.02, atthe .01 level = 2.66, at the .001 level = 3.55).3Shift-Base Model Level of Significance: Instrumental (F2>120 at the .05% level = 3.07, at the.01 level = 4.70, at the .001 level = 7.32).69TABLE 6.7Comparison of Regression Models:Expressive Network Dependent Variables (n = 242)Moderator vs. Moderator vs. Shift vs. BaseBase Model: Shift Model: Model:Dependent Variables F-Statistic1 F-Statistic2 F-Statistic3if of Individuals Listed 0.696 0.334 2.194#of Males 4.047*** 3.843*** 4.518*# of Females 4.053*** 3.848*** 4.527*if of Overlapping Ties 2.355** 1.383 7.098**if of Overlapping Females 1.214 1.349 0.530if of Individuals Listed Seen 1 .379 1 .445 1 .031Outside of Workif of Female Seen Outside of Work 1 .896* 1 .872 1 .946Location of Individuals Listed:Same Site 3.017** 1.639 9.634***Same City/Different Site 2.42 1 ** 1.501 6.87 1 **Different City/Different Site 3.61 7*** 1.148 1 5.862***if of Individuals in Same Function 3.563*** 2.451 * 8.581 ***if of Different Functions 3.462*** 2.479** 7.864***Hierarchical Rank 1.100 1.004 1.581Hierarchical Range 3.232*** 1.685 10.648***if of Supervisors 1 .898* 2.20 1 * 0.363if of Subordinates 2.274* 1 .963* 3.676*Density 0.684 0.665 0.789Frequency of Contact 1.199 0.888 2.770* p < .05; ** p < .01; p < .0011Moderator-Base Model Level of Significance: Instrumental (F12>20 at the .05% level = 1 .83,at the .01 level = 2.34, at the .001 level = 3.02).2Moderator-Shift Model Level of Significance: Instrumental (F10>20 at the .05% level = 1 .91,at the .01 level = 2.47, at the .001 level = 3.24).3Shift-Base Model Level of Significance: Instrumental (F2>120 at the .05% level = 3.07, at the.01 level = 4.70, at the .001 level = 7.32).70TABLE6.8ComparisonofRespondentSampleversusIndividualCompanyComparisonsResults:InstrumentalNetworkDependentVariables(n=242)CombinedCombinedRegressionBankForestryInsuranceNon-ShiftNon-Non-Non-Parametric1ModelParametric1Parametric1Parametric1DependentVariablesp-valuesp-valuesp-valuesp-valuesp-values#ofIndividualsListed.016#ofMales.007<.001.003.048#ofFemales.003<.001.010.001LocationofIndividualsListed:SameSiteSameCity/DifferentSiteDifferentCity/DifferentSite.024#ofIndividualsinSameFunction#ofDifferentFunctions.064.061.061HierarchicalRankHierarchicalRangeDensity.039<.05.016‘Mann-WhitneyUnon-parametrictest;p<.10listed.TABLE6.9ComparisonofRespondentSampleversusIndividualCompanyResults:ExpressiveNetworkDependentVariables(n=242)CombinedCombinedRegressionBankForestryInsuranceNon-ShiftNon-Non-Non-Parametric1ModelParametric1Parametric1Parametric1DependentVariablesp-valuesp-valuesp-valuesp-valuesp-values#ofIndividualsListed#ofMales.000<.001.005.000#ofFemales.000<.<.01.002.020#ofIndividualsListedSeenOutsideofWork#ofFemaleSeenOutsideofWork.000<.<.05SameCity/DifferentSite.007.020DifferentCity/DifferentSite#ofIndividualsinSameFunction.10#ofDifferentFunctions.044.017.057HierarchicalRank.073HierarchicalRange.028<.10#ofSupervisors.034<.05.071.058#ofSubordinatesDensity.002<.05.020.071FrequencyofContact.000<.;p<.10listed.significant (p < .05) bank and/or forestry company main effect was found for 10 of theeleven instrumental dependent variables4and for 8 of 1 8 expressive network dependentvariables5.Gender differences. The regression results confirm the gender instrumental networkdifferences found through the Mann-Whitney U non-parametric tests (Table 6.10). Men haddeveloped more homophilous instrumental networks (consisting of 7.24 men and 3.64women, compared to 5.62 men and 4.61 women for the women respondents) that werehigher in density (0.29 versus 0.21; p = .039). Research Question la asked: Do theinstrumental work networks of men differ from the instrumental work networks of women?Beyond the gender composition and density differences, men and women did not differsignificantly on the remaining instrumental network characteristics.The results from the expressive network non-parametric tests and regressionanalyses (Table 6.11) are not as consistent as the instrumental network findings. For themost part, the regression models confirm the Mann-Whitney U non-parametric results. Themajority of men’s and women’s expressive ties were homophilous. Men’s expressivenetworks averaged 5.92 males, women 3.50 males (p < .001). Women’s expressivenetworks averaged 5.31 females, men 2.46 females (p < .001). Women saw more womenoutside of the office (3.1 2 versus 1 .00; p < .001) and had more female overlapping ties(between their instrumental and expressive networks) than did men (1 .20 versus 0.69, p <.001), and women had less contact with their friends in comparison to men (2.97 versus3.66, p < .001).The results were not as consistent between the non-parametric tests and regressionanalyses when one considers the location and hierarchical range of the people listed, as well4The lone exception was density.5These nine expressive network dependent variables included: number of overlappingexpressive and instrumental ties, number of individuals listed at same site, number ofindividuals listed at different site/same city, number of individuals listed at differentsite/different city, number of individuals listed at same function, number of differentfunctions listed, hierarchical range of individuals listed, and frequency of contact.73TABLE6.10GenderComparisonofNon-ParametricandRegressionModelsResults:InstrumentalNetworkDependentVariables(n=242)Non-DependentVariablesMaleFemaleParametric’BaseModelShiftModelBetasModeratorModelMeansMeansp-valuesBetasBetas#ofIndividualsListed10.8910.22#ofMales7.245.62.0071.043***O.898***O.977***#ofFemales3.644.61.0031.043***0.898***0.978***LocationofIndividualsListed:SameSite6.205.43SameCity/DifferentSite2.502.66DifferentCity/DifferentSite2.182.18#ofIndividualsin4.513.88SameFunction#ofDifferentFunctions3.723.96HierarchicalRank-0.08-0.05HierarchicalRange0.860.89Density0.290.21.0390.077*.0.70**p<.05;**p<.01;***p<.0011Mann-WhitneyUnon-parametrictest;p<.10listed.TABLE6.11GenderComparisonofNon-ParametricandRegressionModelsResults:ExpressiveNetworkDependentVariables(n=242)Non-DependentVariablesMaleFemaleParametric1BaseModelShiftModelModeratorModelMeansMeansp-valuesBetasBetasBetas#ofIndividualsListed8.388.81#ofMales5.923.50.0002.61O***2.425***2.341***#ofFemales2.465.31.0002.610***2.425***2.341***#ofOverlappingTies2.612.34#ofOverlappingFemales0.691.20.0010.437**0.417**0.373*#ofIndividualsListedSeen4.214.70OutsideofWork#ofFemaleSeenOutsideofWork1.003.12.0001.958***1.817***1.765***LocationofIndividualsListed:SameSite5.794.70.0621.364**0.924*0.915*SameCity/DifferentSite1.602.55.0070.810*DifferentCity/DifferentSite1.211.58#ofIndividualsinSameFunction5.284.640.944*0.725#-0.791*#ofDifferentFunctions2.182.72.044HierarchicalRank-0.31-0.41HierarchicalRange0.911.02.0280.120*0.089##ofSupervisors0.380.23.0340.143*0.151*-0.127##ofSubordinates1.371.31Density0.240.16.0020.081**0.072*0.O75*FrequencyofContact3.662.97.0000.727***0.687***O.724***#p<.10;*p<.05;**p<.01;***p<.0011Mann-WhitneyUnon-parametrictest;p< the number of supervisors considered to be friends, and the density of the networks.One particularly anomalous finding was that the number of different functions from which arespondent drew his or her expressive ties was significant in the regression analyses, butnot in the non-parametric tests. However, the number of individuals working in the samefunction as the respondent was significant in the non-parametric tests, but not in theregression analyses.Hypothesis 1 a posited that men’s expressive work networks would differ fromwomen’s expressive networks. One could conclude that the expressive networks of mendo differ from the expressive networks of women. The “strongest” differences involve thegender composition network characteristics (e.g., number of males listed, number ofoverlapping female ties); however, the number of male-female personal work networkdifferences diminishes greatly when one excludes the results which attest to the degree ofsame-sex individuals found in the expressive networks.Control variables and interaction terms. Few control variables’ (e.g., child-rearingand job level) beta coefficients were significant (see Appendix D). Considering the ShiftModel (which controlled for company main effects), the child-rearing beta coefficients werenot significant on any of the expressive network dependent variables. In comparison to thechild-rearing control variable, more job category control variables (i.e., the administrative dichotomy) were significant. Holding an administrative position impactedsignificantly the number of males and females in both the instrumental and expressivenetworks and the number of expressive individuals who worked at the same sites as therespondent. In fact, women who held administrative positions had, on average, 2.5 lessmales in their instrumental networks and 3.7 less males in their expressive networks thandid their male counterparts. Furthermore, the location of expressive individuals, thepositional rank of the expressive individuals, and the number of subordinates included in theexpressive networks were affected if the individuals were working in non-administrativemanagement positions.Overall, there were more significant (p < .05) interaction beta coefficients, but the76majority of these significant coefficients involved the “company x covariate (number ofindividuals listed)” interaction terms. Only 9 of the possible 58 “gender x company” betacoefficients were significant.6 There were only 5 significant “company x control variable”beta coefficients.7 Finally, 11 of 54 “company x covariate” interaction terms weresignificant.8THE PROMOTION STUDYThe promotion study first examined the same research question and hypothesis asdid the non-transition study, the only difference between the two studies being thepopulation studied. Male and female dependent variable means are listed in Tables 6.12and 6.13. These means are from the data collected in the second (or follow-up)questionnaire, and the results from the non-parametric tests and regression analyses arealso provided in the same tables.There were no significant differences (at the .05 level) between men’s and women’sinstrumental networks, though the size of the networks, the number of males in thenetworks, and the number of individuals listed at a different site and different city didapproach significance (p < .10). Expressive network significant differences included the6The significant “gender x company” beta coefficients included: instrumental (gender xbank -- size); expressive (gender x bank -- number of males listed, number of females listed,number of individuals listed working at a different site in a different city, number ofindividuals listed working in the same function, and number different functions listed;gender x forest -- number of males listed, number of females listed, and number of femalesseen outside of work).7The significant “company x control variable” beta coefficients included: instrumental(manager x forest -- same function); expressive (children x bank -- number of males listed,number of females listed; administrative x forest -- hierarchical range; and manager x bank-- hierarchical range).8The significant “company x covariate” beta coefficients included: instrumental (bank --number of individuals listed at same site; forest -- number of individuals at same hierarchicalrank); expressive (bank -- number of males listed, number of females listed, number ofindividuals listed at same site, number of individuals listed at different site/different city,and number of individuals listed in same function; forest -- number of overlappingexpressive and instrumental ties, number of individuals listed working at a differentsite/same city, number of individuals listed in same function, number of different functionslisted).77TABLE 6.12Gender Non-Parametric and Regression Results:Promotion Sample Instrumental Network Dependent Variables (n = 33)Non-Male Female Parametric RegressionDependent Variables Means/(SD) Means/(SD) p-values Model Betas# of Individuals Listed 10.23 (4.89) 7.10 (4.12) .056# of Males 6.00 (3.63) 3.35 (3.15) .053# of Females 4.23 (2.24) 3.75 (2.57)Location of Individuals Listed:Same Site 1.54 (2.26) 2.85 (3.65) 2.12#Same City/Different Site 4.46 (5.46) 3.30 (3.61)Different City/Different Site 4.23 (4.87) 0.95 (1 .23) .091 2.86*# of Individuals in 3.46 (3.99) 3.05 (2.63)Same Function# of Different Functions 3.38 (1 .76) 2.60 (2.62)Hierarchical Rank 0.02 (0.86) -0.06 (0.81)Hierarchical Range 0.98 (0.42) 1 .01 (0.47)Density 0.27 (0.27) 0.38 (0.34)#p<.lO; *p<.0S1Mann-Whitney U non-parametric test; p < .10 listed.78TABLE 6.13Gender Non-Parametric and Regression Results:Promotion Sample Expressive Network Dependent VariablesNon- RegressionMale Female Parametric ModelDependent Variables Means/(SD) Means/(SD) p-values Betasif of Individuals Listed 9.23 (5.04) 8.05 (4.24)if of Males 4.23 (3.12) 2.05 (1.76) .035 1.79*# of Females 5.00 (3.79) 6.00 (3.42) 1 •79*# of Overlapping Ties 3.08 (2.36) 2.35 (1 .46)if of Overlapping Females 1 .69 (1 .49) 1 .75 (1 .29)# of Individuals Listed Seen 5.62 (4.43) 4.10 (2.83)Outside of Workif of Female Seen Outside 3.08 (3.66) 3.00 (2.42)of WorkLocation of Individuals Listed:Same Site 2.08 (1.61) 3.20 (3.16)Same City/Different Site 3.15 (3.89) 3.75 (3.57)Different City/Different Site 4.00 (4.44) 1 .10 (2.05) .034 -2.61 *if of Individuals in 5.15 (4.53) 4.1 5 (3.23)Same Function# of Different Functions 2.08 (1.80) 2.15 (1 .60)Hierarchical Rank -0.58 (0.56) -0.37 (0.62)Hierarchical Range 1 .06 (0.33) 0.94 (0.40)if of Supervisors 0.38 (0.51) 0.25 (0.44)if of Subordinates 1 .23 (1 .30) 1 .75 (2.05)Density 0.26 (0.25) 0.27 (0.30)Frequency of Contact 3.07 (0.85) 3.30 (1.07)#p<.lO; *p<•051Mann-Whitney U non-parametric test; p < .10 listed79number of males and the number of individuals listed at a different site in a different city onthe basis of Mann-Whitney U non-parametric tests.With only 33 respondents, the promotion models, which included a covariate, anindependent variable and control variables (as outlined in Chapter 3), could not be testedusing regression analysis without compromising statistical power. Consequently, only thecovariate and the gender of the individuals were included as predictor variables in theregression analyses. The lone significant result for both the instrumental and expressivenetworks (Tables 6.1 2 and 6.13) was that the women tended to draw fewer expressive tiesfrom different sites in different cities in comparison to the men. On the basis of theregression analyses, the gender composition of the expressive networks differed betweenmen and women. Basically, the personal work networks of the men and women did notdiffer.Hypothesis 2. Two additional hypotheses were tested in the promotion study. Thefirst studied the turnover in men’s and women’s expressive networks following promotions.Analysis was limited by the operationalization of turnover and the small respondent sample.The 33 respondents were compared9on the number of individuals added to, dropped from,and remaining in their expressive networks after their promotions. Means are listed in Table6.14. The expressive networks for both genders remained relatively the same size from thefirst data collection to the second. However, there was turnover in the expressive networksof both men and women. Though there were no significant gender differences as to thedegree of turnover, men and women added approximately four individuals on average whiledropping approximately four people from their expressive networks. Regression analyseswere conducted with the gender of the respondent and time with employer serving aspredictor variables. There were no significant beta coefficients for the dependent variables:individuals added to, dropped from, or remaining in the network.No formal hypothesis was offered as to the extent of turnover in men’s and9The Wilcoxon related-sample method was used to compare men and women.80TABLE 6.14Gender Comparison of Turnover in Expressive and Instrumental Networks:Promotion Respondent Sample (n = 33)Ties Ties Ties Time 1 Time 2Added Dropped Remaining Size SizeExpressive Network:Male Means 4.08 4.25 3.58 7.83 7.67Female Means 3.86 4.24 5.14 9.38 9.00p-value1 .94 .64 .12 .31 .47Instrumental Network:Male Means 7.25 5.33 2.08 7.42 9.33Female Means 6.33 5.76 1 .71 7.48 8.05p-value1 .46 .46 .46 .68 .52U non-parametric test.81women’s networks; however, the results are provided in Table 6.14. Again, there were nogender differences in the number of individuals added, dropped, and remaining 8 monthsfollowing the promotions. Only two individuals on average remained in the instrumentalnetworks, and overall, the networks grew in size following the promotion, though notsignificantly.Hyrothesis 3. This hypothesis considered the change in the hierarchical rank andrange of women’s expressive networks. It was posited that following promotions thepositional range of women’s ties would increase. Related-sample, non-parametric testswere conducted comparing the hierarchical rank and range of women’s ties between Time 1and Time 2. There was no significant hierarchical rank change, -0.40 in Time 1 versus-0.37 in Time 2. A significant hierarchical range change was found as the positional rangeof the women’s ties dropped from 1.20 to 0.94 (p = .028).In summary, this thesis set out to explore the extent of similarities and/ordifferences in men’s and women’s personal work networks. Few differences were found inmen’s and women’s instrumental networks. There were more differences between theirexpressive networks. Discussion of the results follows in the next chapter.82Chapter Seven:DISCUSSION and SUMMARYThe tables presented in Chapter Six and its corresponding appendices belie thesimple intent of this thesis, which was to study the similarities and differences of men’s andwomen’s personal work networks. The assumption currently established in the literature isthat there are differences between men’s and women’s personal work networks.Considering the results from the present non-transition study, there were few differencesbetween men’s and women’s instrumental work networks; however, there were differencesbetween men’s and women’s expressive work networks. In discussing the results and theirpossible implications, this chapter is divided into four sections: (1) the non-transition study,(2) the promotion study, (3) the thesis’ limitations, and (4) the status of (and future for)personal work network research.THE NON-TRANSITION STUDYThis study was designed to explore gender personal network similarities anddifferences, and I begin by reviewing the gender difference results.Gender EffectsThe lack of significant gender differences in the instrumental networks was notsurprising. In order to effectively and efficiently complete the tasks associated with theirjobs, men and women should have relatively similar instrumental networks. In this study,men and women reported similar size networks, comprised of people from similar locations,functions, and hierarchical levels. The men’s networks had greater density, and this resultcorroborates past research that demonstrated that men tend to be more activity- and grouporiented with others in their networks (Aukett et al., 1988; Sapadin, 1988).The intriguing instrumental network result was the gender composition of men’s andwomen’s networks. Men’s instrumental work networks displayed greater homophily thandid the women’s instrumental work networks. This finding coincides with those of Ibarra83(1992, 1993b). At issue is whether it is “problematic” for women’s networks to have ahigher proportion of women than men. Because women tend to be outside the dominantcoalition, thanks in part to the glass ceiling effect (see Morrison, White, & Van Velsor,1987), networks with greater female representation could be dominated by women withlower positional rank. Yet, this was not the case in this study, as the men and womenrespondents did not differ in the hierarchical rank or range of their instrumental ties.Moreover, in both the men’s and women’s expressive networks, women accounted forfewer than 50% of the individuals listed (45% for the women respondents and 33% for themen).Gender differences were most evident in men’s and women’s expressive networks,as discussed below.Expressive network size. The men from the three organizations did not list moreexpressive ties than did the women. Though not significant, women had slightly largerexpressive networks (8.81 versus 8.38). This finding is contrary to that of Mayhew andLevinger (1976), who concluded that women tend to develop more intense ties than men,and consequently, it may be harder for women to maintain as many expressive ties as men.One plausible explanation for the lack of differences between the sizes of men’s andwomen’s expressive networks is the expressive tie definition provided in the questionnaires(see Chapter Five). The definition is an “activity-oriented” (or allegedly male) definition, andconsequently, the definition ignores the “emotion-sharing” (or allegedly female) aspects offriendship. Specifically, women tend to disclose more personal information and establishmore emotionally intimate friendships (Aukett et al., 1988; Sapadin, 1988). It is possiblethat a more complete definition of friendship may have led the sample respondents --especially the women -- to list different individuals (while excluding others), thus resulting inexpressive networks differing in size between the genders.Homophily. The results of this study confirmed the differentiated gender pattern ofmen’s and women’s expressive networks. Men’s expressive networks had, on average,5.92 males and 2.46 females; women’s expressive networks, 3.5 males and 5.31 females.84This homophilous pattern is particularly significant when one considers the sex of thoseindividuals who were listed as both instrumental and expressive ties, and the sex of thoseindividuals seen socially outside of work. lbarra’s work (1992, 1993a) demonstrated thatmen and women maintain more homophilous ties in their expressive networks than in theirinstrumental networks, on the basis of the exchange of resources. Lincoln and Miller(1979) concluded that ascribed attributes influence expressive ties more than instrumentalties because social homogeneity makes communication easier and behavior morepredictable, and fosters relationships of trust and reciprocity.When considering a respondent’s expressive network, the individuals listed weremore likely to be of the same gender as the respondent. This trend of same-sex friendswas also evident if these friends were seen outside of the workplace and/or were also listedas instrumental. It is possible that those ties, which overlap between the instrumental andexpressive networks and are also seen outside of work, represent the strongest ties.1 Thisis critical because Granovetter (1982) indicates that stronger ties invariably exhibit a greatermotivation to be of assistance, and thus, this would mean that men and women end uprelying the most on same-sex individuals within the workplace.Overlap between instrumental and exDressive networks. The male and femalerespondents did not differ significantly on the number of overlapping ties between theirinstrumental and expressive networks. This result is also contrary to past research, as mensupposedly acknowledge who their friends are on the basis of what they do with them(Roberto & Kimboko, 1 989), and consequently, men tend to develop expressive tiesthrough their involvement in various organizational activities (i.e., work activities). Ibarra(1 993a) argued that women invariably have to look outside their work activities in order todevelop expressive relationships with other women (Ibarra, 1993a), thus decreasing the1Tie strength is established on the basis of: (1) the frequency and recency of contact,(2) the emotional intensity and level of mutual confiding, (3) the level of reciprocity builtinto the relationship and/or (4) the number of roles played by the individuals (e.g.,instrumental and expressive in the case of this study) (Granovetter, 1 973; Krackhardt,1990).85likelihood of overlapping ties.However, this may not necessarily be correct, as individuals may have some choicein deciding with whom they work. The individuals with whom one works are not entirelyprescribed by the task interdependencies of the organization’s work flow, and individuals, ingeneral, may have some leeway in developing instrumental ties. The fact that there may beworking relationships, prescribed by the company’s structure, does not preclude individualsfrom developing additional, non-prescribed instrumental ties. If so, these additionalinstrumental ties could come from already established expressive ties. Expressive networksare invariably more homophilous, and overlapping ties tend to be homophilous. Therefore,the extent of homophily in overlapping relationships could be the result of homophilousexpressive ties becoming instrumental ties instead of the instrumental ties becomingexpressive ties. For example, a woman could develop a friendship with another woman inthe organization, and even though there was no task interdependence between the two atthe start of their friendship, the friendship could end up being instrumental to theirperforming their respective jobs.Range. The results of this study attest to the greater functional range in women’sexpressive networks. Women tend to have expressive networks with higher functionalrange because they must look outside their work activities in order to develop friendships(lbarra, 1993a, 1993b). Though not significant, the number of expressive ties working inthe same function was lower for women in comparison to men (4.64 versus 5.28, p =.280), but the number of different functions from which the expressive ties were drawnwas greater for women (2.72 versus 2.18, p = .044). Furthermore, the men were morelikely to include their supervisors (but not their subordinates) in their personal worknetworks than were women (.38 versus .23, p = .034). This finding corresponds to thebelief that men supposedly develop friendships on the basis of doing things with people(Sapadin, 1988). Yet, with a higher proportion of supervisors in their expressive networks,the hierarchical rank of men’s expressive ties did not differ significantly in comparison tothe women. This is contrary to Brass’ conclusion (1985) that men tend to have ties with86higher positional range because men invariably comprise the dominant coalition in mostorganizations.Density. Men and women did differ significantly in the density of their expressivenetworks (.24 for men versus .16 for women, p = .002). Logically, if men do developrelationships through activities and shared experiences, then one could posit that men’swork networks are more dense than those of women. On the other hand, women developmore dyadic relationships that would not necessarily develop into dense clusters ofrelationships.Frequency of Contact. Finally, men and women did differ significantly in thefrequency of contact that they had with the individuals listed in their respective expressivenetworks. Through their various activities (including work), men would seemingly have agreater chance to be in contact with their friends. Women would not necessarily have thesame frequency of contact with their friends because of the dyadic, intense nature of theirties. This difference in the frequency of contact could partially explain why men andwomen have expressive personal work networks of similar size. Women could maintain asmany ties as men, even though these ties may be more intense and dyadic, because thefrequency of contact is lower for women in comparison to men.Interpreting the Instrumental and Expressive Network DifferencesThe expressive network non-parametric tests and regression analyses wereconsistently significant when one considers the gender of the ties, the density of the ties,and the frequency of the contact with the individuals listed. These results paralleled theinstrumental network results, and corroborated previous research findings (Ibarra, 1 992,1993b; Lincoln & Miller, 1979). Both Ibarra (1992, 1993b) and Lincoln and Miller (1979)found that men’s instrumental and expressive networks were homophilous, whereaswomen’s expressive networks were homophilous but their instrumental networks exhibiteda more differentiated pattern.The lack of extensive gender personal work network differences may appearcounter-intuitive to some. This is especially true if we start with the assumption that there87should be differences. From the outset, I have argued that such an assumption may beincorrect, and the results of the non-transition study cast doubt on the viability of thegender personal work network differences assumption. Disregarding the possible limitationsor methodological deficiencies of this study, company effects and time effects provide twopossible explanations for the lack of differences between men and women.Company effects. Companies will differ in the total number of women employed,the number of women in management positions, the number of operating sites, the numberof hierarchical levels, the number of functional departments, and the emphasis placed onworking with others. These company differences will impact the networks that developwithin the workplace. In testing for company main and interaction effects, I found that anemployee’s company was an important predictor variable. Company characteristics mayimpact the structure of the personal work networks, and consequently, companycharacteristics will need to be considered in future studies that compare and contrast men’sand women’s personal work networks.The data analyses did not find consistent gender differences among the threecompanies. The only consistent gender finding (see Tables 6.8 and 6.9) was that,regardless of the company, there were more gender differences among the dependentvariables in the expressive networks than in the instrumental networks. The instrumentalnetwork differences included: the number of males and females listed and the density ofthe networks. On the other hand, the expressive network differences included: the numberof males and females listed, the number of overlapping females, the number of women seenoutside of work, the work location of the individuals listed, the number of individualsworking in the same function as the respondent, the number of different functions listed,the number of supervisors listed, the density of the network, and the frequency of contact.The forestry company results provide a graphic example of how companycharacteristics can impact the networks of both men and women. The forestry company isa male-dominated industry. There are few women managers and supervisors, and there arefew women employees who work in the mills. The small proportion of women employees88has an impact on the number of women listed in women’s expressive networks. Within theforestry company, women’s expressive work networks, on average, had nearly five menand four women. The women in the insurance and banking companies had a clear majorityof female-to-male friends (bank: 5.3 females to 3.1 males; insurance: 6.0 females to 3.3males).Lincoln and Miller (1979) concluded that one’s sociometric characteristics (e.g., sex,race, level of authority) tend to have a greater impact on the expressive relationships that aperson develops in the workplace, whereas structural characteristics (e.g., the formaldivision of labor) tend to have a greater impact on the development of instrumentalnetworks. Yet, I would argue that explaining differences in personal work networks isdependent upon the characteristics of both the individual and the company. For example,the extent that individuals develop ties with individuals of the same sex is the result ofinduced homophily and choice homophily (Ibarra, 1993b). Women associating with otherwomen is not only dependent upon having just men to work with (i.e., induced homophily),but is also dependent upon their wanting to associate with other women (i.e., choicehomophily).The relative effect (and importance) of choice versus induced homophily is open todebate and further study. What cannot be overlooked in future research is the role thatcompany or industry-specific characteristics play. In this thesis, company was confoundedby industry. The response sample was drawn from three companies with very differentnormative missions that affected the gender composition of the organizations. The forestrycompany represents an industry that is production-oriented, where the work is physicallydemanding. The insurance company, though providing a service, is cost-control oriented,while the bank is service-convenience oriented.Future research will have to consider the effects the company (or industry) have onpersonal work networks; however, there is no clear answer, on the basis of this study’sresults, as to whether there is a company main effect (shift model) or a company interactioneffect (moderating effect). With a main effect, the inclusion of company dummy variables89would explain additional variance for each of the dependent variables, whereas a moderatoreffect would mean that the relationships between the predictor variable (i.e., gender) andthe dependent variables would depend upon the moderator variable (i.e., the company).Minimally, there is a company main effect (shift model) because at least one of thecompany dummy variables was significant in all eleven instrumental network variables and10 of the 18 expressive network variables. Either the bank and/or the forest companydummy variable was significant on 9 of 11 instrumental variables and on 8 of 1 8 expressivevariables, adding further evidence of company main effects. The introduction of interactionterms (in the Moderator Regression Model) resulted in significant company interactioneffects for three instrumental network and six expressive network variables. However, theprobability of finding these significant company interaction effects was increased for tworeasons. First, in creating the Moderator Model, up to 10 interaction terms were added tothe Shift Model. Second, the majority of the R-squared in the Shift Model were below .50.It is likely that adding that many interaction terms when much of the variance remainedunexplained (on all of the dependent variables studied) would increase the likelihood of asignificant interaction effect.There is limited evidence to suggest that attention should be paid to the possibilityof interaction effects. However, these company interactions are difficult to interpret. Yet,company as a main effect is a definite confound and will have to be considered in futureresearch. Of the three models tested, I consider the shift model, which controls forcompany main effects, the most viable for future research.Time effects. In discussing the results with the individuals who served as contactsat the data sites, I found that they were surprised by the lack (and size) of differencesbetween men’s and women’s personal work networks, and one question kept coming up:“I wonder what the results would have looked like 10 years ago? These individualshonestly believed that there may have been more gender differences in personal worknetworks 10 years ago, and they may have been right. Consider for a moment that therehave been profound changes in the economy (i.e., changes in the supply and demand of90certain jobs), workforce demographics (i.e., more women moving into the labor force formyriad reasons), and the law (i.e., equal opportunity and affirmative action legislation in theUnited States). We have also seen an influx of women attending business schools and anincrease in networking seminars for women. Finally, over the past 10 years, more womenhave been moving into management positions, though still not in proportion to theirrepresentation in the workforce.Given these on-going changes, one could ask how long will the personal worknetworks differences discussed above generalize to the organizations in this study?Cronbach (1975, pp. 122-1 23) writes:Generalizations decay. At one time a conclusion describes the existing situationwell, at a later time it accounts for rather little variance, and ultimately it is validonly as history. The half-life of an empirical proposition may be great or small.The more open a system, the shorter the half-life of relations within it are likelyto be.An example of Cronbach’s point is Kanter’s assertion of gender personal work networkdifferences. In 1 977, Kanter’s assertions were correct. However, in the almost 20 yearssince her seminal piece, Men and Women of the Corporation, was published, no one woulddare to argue that times have not changed. With changing times have come changingpersonal work networks. There may have been more pronounced personal work networkdifferences between men and women 10 years ago (e.g., average hierarchical rank ofindividuals listed); however, today, considering these organizations only, gender may not beas important as company characteristics in explaining the differences in personal worknetworks.If men’s and women’s personal work networks have become more comparable inthe past 10 years, the question of interest is “Why?” Possible explanations include, but arenot limited to: (a) the effects of career transitions, (b) changing organizational structures,and/or (c) changing individual characteristics. For example, more women are moving intosupervisory and managerial positions, organizations have become leaner with a greateremphasis on teamwork, and consequently, men and women have become more accustomed91to working together.Significant differences in men’s and women’s personal work networks were found,and these differences were limited to the gender composition, density, and frequency ofcontact. Consequently, the results do not attest to the current assumption that there arevast differences between men’s and women’s personal work networks.I have argued that the lack of extensive differences may be due tocompany/industry or time effects. However, true differences may have also been maskedby variable operationalizations used in the analysis of the data. The lack of significant childrearing responsibilities findings provides one such example. In testing for child-rearingresponsibilities, a dummy variable was created. Individuals with children 18 years old oryounger constituted one group. Individuals with no children or children older than 1 8 yearsof age comprised the other. It is possible that having young children may have a greaterimpact on expressive networks than having teenage children. If so, differences may havebeen concealed by creating one category for child-rearing responsibilities.THE PROMOTION STUDYThe promotion study results provide little information, given the research questionsand hypotheses and the small sample size. There were no significant differences foundbetween the instrumental and expressive networks of the 33 respondents; however, thatdoes not mean that there were no differences. Differences may not have been detecteddue to sampling deficiencies. The respondent sample was very small (33 individuals out ofa possible sample of 162), and these results may not adequately represent the companyfrom which the sample was drawn. Yet, even if there had been significant differences, thegeneralizability within the bank would have been limited because of the sample size. Therewere few male and female differences prior to the promotions,2and one would think that,2From the data provided at (or near) the time of their promotions, the men and womendiffered significantly on: the number of individuals listed and the number of males listed intheir instrumental networks; and the number of males and the hierarchical range of the tieslisted in their expressive networks.92prior to the promotions and even 8 months after the promotions, the respondents’expressive and instrumental networks would exhibit some of the same differences found inthe non-transition study. This is especially true when we consider homophily. The non-transition study corroborates the significant findings from past research that men andwomen develop homophilous expressive networks; however, this association pattern wasnot evident in the promotion study expressive networks.Hypothesis 2 posited that there would be more turnover (i.e., the number ofindividuals added and dropped) in the expressive networks of the male respondents than inthe women’s expressive networks, but the results did not support this hypothesis. (Therewere no significant differences in the changes between men’s and women’s instrumentalnetworks either.) Again, the small sample may have adversely impacted my ability todetect gender differences, but the results are still interesting because of the extent of thechanges. Eight months following the promotions, the average number of instrumental tiesremaining was approximately two (or roughly 27% of the individuals listed as instrumentalprior to the promotion). The percentage of ties remaining in the expressive networks was45% for the men and 55% for the women.It would seem likely that these changes would affect the social support present andavailable in the individuals’ instrumental and expressive personal work networks. Followingtheir promotions, individuals are “learning the ropes of their new jobs; however, fellowworkers, who in the past have provided advice, support, and friendship, are not available orable to provide support. It is ironic that at the very time an individual would need socialsupport, changes to his or her network may prevent him or her from receiving the support.Finally, Hypothesis 3 considered the change in hierarchical rank in women’sexpressive networks. It was posited that, following their promotions, the positional rank ofthe women’s ties would increase (and thus become more comparable) to those of the men.However, following the promotions, the hierarchical ranks of the women’s ties were higherthan the men’s, though the difference was not significant. The small respondent samplecould explain this. Alternatively, this hypothesis was based upon the assumption that the93positional rank of women’s expressive networks would be lower than that of the men. Thenon-transition study found there to be no hierarchical rank differences between men andwomen, even though the men and women of the bank non-transition respondent sampleexhibited a greater positional rank difference that neared significance (p = .073). In otherwords, this hypothesis now appears meaningless when we consider the hierarchical rankfindings of the non-transition study. If the average hierarchical rank of men’s and women’sexpressive networks did not differ greatly prior to the promotion, there was no opportunityfor women to improve the hierarchical rank of their expressive networks followingpromotions when compared to men.The findings of the non-transition study would have definitely helped in the designof the promotion study. The non-transition study results, coupled with the turnoverfindings in the promotion study, may help in the design of future studies, as the needremains for further research on the effects of promotions (and other career transitions) onmen’s and women’s personal work networks.3Personal work network ties do not remain static, and this poses an interestingchallenge to network research. At the individual level, networks change or evolve eithergradually or quite abruptly, depending upon individual career transitions and/or organization-wide reorganizations. Progress in theory development can come about only with a betterunderstanding of how personal work networks change and how quickly they change overtime. Such an understanding is important if we are to understand the impact changes topersonal work networks have on social support availability. However, changes to personalwork networks are also occurring as the result of societal changes. In effect, there couldbe events occurring at the individual, company/industry, and societal level that affect thecharacteristics of personal work networks. Any theory development will have to addressthe evolution of networks over time if the findings are to have any sustained3Research on the evolution of networks will have to incorporate frequent, repeatedmeasures. Testing for differences eight months apart will not be sufficient, as subtlechanges (and the reasons behind these changes) may be missed.94generalizability.Future research questions or issues have become evident. First, does the amount ofturnover (i.e., the number of ties lost and added) in individuals’ personal work networksdepend upon the type of career transition (e.g., promotion versus lateral move)? Second,how much time is needed to “rebuild” one’s personal work network following a careertransition, and does the amount of time differ between companies/industries? Third, howquickly are expressive ties lost following a career transition that involves a geographicmove? Finally, how do the changes in a personal work network, following a careertransition, compare to the normal changes that networks undergo?LIMITATIONSThere are a number of limitations associated with the research reported in thisthesis. First, both the non-transition and promotion studies used convenience samples, sothe generalizability of the results beyond the three companies comprising the sample islimited. Moreover, the promotion study respondent sample was not sufficiently large, soagain, the findings cannot be generalized beyond the actual respondents.Second, the correlation design prevents any discussion of the causation ofdifferences (e.g., structural or dispositional factors) between men’s and women’s personalwork networks. Instead, the cross-sectional design provides a “snapshot” of personal worknetworks, and “... accepts the natural range of variables, instead of shaping conditions (likemanipulative research) to represent a hypothesis” (Cronbach, 1975, p. 124). In thepromotion study, data were collected over a period of time; however, the tracking of thechanges was hampered by the sample size.Third, the thesis considered several sociometric variables (e.g., age, job category,child-rearing responsibility) in an attempt to isolate personal work network differences.However, there was no attempt to study individual psychometric differences (e.g., shyness,aggressiveness, sociability) and/or skill-level differences (e.g., ability to meet people andmaintain on-going relationships). Furthermore, there was evidence of company effects on95personal work networks; however, no specific company variables (e.g., the number ofhierarchical levels at each organization, the number of women employed, and/or the numberof women in managerial or supervisory positions) were studied. All that is known is thatthe gender of the individual and the company he or she works for may affect thedevelopment of his or her personal work network. We do not know what characteristicsspecific to the gender of the individual or to the company affect the make-up of personalwork networks.Fourth, I did not consider indirect personal network links, and consequently, thedetection of deeper structural differences between genders was not possible (Ibarra,1993b). Individuals are indirectly tied through their direct ties to many other individualswho work in (and outside) the organization. What differences there are in one’sinstrumental and/or expressive networks (and the impact of these differences) may be eithermagnified or attenuated by studying both indirect and direct ties. For example, one’s directties may not be able to provide needed information that will help him or her to land animportant account; however, someone he or she is directly tied with may know someoneelse in the company who can provide the information. In effect, indirect ties can make upfor deficiencies (or gender differences) in the direct ties of individuals, and consequently,studying the impact of direct ties and not indirect ties on social support would beincomplete (Ibarra, 1993b).Finally, there are concerns surrounding the reliability of individual networkrepresentations of direct ties via questionnaires. The reliability check conducted in thisthesis demonstrated that a large number of direct ties were added and dropped4only onemonth after the respondents had first provided a representation of their instrumental andexpressive networks. There is no concrete evidence that these changes to their personal4Approximately 3.5 ties were added and 3.8 dropped, on average, from the instrumentalnetwork. In the expressive network, roughly 1 .7 ties were added and 2.0 were dropped.96work networks in one month’s time were due to actual network changes5,to oversight, orto both. Until respondent oversight can be ruled out, the reliability of participant self-reports remains a concern. In the future, researchers may have to provide respondents withlists of names or with the chance to revisit and amend the lists at a later date to ensure anaccurate representation of their personal work networks.THE STATUS AND FUTURE OF PERSONAL WORK NETWORK RESEARCHThis thesis set out to address gaps and assumptions in the literature regardingmen’s and women’s personal work networks. In conducting the non-transition andpromotion studies, more questions about personal work networks (and weaknesses incurrent research) have been uncovered, and these questions (and research weaknesses)attest to the great research potential surrounding personal work networks. What follows isan attempt to structure these myriad questions with two global questions. The firstquestion is: What is meant by the term “personal work networks?” The second is: Whatrole(s) do personal work networks play in individuals’ lives (i.e., what do we get frompersonal work networks)?Personal Work NetworksCurrent research, including this thesis, has simplified what is meant by the term“personal work networks” in order to collect data. Consequently, this research has focusedon within-company, direct ties, which simplifies the construct of personal work networks.This simplification has made the study of personal work networks easier; however, thissame conceptualization has failed to address and to acknowledge adequately the complexityof personal work networks.What is clearly missing in the study of personal work networks is an understandingof how and where people work. Researchers (myself included) have studied large,5Test-retest methods used to establish self-report reliability may be inadequate. A testretest correlation may be attentuated because personal work networks may undergoconsiderable membership changes over a period of time (e.g., one month).97hierarchically organized companies, thus aiding themselves in the study of within-company,direct ties. Work, though, is not necessarily hierarchically organized. For example, withinmany organizations, there are ever-changing project management teams and/or temporarycommittee assignments. Many companies are forming domestic and international allianceswith other organizations. Also, the increasing numbers of people telecommuting havealtered work arrangements. How and where people work with others has varying effectson the characteristics of individuals’ personal work networks (i.e., the gender and racialmake-up, range, density, and frequency of contact).Instrumental and expressive work relationships are not limited to being within asingle organization. Individuals can have external and internal instrumental and expressivework relationships. Aldrich’s research (see Aldrich, 1989; Aldrich et al., 1989) studiedentrepreneurs’ instrumental relationships (e.g., accountants, bankers, lawyers, suppliers)outside of the organization.External work relationships may be especially critical to individuals who own or whoare employed by small organizations.6 Smaller businesses may have to rely on outsourcingof certain functions (e.g., marketing, sales, production, or administration, for example)because they do not have the human resources to staff each and every position. Manylarge businesses are also relying more and more on outsourcing (Harris, 1993). There areestimates that three out of ten large American industrial firms now outsource half theirmanufacturing (Bridges, 1994). Some interesting questions include:What do the personal work networks of individuals working for small versuslarge organizations look like (i.e., how similar or different are the personalwork networks on the various network characteristics)?What is the percentage of internal to external relationships in theinstrumental and expressive work networks of individuals working for smallversus large organizations?Do individuals working in different functional departments of largeorganizations have differing percentages of internal to external relationships6Small businesses employ more individuals than do large businesses. In fact, within thepast 10 years, the proportion of the work force employed by Fortune 500 companies hasfallen from 30% to 13% (Harris, 1993).98in their instrumental and expressive work networks?Do the instrumental and expressive work networks differ betweenindividuals working part-time versus full-time?Such questions would force researchers to consider both internal and external relationshipsand the personal work network differences between individuals working for small or largeorganizations.Direct versus indirect ties. Thus far, I have discussed internal-external andinstrumental-expressive relationships, but these are only two ways to represent ties.Current research focuses on direct ties, but as noted in the limitations section, researchcould also study indirect ties, as support and resources can flow to people from individualswith whom they are indirectly tied. However, who qualifies as an indirect tie is at issue.Milgram (1967) concluded from his small-world problem studies that each individual isseparated, on average, from every other person by six individuals. Consequently, should allindividuals not directly tied to a focal person be considered as an indirect tie? I would argueno.The indirect ties most salient to our understanding are likely those individualsseparated by one person (i.e., a direct tie) or two people (i.e., a direct and indirect tie).These are the individuals who could, more than likely, provide us with needed support orbenefits in a timely fashion. The remaining individuals would comprise a pool of potentialrelationships. These are individuals who could eventually become part of our networks (aseither direct or indirect ties) depending upon the circumstances. Such circumstances couldinvolve, for example, a person changing professions and coming to work for the samecompany and in the same department which employs me. This would be an example of apreviously potential tie changing into a direct tie. Or, this person could form a workingrelationship with my superior, thus turning a once potential tie to an indirect tie.7Further complicating the study of personal work networks is the fact that direct ties7This example highlights how quickly a person’s personal work network can changewhen one considers both direct and indirect ties.99can become latent. For example, an individual could be instrumental to my jobperformance, and I would list him/her when asked to generate a list of instrumental ties.However, this individual may have been instrumental at one time to my job performance,but is no longer. Such a relationship would now be latent, and I would not, in all likelihood,list him/her when asked to generate a list of instrumental ties. Yet, though this relationshipis currently latent, it is relatively easy to re-establish the relationship in the future becauseof my past experience working with the person. In effect, personal work networkrelationships can be labelled as direct, direct but latent, indirect, or potential.Ultimately, the study of “personal work networks” will have to focus on more thanjust direct, within-company relationships. By expanding what is meant by a personal worknetwork, the characteristics of an individual’s personal work network may change. Suchchanges will have an effect on the comparisons subsequently made between individuals onthese personal work network characteristics. First, we have no idea how large personalwork networks are when considering both direct and indirect ties. Second, questionsremain as to how many direct ties a person can handle, as previous studies have limited thenumber of instrumental and expressive relationships a respondent could provide. Finally,the inclusion of external and indirect ties (in conjunction with internal, direct ties) mayincrease the location, functional, and hierarchical individuals in the networks along withattenuating the homophily, density and frequency of contact. (As will be discussed below,a more complete representation of what a personal work network is will also affect the role(and our interpretation of that role) of personal work networks.)The evolution of personal work networks. I have been arguing to this point that thestudy of personal work networks (and its focus on direct, within-company ties) has beenincomplete. What is also missing from our understanding of personal work networks is howthey develop and change over time. Questions can be generated concerning the fluidity ofpersonal work networks. For example, how many individuals are added and become latentin a given time period, say a month, six months, or a year.Numerous questions also remain regarding how individuals become part of (or leave)100another individual’s personal work networks. What or who plays a critical role in thedevelopment of a relationship between two people? Is it the company’s structure? Thematching of comparable traits or values between two individuals through socialcomparison? Or, is it events such as transfers or promotions? To what extent does anindividual have a say in the building of his/her personal work networks? How strategic canindividuals be in building instrumental versus expressive relationships with other people?Roles of Personal Work NetworksA better understanding of what is meant by the term “personal work network” andhow individual personal work networks are similar or different and how they change overtime is necessary if we are to understand the roles of personal work networks. The studyof personal work network differences between individuals (especially between males andfemales) in isolation is informative and interesting to a point. It was noted in theintroduction that understanding the similarities and differences is critical, becausedifferences in association patterns at work can affect the opportunities8and amount ofsocial support available to individuals. In other words, having information on the similaritiesand differences of personal work networks may provide a better understanding of theavailability of social support, as gender research has overlooked the differences in socialresources that men and women have at their disposal (Hare-Mustin & Marecek, 1 990b).More than lust social suprort. The real potential for personal work networkresearch is through the combining of the study of personal work network characteristicsand the affect of these characteristics on what individuals receive from their personal worknetworks. Up to this point, the term “social support” has been used to represent the“benefits” received by individuals through their associations with other people. Like theterm “personal work networks,” social support is a complex construct.Support can be viewed as a physical, emotional, or symbolic contribution to8Recent research (Ibarra, 1993b, 1995) has focused on explaining the effect personalwork network differences have on differential career opportunities and outcomes for womenand minorities.101individuals increasing their net stockpile of emotional capacity to cope with change (Walter& Marks, 1981), thus intended to enhance the well-being of the recipient (Shumaker &Brownell, 1984). Heller and Swindle (1983) note that “... social support is increasinglyviewed as a multidimensional construct, consisting of social network resources, types ofsupportive exchanges, perceptions of support availability, and skills in assessing andmaintaining supportive relationships.” Types of social support include emotional, appraisal,informational (including feedback), and instrumental (Walter & Marks, 1981).However, by focusing solely on social support (when studying personal worknetworks), researchers will miss some very important positive and negative benefits and/oroutcomes. It is true that networks provide for the exchange of support, but personal worknetwork relationships are also sources of status, innovations, competition, motivation, peerpressure, and conflict. Moreover, personal work network relationships allow individuals topool resources when attempting to complete complex tasks.Consequently, personal work networks can be portrayed as systems -- exchangesystems, diffusion systems, learning systems, and/or activation systems. Tangible andintangible benefits and outcomes, including social support, flow through these systems.The effect of one individual becoming part of or leaving one’s network can be quiteprofound. On the positive side, an individual, who had recently undergone some advancedtraining, could share this new information with fellow work colleagues and friends. Also, aperson looking for a job could increase his/her credibility (and ultimately his/her potential jobprospects) by using the name of a direct tie (“So-and-so said that you would be the bestperson to talk with regarding possible job opportunities.”) On the negative side, personalwork networks can also be sources of conflict, peer pressure, and norms. Smith (1989)provides an excellent illustration of the effects of conflict movement through a socialnetwork.This discussion on personal work networks and what personal work networksprovide to individuals is by no means comprehensive. Instead, it should provide theimpression that personal work networks are sources of support, opportunities, benefits, and102outcomes that are not always completely positive.I return now to an earlier point, that the study of network similarities anddifferences is important, but incomplete. A clearer understanding of personal worknetworks and their characteristics is essential if we are to study the impact networkcharacteristics have on the resources9provided by personal work networks. The realpotential of and interest in personal work network research lies in researching suchquestions as:Which network, instrumental or expressive, is best at providing resources, such assupport, status, or credibility?Which resources (and in what proportions) emanate from internal versus externalpersonal work network relationships?Which resources (and in what proportions) emanate from direct versus indirectrelationships?How important is tie strength in the exchange or transfer of various resources?How important is location, functional, and/or hierarchical range in the exchange ortransfer of various resources?How important is the frequency of contact in the exchange or transfer of variousresources?Are the resources provided by personal work network relationships to individualsreally comparable when their personal work networks are similar?What are the characteristics of personal work networks that shield individualsfrom negative resources?Is one type of network (i.e., instrumental-expressive or internal-external) moreimportant in the resources provided to one gender than that of the other gender?How immune is the availability of resources to changes (specifically the loss ofindividuals) to personal work networks?The number of questions increases dramatically when we consider internal-external,instrumental-expressive, and/or direct-indirect relationships. What I am suggesting in raisingthese questions is that the study of personal work network characteristics provides9From this point forward, I use the word “resources” to reflect the support,opportunities, and outcomes that are exchanged, transferred, or shared in personal worknetworks.103researchers with many research opportunities.Two warnings. When considering these questions, however, there are twoimportant warnings. First, there is the tendency for researchers to focus on the effect ofdifferences without considering the similarities. This is particularly true in gender research.Hare-Mustin and Marecek (1990b, p. 30) write:the view of male and female as different and opposite and thus as havingmutually exclusive qualities transcends Western culture and has deephistorical roots... (there is an) inclination to emphasize differences...The examination of gender differences obscures the examination of gender similarities, andsuch questions about gender differences “often imply a trait view of behavior that obscuressituational influences on behavior” (Unger, 1990, p. 104). Similarities in networkcharacteristics may not necessarily guarantee similar benefits to individuals (Burt, 1992).Personal network characteristic differences need to be studied for their effect on resources;however, individual differences and situational differences must also be examined at thesame time.Second, to talk of personal work network relationships without differentiatingwhether they are instrumental or expressive, internal or external, or direct or indirect maymask important similarities and/or differences between individuals (e.g., genders,minorities). For example, how would a researcher explain differential resource outcomesafter finding that the personal work networks of these individuals did not differ?Notwithstanding possible methodological deficiencies, individual differences in recognizing,accepting, and reporting these resources may provide one explanation. However, theresponse to the question would differ if I added that, in conducting the study, theresearcher considered only internal and direct instrumental ties, ignoring indirect, externalties.SUMMARYThe non-transition study, even with its limited generalizability, does provide104evidence that men’s and women’s personal work networks may not be all that differentwhen we consider personal work network characteristics other than gender compositionand density and the frequency of contact between individuals in the network. The non-transition study also attests to possible company-industry and/or time effects on thecharacteristics of personal work networks.The future study of personal work networks will not be easy, as there is more to thestudy of personal work networks than just direct, within-company relationships. External,indirect ties also play critical roles in the resources provided to individuals because of theirassociation patterns. Ultimately, research will have to focus on the positive and negativeresources that flow through personal work networks, regardless of network similarities anddifferences. Differences between men’s and women’s personal work networkcharacteristics (and how men and women develop and nurture their personal worknetworks) may impact the resources that are derived from a person’s network; however,differences in how men and women use their personal networks are also important to ourunderstanding of what resources reach individuals.105REFERENCESAiken, L.S., & West, S.G. (1991). Multiple regression: Testing and interpretinginteractions. Newbury Park, CA: Sage.Aldrich, H., & Reese, P.R. (October 1994). Gender gap, gender myth: Does women’snetworking behavior differ significantly from men’s? Presented at the University of BritishColumbia, Faculty of Commerce and Business Administration, Vancouver, B.C.Aldrich, H. (1989). Networking among women entrepreneurs. In 0. Hagan, C. Rivchun& D. Sexton (Eds.), Women-owned businesses (pp. 103-132). New York: Praeger.Aldrich, H., Reese, P.R., & Dubini, P. (1989). Women on the verge of a breakthrough:Networking among entrepreneurs in the United States and Italy. Entrepreneurship &Regional Development, 1, : 339-356.Aukett, R., Ritchie, J., & Mill, K. (1988). Gender differences in friendship patterns.Sex Roles, 19, 5 7-66.Barrera, M., Jr., & Ainlay, S.L. (1983). The structure of social support: A conceptualand empirical analysis. Journal of Community Psychology, 11, 133-1 43.Bell, R.A. (1991). Gender, friendship network density, and loneliness. Journal ofSocial Behavior and Personality, 6, 45-56.Bernard, H.R., Killworth, P.D., Kronenfeld, 0., & Sailer, L. (1984). The problem ofinformant accuracy: The validity of retrospective data. Annual Review of Anthropology,13, 495-517.Blau, P.M. (1988). Inequality and heterogeneity: A primitive theory of social structure.New York: Free Press.Blau, P.M. (1955). The dynamics of bureaucracy. Chicago: University Press.Boneau, C.A. (1960). The effects of violations of assumptions underlying the ttest.Psychological Bulletin, 57, 49-64.Brass, D.J. (1985). Men’s and women’s networks: A study of interaction patterns andinfluence in an organization. Academy of Management Journal, 28, 327-343.Brett, J.M. (1984). Job transitions and personal and role development. Research inPersonnel and Human Resources Management, 2, 155-1 85.Bridges, W. (1994). Job shift: How to prosper in a workplace without iobs. Reading,MA: Addison-Wesley Publishing Co.Burt, R.S. (1992). Structural holes: The social structure of competition. Cambridge,MA: Harvard University Press.Cohen, J. (1968). Multiple regression as a general data-analytic system. PsychologicalBulletin, 70, 426-443.Cronbach, L.J. (1975). Beyond the two disciplines of scientific psychology. American106Psychologist, 30, 116-1 27.Dreher, G.F., & Ash, R.A. (1990). A comparative study of mentoring among men andwomen in managerial, professional, and technical positions. Journal of Applied Psychology,75, 539-546.Dickens, W., & Perlman, D. (1981). Friendship over the life cycle. In S.W. Duck & R.Gilmour (Eds.), Personal relationships 2: Developing personal relationships (pp. 91-1 22).New York: Academic Press.Ely, R. (1994). The effects of organizational demographics and social identity onrelationships among professional women. Administrative Science Quarterly, 39, 203-238.Fenlason, K.J., & Beehr, T.A. (1994). Social support and occupational stress: Effectsof talking to others. Journal of Organizational Behavior, 15, 157-175.Fischer, C. (1982). To dwell among friends. Chicago: University of Chicago.Fischer, C., & Oliker, S. (1983). A research note on friendship, gender, and the lifecycle. Social Forces, 62, 124-1 32.Fischer, J., & Narus, L. (1981). Sex roles and intimacy in same sex and other sexrelationships. Psychology of Women Quarterly, 5, 444-455.Freeman, L.C. & Romney, A.K. (1987). Words, deeds and social structure: Apreliminary study of the reliability of informants. Human Organization, 46, 330-334.Harris, T.G. (1993). The post-capitalist executive: An interview with Peter F. Drucker.Harvard Business Review, 71(3), 115-122.Heller, K. & Swindle, R.W., Jr. (1983). “Social networks, perceived social support, andcoping with stress.” In R.D. Felner, L.A. Jason, J. Moritsugh, & S.S. Farber (Eds.),Preventive psychology: Theory, research, and practice in community intervention (pp. 87-103). New York: Pergamon Press.Ganster, D.C., Fusilier, M.R., & Mayes, B.T. (1986). role of social support in theexperience of stress at work. Journal of Applied Psychology, 71, 102-110.George, J.M. (1992). The role of personality in organizational life: Issues andevidence. Journal of Management, 18, 185-21 3.George, J.M. (1991). State or trait: Effects of positive mood on prosocial behaviors atwork. Journal of Applied Psychology, 76, 299-307.Gouldner, A.W. (1954). Patterns in industrial bureaucracy. New York: Free Press.Granovetter, M. (1982). The strength of weak ties: A network theory revisited. InP.V. Marsden & N. Lin (Eds.), Social structure and network analysis (pp. 105-133). BeverlyHills, CA: Sage.Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology6, 1360-1 380.Hare-Mustin, R.T., & Marecek, J. (1990a). On making a difference. In R.T. HareMustin & J. Marecek (Eds.), Making a difference: Psychology and the construction of107gender (pp. 1-21). New Haven, CN: Yale University Press.Hare-Mustin, R.T., & Marecek, J. (1990b). Gender and the meaning of difference:Postmodernism and psychology. In R.T. Hare-Mustin & J. Marecek (Eds.), Making adifference: Psychology and the construction of gender (pp. 22-64). New Haven, CN: YaleUniversity Press.Hays, R.B., & Oxley, D. (1986). Social network development and functioning during alife transition. Journal of Personality and Social Psychology, 50, 305-313.House, J.S. (1981). Work stress and social support. Reading, MA: Addison-Wesley.Ibarra, H. (1993a). Personal networks of women and minorities in management: Aconceptual framework. Academy of Management Review, 1 8, 56-87.Ibarra, H. (1993b). Untangling the web of interconnections: An exploration ofcompeting explanations for gender differences in Managerial Networks. Working Paper#93-044, Harvard Business School.lbarra, H. (1992). Homophily and differential returns: Sex differences in networkstructure and access in an advertising firm. Administrative Science Quarterly, 37, 422-447.Judge, T.A. (1992). The dispositional perspective in human resource management. InG.R. Ferris & K.M. Rowland (Eds.), Research in Personnel and Human ResourcesManagement (pp. 197-232). Vol. 10. Greenwich, CN: JAI Press.Kanter, R.M. (1977). Men and women of the corporation. New York: Basic Books.Kaplan, R.E. (1984). Trade routes: The manager’s network of relationships.Organizational Dynamics, 12, 37-52.Katz, R. (1980). Time and work: Toward an integrative perspective. Research inOrganizational Behavior, 2, 81-1 27.Kaufman, G.M., & Beehr, T.A. (1986). Interactions between job stressors and socialsupport: Some counterintuitive results. Journal of Applied Psychology, 71, 522-526.Kerlinger, F.N., & Pedhazur, E.J. (1973). Multiple regression in behavioral research.New York: Holt, Rinehart, and Winston, Inc.Kmenta, J. (1971). Elements of econometrics. New York: Macmillan.Krackhardt, D. (1992). The strength of strong ties: The importance of philos inorganizations. In N. Nohria & R.G. Eccles (Eds.), Networks and organizations: Structure,form and action (pp. 216-239). Cambridge, MA: Harvard Business Press.LaRocca, J.M., House, J.S., & French, J.R.P., Jr. (1980). Social support, occupationalstress, and health. Journal of Health and Social Behavior, 21, 202-218.Lincoln, J.R., & Miller, J. (1979). Work and friendship ties in organizations: Acomparative analysis of relational networks. Administrative Science Quarterly, 24, 1 81-199.Lindley, P., & Noble Walker, S. (1993). Theoretical and methodological differentiationof moderation and mediation. Nursing Research, 42, 276-279.108Louis, M.R. (1980). Career transitions: Varieties and commonalities. Academy ofManagement Review, 5, 329-340.Lynch, J.J. (1977). The broken heart: The medical consequences of loneliness. NewYork: Basic Books.Maltz, D.N., & Borker, R.A. (1982). A cultural approach to male-femalemiscommunication. In J.J. Gumperz (Ed.), Language and social identity (pp. 196-21 6).Cambridge: Cambridge University Press.Markiewicz, D., & Devine, I. (June 1994). Women’s friendships in the changing workenvironment: Implications for organizational settings and employee well-being. Presentedat ASAC, Halifax, Nova Scotia.Marsden, P.V. (1988). Network data and measurement. Annual Review of Sociology,16, 435-463.Marsden, P.V. (1987). Core discussion networks of Americans. American SociologicalReview, 52, 122-131.Mayhew, B.H., & Levinger, R.L. (1976). Size and density of interaction in humanaggregates. American Journal of Sociology, 82, 86-110.Milgram, S. (1967). The small world problem. Psychology Today, 1, 62-67.Mitchell, J.C. (1969). The concept and use of social networks. In J.C. Mitchell (Ed.),Social networks in urban situations (pp. 1-50). Manchester, England: University ofManchester Press.Morrison, A.M., White, R.P., & Van Velsor, E. (1987). Breaking the glass ceiling: Canwomen reach the top of America’s largest corporations? Reading, MA: Addison Wesley.Moore, G. (1990). Structural determinants of men’s and women’s personal networks.American Sociological Review, 55, 726-735.Neter, J., Wasserman, W., & Kutner, M.H. (1985). Applied linear statistical models:Regression, analysis of variance, and experimental design. Homewood, IL: Irwin.Nieva, V.F., & Gutek, B.A. (1981). Women and work: A psychological perspective.New York: Praeger.Pinder, C.C., & Schroeder, K.G. (1987). Time to proficiency following job transfers.Academy of Management Journal, 30, 336-353.Reisman, J. (1981). Adult friendships. In S.W. Duck & R. Gilmour (Eds.), Personalrelationships 2: Developing personal relationships (pp. 205-230). New York: AcademicPress.Roberto, K.A., & Kimboko, P.J. (1989). Friendships in later life: Definitions andmaintenance patterns. International Journal of Aging and Human Development 28, 9-19.Sapadin, L.A. (1988). Friendship and gender: Perspectives of professional men andwomen. Journal of Social and Personal Relationships, 5, 387-403.Shumaker, S.A. & Brownell, A. (1984). Toward a theory of social support: Closing109conceptual gaps. Journal of Social Issues, 40, 11-36.Siegel, S. (1956). Nonparametric statistics for the behavioral sciences. New York:McGraw-Hill.Smith, K.K. (1989). The movement of conflict in organizations: The joint dynamics ofsplitting and triangulation. Administrative Science Quarterly, 34, 1-20.Sollie, D.L., & Fischer, J.L. (1988). Career entry influences on social networks ofyoung adults: A longitudinal study. Journal of Social Behavior and Personality 3, 205-225.Spierer, H. (1981). Coping with transitions. In A.C. Eurich (Ed.), Maior transitions inthe human life cycle (pp. 45-52). Lexington, Mass.: Lexington Books.Tannen, D. (1990). You iust don’t understand. New York: Ballentine Books.Tannen, D. (1986). That’s not what I mean! New York: Ballentine Books.Tracy, E.M., & Whittaker, J.K. (1990). The social network map: Assessing socialsupport in clinical practice. Families in Society: The Journal of Contemporary HumanServices, 71, 461-470.UIlah, P., Banks, M., & Warr, P. (1985). Social support, social pressures andpsychological distress during unemployment. Psychological Medicine, 15, 283-295.Unger, R.K. (1990). Imperfect reflections of reality: Psychology constructs gender. InR.T. Hare-Mustin & J. Marecek (Eds.), Making a difference: Psychology and theconstruction of gender. New Haven, CT: Yale University Press.van Groenou, M.B., van Sonderen, E., & Ormel, J. (1990). Test-retest reliability ofpersonal network delineation. In K.C.P.M. Knipscheer & T.C. Antonucci (Eds.), Socialnetwork research: Substantive issues and methodological ciuestions (pp. 121-1 36).Rockland, MA; Swets & Zeitlinger.Vaux, A. (1985). Variations in social support associated with gender, ethnicity andage. Journal of Social Issues, 41, 89-1 10.Walter, G.A. & Marks, S.E. (1981). Experimental learning and change. New York:John Wiley & Sons.Wellman, B. (1985). Domestic work, paid work and net work. In S. Duck & D.Penman (Eds.), Understanding personal relationships (pp. 159-191). London: Sage.Wiseman, J.P. (1986). Friendships: Bonds and binds in a voluntary relationship.Journal of Social and Personal Relationships, 3, 191-211.110Appendix A:LETTERS AND QUESTIONNAIRES111LETTER 1Dean Goldberg’s Data Site Approach LetterDearOne of my senior faculty members in the HRM/Organizational Behavior group -- Prof. CraigPinder -- is currently initiating two new research projects related to [1] intra-companytransfers and [21 men’s and women’s work networks. I am writing to you on his behalf tosee whether I can interest you and your company in these studies and to secure theparticipation of a number of your employees.The first study is directed at the issue of whether intra-company transfer experiences cancontribute to growth, learning, and development of individual employees who are moved.Craig’s work on this topic over the past 17 years has impressed upon him how widely-spread the belief is that moving people around the various locations at which anorganization conducts business is “good for” both the organization and its employees. Thepurpose of the new project is to examine the actual learning process of employees -- howand what employees learn -- so that the degree to which the transfer experiences havecontributed to the employees’ personal and professional development can be assessed. Inview of the enormous financial and human costs associated with moving people, Craigbelieves employers can stand to benefit from an assessment of the value they receive inreturn for the considerable costs of transfers.The second study is actually doctoral dissertation research being conducted by Mr. RichardStackman, one of Prof. Pinder’s Ph.D. students, working under his supervision. Thisdissertation is ground-breaking in that it is one of the first studies to answer some verybasic questions in regards to work networks. Richard is interested in studying thesimilarities and differences of male and female work networks as they exist withinorganizations -- both prior to and following vertical promotions. This dissertation could bethe basis for future studies which examine how people’s networks impact employees’ability to become and remain proficient in their jobs as well as secure future advancementwithin their organizations. One of the ultimate goals for this stream of research is todemonstrate that these differences may be beneficial, and that both sexes may actuallylearn something from the other in how they should go about building and maintaining theirwork networks. Finally, Craig and Richard hope to learn a great deal about the way peopledeal with being promoted, regardless of their gender.To reiterate, the purpose of these two studies is to [1] advance the understanding of therole of transfer experiences in the personal and professional development of Canadianmanagerial, professional, and technical personnel and [2] advance the understanding of thesimilarities and differences of men’s and women’s work networks.Should you elect to have your organization participate, a sample of your employees wouldbe administered questionnaires. The particular questionnaire used would depend uponwhether the employee [11 is being transferred, [21 is being promoted, or [31 has notundergone any career transition in the past year. As has been the case in his past research,Craig is seeking to gain the involvement of employees of several organizations from avariety of industries within Canada.As you know, one of the primary goals of my deanship is to further develop the workingrelationships between the academic and business communities. One way of accomplishing112LETTER 1 (continued)Dean Goldberg’s Data Site Approach Letterthis is for academics to conduct research within the business community that is ofparticular interest to the business community. I think that the research projects Craig isproposing are directed toward this goal.Professor Pinder ensures me that the amount of time and effort required by your own staffin executing the study will be minimal. It will be Mr. Stackman’s job over the comingacademic year to make these projects “go.” Craig wants me to emphasize that theconfidentiality of data provided by individual employees will be ensured. Likewise, noorganizational data [or results] will be provided to anyone outside the respectiveparticipating companies: results will be reported by industry or in aggregate form only.Both Craig and Richard say that they can make participation attractive to thoseorganizations that do participate through what is learned about men’s and women’s worknetworks as well as about the developmental value of the transfer experience of yourparticular personnel. In previous studies, Craig has obtained complete data from over 800employees, working for seven different firms in a variety of industries. Dr. Pinder hasconsiderable experience at presenting the results of both current and previous findings thathave emerged from his 1 7 years of work in this area, and he is prepared to provide suchinformation with any of the participating companies. I am enclosing a copy of an articleCraig wrote for the management periodical Organizational Dynamics that summarizes manyof his past findings.In closing, I hope you will consider engaging your company in our new studies, and that youwill contact Craig Pinder or myself if you have any questions about them. Both Craig andRichard are willing to meet with you or any representatives of your company for the sake ofsecuring your participation in either or both of these projects. Craig Pinder can be reachedat [604] 822-8374; Fax 822-8517. Thank you for considering this request.Sincerely yours,Michael A. GoldbergDean113LETTER 2Dr. Pinder’s Data Site Approach LetterDearOver the past 1 7 years, I have been involved in a series of studies dealing with employeemobility and organization policies concerning the transfer of personnel. I am presentlyundertaking, with one of my Ph.D. students, several new projects related to [1] transfersand [2] men’s and women’s work networks. I am writing to see whether I can interest youand your firm in these studies and to secure the participation of a number of youremployees. As in the previous studies, I hope to gain the involvement of employees ofseveral organizations from a variety of industries throughout Canada.Let me describe the projects. The first study is directed at the issue of whether intracompany transfer experiences can contribute to growth, learning, and development ofindividual employees who are moved. My work on this topic over the years has impressedme with how widely-spread the belief is that moving people around the various locations atwhich an organization conducts business is “good for” both the organization and itsemployees. The purpose of my new project is to examine the actual learning process ofemployees -- how and what employees learn -- so that I can assess the degree to whichtheir transfer experiences have, in fact, contributed to their personal and professionaldevelopment. In view of the enormous financial and human costs associated with movingpeople, I believe employers can stand to benefit from an assessment of the value theyreceive in return for the considerable costs of transfers.The second study is being undertaken by Mr. Richard Stackman, one of my Ph.D. studentsworking under my supervision, as his doctoral dissertation. This dissertation is ground-breaking in that it is one of the first studies to answer some very basic questions in regardsto work networks. He is interested in studying the similarities and differences of male andfemale work networks as they exist within organizations -- both prior to and followingvertical promotions. Moreover, having begun to isolate these differences and similarities,this dissertation could be the basis for future studies which examine how people’s networksimpact employees’ ability to become and remain proficient in their jobs as well as securefuture advancement within their organizations. One of the ultimate goals for this stream ofresearch is to demonstrate that these differences may be beneficial, and that both sexesmay actually learn something from the other in how they should go about building andmaintaining their work networks. We also hope to learn a great deal about the way peopledeal with being promoted, regardless of their gender.We are currently developing new questionnaires to address the issues associated with thetwo studies. (We will be happy to send you draft copies, when they are ready, should youwish to pursue the matter of participating in the research.) Should you elect to have yourorganization participate, the questionnaires would be administered to a sample of youremployees who [1] have been transferred at least once, [2] are being promoted, or [3] havenot undergone any career transition in the past year. We will minimize the amount of timeand effort required of your own staff in executing these studies: Mr. Stackman’s full-timejob over the coming year is to make these projects “go.”I wish to make it clear that our goals in these two projects are purely academic rather thaneconomic, although the findings should have considerable applied value for participatingcompanies. Our purpose is to [11 advance our understanding of the role of transferexperiences in the personal and professional development of Canadian managerial,114LETTER 2 (continued)Dr. Pinder’s Data Site Approach Letterprofessional, and technical personnel and [2] advance our understanding of the similaritiesand differences of men’s and women’s work networks. I also believe that we can makeparticipation attractive to those organizations that do participate through what is learnedabout men’s and women’s work networks as well as about the developmental value of thetransfer experience of your particular personnel.In previous studies I have obtained complete data from over 800 employees, working forseven different firms in a variety of industries. I have considerable experience at presentingthe results of both current and previous findings that have emerged from my 1 7 years ofwork in this area, and I am quite prepared to provide such information with any of theparticipating companies. I have published an article in the management periodicalOrganizational Dynamics (1989) that summarizes many of my findings. I would be pleasedto send you a copy of that paper to illustrate where my work has been in the past.In closing, I hope you will consider engaging your company in our new studies, and that youwill contact me if you have any questions about them. I am willing to meet with you forthe sake of securing the participation of any organizations that might consider. Feel free tocall me at [604] 822-8374; Fax 822-8517.Sincerely yours,Craig C. PinderProfessor115LETTER 3First Cover Letter Included with QuestionnaireSigned by Data Site Company OfficialMr. Richard Stackman, a doctoral student at the University of British Columbia, hasapproached [company’s name] with the intention of conducting a research study using[company’s name] managers and supervisors as participants. After careful review of theresearch proposal, the decision was made to grant Mr. Stackman access to our employees.I would appreciate it if you gave careful consideration to filling out the enclosedquestionnaire at your earliest convenience. Granted, you are busy with your job. However,the company receives numerous requests each year from prospective researchers, and thisrequest is one research study in which [company’s name] is definitely interested inparticipating.A cover letter from Mr. Stackman’s advisor, Dr. Craig Pinder, is enclosed, and I would liketo emphasize two points made in that letter. First, no one in the company will have accessto the responses you provide. You will be mailing the questionnaire directly back to Mr.Stackman. When the data collection is complete, all [your company’s name] will receivefrom Mr. Stackman is aggregate results. No individual results will ever be provided to us.Second, Mr. Stackman’s main goal in pursuing this stream of research is to learn moreabout the similarities and differences between men’s and women’s work networks. Suchinformation should prove beneficial in that both sexes may have something to learn fromthe other in how they ultimately go about building and maintaining their work networks.Should you have any questions, please feel free to contact me at [phone number] or Dr.Pinder at [604] 822-8374.Thank you for your cooperation.Sincerely yours,116LETTER 4Second Cover Letter Included with QuestionnaireSigned by Dr. PinderI am writing to request your help in a research study one of my Ph.D. students, Mr. RichardStackman, is conducting as his dissertation. This study is being undertaken in conjunctionwith your organization which has given Richard and myself permission to contact you in thehope of persuading you to participate in this study.The title of Richard’s thesis is “The Similarities, Differences, and Fluidity of Men’s andWomen’s Work Networks.” This dissertation is ground-breaking in that it is one of the firststudies to answer some very basic questions in regards to work networks. He is interestedin studying the similarities and differences of male and female work networks as they existwithin organizations -- both prior to and following vertical promotions. Moreover, havingbegun to isolate these differences and similarities, this dissertation could be the basis forfuture studies which examine how people’s networks impact employees’ ability to becomeand remain proficient in their jobs as well as secure future advancement within theirorganizations. One of the ultimate goals for this stream of research is to demonstrate thatthese differences may be beneficial, and that both sexes may actually learn something fromthe other in how they should go about building and maintaining their work networks.All that is required to participate in this study is for you to complete the enclosedquestionnaire. It should take you, on average, 30 minutes to complete. When you aredone, please return the completed questionnaire in the self-addressed stamp envelopeprovided. By completing the questionnaire, it is assumed that you have given your consentto participate in the study. Please note that you retain the right to refuse to participate orwithdraw from the study at any time.Your questionnaire will remain confidential. When the completed questionnaire is returnedto us, the only identification on the questionnaire will be a number to ensure anonymity.Your organization will not have access to your individual questionnaire. Only organizationalor aggregate results will ever be reported. No individual data will ever be reported.In closing, I hope you will consider participating in the study and will fill out thequestionnaire according to the instructions provided. If you should have any questions,feel free to call me at [6041 822-8374.Sincerely yours,Craig C. PinderProfessor117LETTER 5Follow-up LetterOver two weeks ago you received a request from me to complete a questionnaireconcerning your work relationships. This questionnaire is part of a project representing thefinal hurdle on the path to my Ph.D. For me to successfully complete my degree, I need anadequate response from the individuals who were sent the questionnaire.It is possible that given your work schedule you have put the questionnaire aside to fill outlater. If so, this letter will serve as a reminder.Moreover, some individuals may be unwilling to fill out the questionnaire because they fearthat the individuals they list could be contacted in the future. This is not the case.guarantee you that none of the co-workers or friends you list will be contacted.Should you need a new questionnaire, please call me at 822-8504, and I would be morethan happy to provide you with a new copy. Also, should you have any questions orconcerns regarding this project, feel free to call my advisor, Dr. Craig Pinder, at 822-8374.Thanks for your help.Sincerely yours,Richard W. StackmanPh.D. Candidate118EXHIBIT 1Non-Transition Sample Questionnaire119SURVEYOFINDIVIDUALWORKNETWORKSThisquestionnaireisdesignedtoexaminetheworknetworksofmanagersandsupervisors.Youremployerhasapprovedthisprojectandquestionnaire.However,Iamcompletelyresponsibleforthecontentsoftheinstrument,aswellasforthedatacollected.Iwillbetheonlyperson--asidefrommyadvisororassistants--whowillhaveaccesstoyourpersonalresponses.Youremployerwillnothaveaccesstoyourpersonalresponses,althoughyouremployerwillreceiveacopyoftheresultsofthesurvey,inaggregatedform.Yourparticipationis,ofcourse,completelyvoluntary.Yourquestionnaireiscomposedoffourparts.Thefirstthreepartsofthequestionnairedealspecificallywithyourworkrelationships.Thefinalsectionasksaboutyouandyourcurrentemploymentstatus.Pleasetrytoanswerofthequestionsasked,andremembermypromiseofcomrleteconfidentiality.Ihopethatyouwilltakethetimetocompletethisquestionnaireandreturnittome,usingthestampedenvelopethatisenclosed.Theinstrument takesbetween30to45minutestocomplete.Youwilllikelyfinditinteresting.Thankyouinadvanceforyourhelp.RichardW.StackmanPh.D.CandidateFacultyofCommerceandBusinessAdministrationTheUniversityofBritishColumbia2053MainMallVancouver,B.C.V6T1Z2Telephone:(604j822-8504Fax:1604]822-8517Advisor:Dr.CraigC.PinderTelephone:[6041822-8374F.’)0IPARTONE:QUESTIONSABOUTYOURWORKRELATIONSHIPS1.Pleaselistthename[ornamesiofyourimmediate3.PleaselistthenameLsiandpositionLs]ofanyoneinyourmanager/supervisor.organizationwhohasbeenusefulorhelpfulinyourlearningtoperformthetasksanddutiesassociated______________________________________withyourcurrentposition.______________________________________NamePosition2.Pleaselist thenameEsiandposition[sJofanyoneinyour_______________________________________________organizationwhohasbeenamentortoyou.Amentorissomeonewhohastakenapersonalinterestinyourcareerandhasguidedorsponsoredyou.Thispersonmayhaveservedasacareerrolemodelandactivelyadvised,guided,and/orpromotedyourcareertraining.4a.PleaseprovidethenumberofindividualsyoudirectlyNamePositionmanage/supervise.b.Approximatelywhatpercentageoftheindividualsyoudirectlysuperviseare...?_____%Men_____%WomenPLEASETURNTOTHENEXTPAGE.2PARTTWO:TASK-RELATEDTIESQUESTIONNAIREBeforebreakingthesealonthispartofthequestionnaire,pleaselistupto15individualswithinyourcompanywhomyouconsidertobeusefulinperformingthetasksrequiredofyourcurrentjob.Task-relatedrelationshipsincludethoseinyourorganizationwhoaidyouand/orarenecessaryforyoutoperformthetasksrequiredinyourjob.Inotherwords,yourelyontheseindividualstohelpyoudoyourjob.IPleasenotethatyoumayinteractwiththeseindividualsonadailybasisorasinfrequentlyasonceaweekormonth.lMoreover,theseindividualsprovideyouwithjob-relatedresourcesleg.,information,expertise,professionaladvice,and/ormaterialresourcesi.F’.)F’)Pleaseexcludeyourmanaperlsl/sui,ervisorlslaswellasallofyoursubordinatesfromthislist.Pleasebreakthesealonthispartofquestionnaireuponcompletingyourlistandprovidetherequestedinformationinrelationtoeachindividualyouhavelisted.PLEASEREADTHEINSTRUCTIONSBEFOREBREAKINGTHESEALSexIMIMale19FemaleHowlongDoesthisWhatIsthetitleofthisIsthisIfthispersonworksinaConsideringtheConsideringonlythepeopleyou(inyears)individualworkindividual’scurrentpositionpersonindifferentfunctionalhierarchyofthehavelistedinthefirstcolumn.haveyouattheintheorganization?thesamedepartmentthanyou,organization,whoelsewouldyouthisknownthisfunctionalpleasenoteinwhIchdoesthisindividualassociatewithinorderIndividual?(1)Samesitedepartmentfunctionaldepartmentindividualhaveaforyoutocompletethetasks(21DIfferentsite,asyou?thisindividualdoespositionthatisassociatedwithyourjob?work.(e.g.,sales,(31Differentsite.lviYeaclaims,personnel)111atleasttwoForexample,ifyouandJaneSmithdifferentcity(NINolevelsabovealsoworkedfrequentlywithyoursindivIduals#5,7and12.thenyouasyou?(21onelevelabovewouldwtiteInthenumbers5.7andyours12inthespaceprovided.(31atthesamelevelasyours(41onelevelbelowyours151atleasttwoleveisbelowyours?- k)Example:JaneSmithF4-1121Sr. AccountantVAccounting25,7,122. 3. 4. 5. 6. 7. 8. 9.,pleaselist upto15individualswithinyourorganizationwhomyouconsidertobe“friendsofyours.Friendsaredefinedasthoseindividualswithwhomyoufrequentlyorofteninteractforpersonalsatisfactionandenjoymentratherthanjustforthefulfilment ofaparticulartaskorgoal.Thoselistedwouldinclude:111peopleyouseesociallyoutsideofwork,and121thosepeopleyouspendtimewithsociallyatworkle.g.,atlunchandcoffeebreaks)butdonotseeoutsideofwork.Therefore,youaretoincludeanyonewhomyouconsidertobeafriend--evenifyoulistedthemasataskrelatedrelationshiponthepreviousquestion.Also,ifyouconsideranysuperiorsand/orsubordinatesasfriends,pleasebesuretolistthem.Pleasebreakthesealuponcompletingyourlistoffriendsandprovidetherequestedinformationinrelationtoeveryindiyidualyouhavelisted.PLEASEREADTHEINSTRUCTIONSBEFOREBREAKINGTHESEALSex(MlMale(FlFemaleHowlongIsthisanDoesthisWhatisthetitleofIsthisIfthispersonConsideringtheIsthisanHowoftenConsideringonlytinyearslindividualindividualthisindividual’spersonInworksInahierarchyoftheindividualareyouInthepeopleyouhaveyouyouseeworkatthecurrentpositionInthesamedifferentfunctionalorganization,youcontactwithhavelistedintheknownthissociallytheorganization?functionaldepartmentthandoesthismanegelthisperson-.firstcolumn,withindividual?outsideof111Samesitedept.asyou,pleasenoteinindividualhaveasupervise?eitherinwhomelsewouldwork?121Differentyou?whichfunctionalpositionthatispersonorbyyouandthissite,departmentthisIYIYesphone?individualsocializeIYIYessamecity(VIYesindividualdoes(11atleasttwo(NINotogether?(NINo131DifferentININowork.(e.g.,sales,levelsaboveIllDailysite.differentclaims,personnellyours121ThreetimesForexample,Ifyou121onelevelaboveperweekandJaneSmithalsocityyours(31Oncepersocializedfrequently131atthesameweekwithIndIviduals#5,asyou?levelasyours(41TwIcea7and12,thenyou14)onelevelbelowmonthwouldwriteIntheyours(51Lessthannumbers5,7and12151atleasttwoonceperinthespacelevelsbelowmonthprovided.yours?N)Example:JaneSmithF4-112V1Sr.AccountantVAccounting2N25,7,122. 3. 4. 5. 6. 7. 8. 9. thepreviouspositionsyouhaveheldwithyouremployer?LExample:3-1/2yearslcurrentemployer.Besuretolist theyearsinwhichyouheldthesepositions.[Example:Junior______yearsAccountant,1989-19921PositionDates2.Whatisthetitleofyourcurrentposition?3.Howmanyyearshaveyoubeenatyourcurrentposition?______________________________________________[Example:3-1/2yearsl______years4.Howmanytotalyearshaveyouworkedfull-timesince______________________________________________turning18?______years5.Whichofthefollowingjobcategoriesbestdescribesyou?______________________________________________[Pleasecheckoneonly.JExecutive—Seniormanager—Middlemanager—First linemanager/supervisor7.Howmanytimeshaveyoubeenpreviouslypromotedby—Other[pleasespecifylyourcurrentemployer?____________________________timesPLEASETURNTOTHENEXTPAGE.58.Pleaseindicatewithacheckmarkthetypeofdepartmentinwhichyoucurrentlywork.[Pleasecheckoneonly.]—Accounting—AccountManagement—BusinessDevelopmentClaims—Communication/PublicRelationsCustomerService—DataProcessing—EngineeringFinance—GeneralAdministrationInformationSystems—InternationalSales/Marketing—LaborRelations—Law/Legal—Leasing—OccupationalHealth/SafetyPersonnellHumanResources—ProductManagerProduction/Manufacturing—Purchasing—RealEstate/PhysicalPlant—Sales/Marketing—StrategicPlanning/Development—TradeRelations—Transportation—Other[pleasespecify]9.Whichofthefollowingjobcategoriesbestdescribesyou?[Pleasecheckoneonly.]—ManagementProfessionalTechnical—Administrative—Professional/Management—Technical/Management—Administrative/Management—Other[pleasespecify]1 2.Whichethnicgroupbestdescribesyourself?[Pleasecheckoneonlyl:—AboriginalAsianBlackCaucasianHispanic—Other[pleasespecify]-10.Whenisyourbirthday?[Month/Date/Year]11.Whatisyoursex?MaleFemalePLEASETURNTOTHENEXTPAGE.613.Whatisyourrelationshipstatus?16.Whatisyourpresentannualgrosssalary[includingcashbonuses]?[Pleasechecktheappropriaterange.]—Marriedorlivingwithalong-termpartner—Notmarriedorlivingwithalong-termpartner—Under$20,000—$20,000to29,999$30,000to39,999—$40,000to49,99914a.Howmanychildrendoyouhave?$50,000to74,999—$75,000to99,999—$100,000to124,999—$125,000to149,999_$150,000to174,999b.Whatistherangeofyourchildren’sages?—$175,000to199,999$200,000andoverFrom_______yearsto_______years;orIhaveonlyonechild,whois_______yearsold.15.Pleaseindicateyourhighestlevelofformaleducation.[Pleasecheckoneonly.]—Somehighschool—Highschoolgraduation—Somecollege/university—Collegediploma—Universitydegree—Somegraduatestudy—Advanceddegree[pleasespecify:e.g.,Ph.D.,M.D.,LL.B.,M.B.A.,M.S.,etc.]Other[pleasespecifylPLEASETURNTOTHENEXTPAGE.Pleaseprovidebelowyouremployer’sname,thePARTICIPANTNUMBERlocationsite,andthenameofthefunctionaldepartmentforyourcurrentposition.Alsonotethataparticipantnumberlandnotyourname]isbeingusedonthisquestionnaire.Thisparticipantnumberwillensuretheconfidentialityandanonymityofyourresponses.Employer’sName______________________________________GeographicLocationofCurrentJob________________________________________Functional Department________________________________I’.)CDThatcompletesthequestionnaire.Ifyouhaveanythingyouwishtoaddinconnectionwithyourrelationshipsatyourworkplace,pleaseusetheblankpageattachedtothebackofthisquestionnaire.Ifyouwishtoreceiveacopyoftheaggregateresults,pleasecheckthebox:DThankyouforyourtime!RichardW.StackmanPh.D.Candidate,UniversityofBritishColumbiaOE[EXHIBIT 2First Promotion Sample Questionnaire131SURVEYOFINDIVIDUALWORKNETWORKSThisquestionnaireisdesignedtoexaminetheworknetworksofmanagersandsupervisors.Youremployerhasapprovedthisprojectandquestionnaire.HoWever,I amcompletelyresponsibleforthecontentsoftheinstrument,aswellasforthedatacollected.Iwillbetheonlyperson--asidefrommyadvisororassistants--whowillhaveaccesstoyourpersonalresponses.YouremployerfflnIhaveaccesstoyourpersonalresponses,althoughyouremployerwillreceiveacopyoftheresultsofthesurvey,inaggregatedform.Yourparticipationis,ofcourse,completelyvoluntary.Yourquestionnaireiscomposedoffourparts.ThefirstthreeRichardW.StackmaripartsofthequestionnairedealspecificallywithyourworkPh.D.Candidaterelationships.ThefinalsectionasksquestionsaboutyouandFacultyofCommerceandBusinessAdministrationyourcurrentemploymentstatus.PleasetrytoanswerllofTheUniversityofBritishColumbiathequestionsasked,andremembermypromiseofcomplete2053MainMallconfidentiality.Vancouver,B.C.V6T1Z2Telephone:[6041822-8504IhopethatyouwilltakethetimetocompletethisFax:[6041822-8517questionnaireandreturnittome,usingthestampedenvelopethatisenclosed.Theinstrument takesbetween30to50minutestocomplete.Youwilllikelyfinditinteresting.Advisor:Dr.CraigC.PinderThankyouinadvanceforyourhelp.Telephone:[6041822-83741PARTONE:QUESTIONSABOUTYOURWORKANDYOURWORKRELATIONSHIPSIfyourecentlyreceived[inthepastmonthorso]apromotion,beginbyansweringQuestion#1.Ifyouhavenotrecentlyreceivedapromotion,pleasebeginwithQuestion#6onPage2.1.Whatisthetitleofyournewposition?4.Howweretheseindividuals[listedinquestion#31helpfultoyouingettingthisnewposition?Inotherwords,whatdidtheydotohelpyougetthisnewposition?Person#1.______________________________________2.Willyoubereceivingaraise?YesNoPerson#2._________________________________________CA)CA)3.Whatindividualsinyourorganizationwerehelpfultoyou__________________________________________ingettingthepromotiontoyournewposition?Pleaseprovidetheindividuals’namesandDositions1,1Person#3.___________________________________________theorganization.Ifyouneedmorespace,pleaseusethebackofthissheetofpaper.NamePositionPerson#4.________________________________________2.__________________________________________Person#5._______________________________________3.____________________________________________________________________________________4._________________________________________Person#6._________________________________________5.____________________________________________6.______________________________________________________________________________PLEASETURNTOTHENEXTPAGE.25.Inyourownwords,pleaseexplainwhyyouconsider9.Pleaselist thename(sI andoositionLslofanyoneinyouryournewpositionapromotion,organizationwhohasbeenamentortoyou.Amentorissomeonewhohastakenapersonalinterest_______________________________________________inyourcareerandhasguidedorsponsoredyou.Thispersonmayhaveservedasacareerrolemodelandactivelyadvised,guided,and/orpromotedyourcareertraining.NamePosition6.Pleaselist thenameLornames]ofyourimmediatesupervisors._______________________________________________10.Pleaselist thename[sIandpositionisiofanyoneinyour______________________________________organizationwhohasbeenusefulorhelpfulinyourlearningtoperformthetasksanddutiesassociated______________________________________withyourcurrentposition.NamePosition7a.Pleaseprovidethenumberofindividualsyoudirectly_______________________________________________supervise.b.Approximatelywhatpercentageoftheindividualsyoudirectlysuperviseare...%Men%WomenPLEASETURNTOTHENEXTPAGE.3PARTTWO:TASK-RELATEDTIESQUESTIONNAIREBeforebreakingthesealonthispartofthequestionnaire,pleaselist upto15individualswithinyourcompanywhomyouconsidertobeusefulinperformingthetasksrequiredofyourjobPRIORTOYOURRECENTPROMOTION.Task-relatedrelationshipsincludethoseinyourorganizationwhoaidyouand/orarenecessaryforyoutoperformthetasksrequiredinyourjob.Inotherwords,yourelyontheseindividualstohelpyoudoyourjob.[Pleasenotethatyoumayinteractwiththeseindividualsonadailybasisorasinfrequentlyasonceaweekormonth.)Moreover,theseindividualsprovideyouwithjob-relatedresources[e.g.,information,expertise,professionaladvice,and/ormaterialC)resources].01Pleaseexcludeyourmanager[s)/supervisorlslaswellasallofyoursubordinatesfromthislist.Pleasebreakthesealonthispartofquestionnaireuponcompletingyourlistandprovidetherequestedinformationinrelationtoeachindividualyouhavelisted.PLEASEREADTHEINSTRUCTIONSBEFOREBREAKINGTHESEALSexIMIMale(FlFemaleHowlongDoesthisWhatIsthetitleofthisIsthisifthispersonworksinaConsideringtheConsideringonlythepeopleyou[inyearslindividualworkindividual’scurrentpositionpersonindifferentfunctionalhierarchyofthehavelistedinthefirstcolumn,haveyouattheintheorganization?thesamedepartmentthanyou,organization,whoelsewouldyouthisknownthisfunctIonalpleasenoteinwhichdoesthisindividualassociatewithinorderindivIdual?(11Samesitedepartmentfunctionaldepartmentindividualhaveaforyoutocompletethetasks[21DIfferentsite,asyou?thisindividualdoespositionthatisassociatedwithyawjob?CItYwork.(e.g.,sales,131DIfferentsilo.(YlYesclaims,personnel]111atleasttwoForexample,ifyouandJan.SnithntY(NJNolevelsabovealsoworkedfrequentlywithyoursIndIvIduals#6,7and12,thenyouasyou?(21onelevelabovewouldwriteInthenumbers5.7andyours12Inthespaceprovided.(31atthesamelevelasyows(4]onelevelbelowyours(51atleasttwolevelsbelowyours?-s CA)0)Example:JaneSmithF4-1121Sr.AccountantVAccounting25,7,122. 3. 4. 5. 6. 7. B. 9.,pleaselist upto15individualswithinyourorganizationwhomyouconsidertobe“friendsofyours.Friendsaredefinedasthoseindividualswithwhomyoufrequentlyorofteninteractforpersonalsatisfactionandenjoymentratherthanjustforthefulfilment ofaparticulartaskorgoal.Thoselistedwouldinclude:(1]peopleyouseesociallyoutsideofwork,and[2]thosepeopleyouspendtimewithsociallyatwork(e.g.,atlunchandcoffeebreaks]butdonotseeoutsideofwork.Therefore,youaretoincludeanyonewhomyouconsidertoC)beafriend--evenifyoulistedthemasatask-relatedrelationshiponthepreviousquestion.Also,ifyouconsideranysuperiorsandlorsubordinatesasfriends,pleasebesuretolistthem.Pleasebreakthesealuponcompletingyourlistoffriendsandprovidetherequestedinformationinrelationtoeveryindividualyouhavelisted.PLEASEREADTHEINSTRUCTIONSBEFOREBREAKINGTHESEALSex(MlMale(FlFemaleHowlongIsthisanDoesthisWhatisthetitleofIsthisIfthispersonConsideringtheIsthisanHowoftenConsideringonly(inyearslindividualIndividualthisindividual’spersoninworksinahierarchyoftheindividualareyouinthepeopleyouhaveyouyouseeworkatthecurrentpositioninthesamedifferentfunctionalorganization,youcontactwithhavelistedintheknownthissociallytheorganization?functionaldepartmentthandoesthismanage)thisperson-.firstcolumn,withIndividual?outsideof(iiSamesitedept.asyou,pleasenoteinindividualhaveasupervise?eitherinwhomelsewouldwork?121Differentyou?whichfunctionalpositionthatjpersonorbyyouandthisSite,departmentthis(YlYesphone?individualsocialize(YlYestamecity(•Yesindividualdoes(11atleasttwo[NINotogether?ININo(31DIfferent(NINowork.[e.g.,sales,levelsabove(11DaIlysite.differentclaims,personnellyours[21ThreetimesForexample,Ifyoucityt21onelevelaboveperweekandJaneSmithalsoyours(31Oncepersocializedfrequently(31atthesameweekwithIndivIduals#5.asyou?levelasyours[41Twicea7and12.thenyou(41onelevelbelowmonthwouldwriteIntheyours(SILessthannumbers5.7and12(51atleasttwoonceperinthespacelevelsbelowmonthprovided.yours?CA)o:iExample:JaneSmithF4-1)2Y1Sr.AccountantYAccounting2N25.7,122. 3. 4. 5. 6. 7. 8. 9. theyearsinwhichyouheldthesepositions.(Example:JuniorAccountant,1989-______years19921PositionDates2.Howmanyyearshaveyouworkedatthisgeographicsite?_______years3.Whatwasthetitleofyourpositionpriortothepromotion?__________________________________________4.Howmanyyearshadyouworkedatthepositionyouheldpriortoyourpromotion?(Example:3-112years]________________________________________________years_________________________________________________7.Howmanytimeshaveyoubeenpromotedbyyourcurrent5.Howmanyyearshaveyouworkedfull-timesinceturning18?employer?_______years_______timesPLEASETURNTOTHENEXTPAGE.68.Consideringyournewposition,pleaseindicatewithacheckmarkthetypeofdepartmentinwhichyouwillwork.IPleasecheckoneonly.1—Accounting—AccountManagement—BusinessDevelopmentClaimsCommunication/PublicRelations—CustomerService—DataProcessing—EngineeringFinance—GeneralAdministration—InformationSystems—InternationalSales/Marketing—LaborRelationsLawlLegal—LeasingOccupationalHealth/Safety—Personnel/HumanResources—ProductManager—Production/Manufacturing—Purchasing—RealEstate/PhysicalPlant—Sales/Marketing—StrategicPlanninglDevelopment—TradeRelations—Transportation—Other(pleasespecify]9.Whichofthefollowingjobcategoriesbestdescribesyou?[Pleasecheckoneonly.]Executive—SeniormanagerMiddlemanager—First linemanager/supervisor—Other(pleasespecify]10.Whichofthefollowingjobcategoriesbestdescribesyou?(Pleasecheckoneonly.]ManagementProfessionalTechnical—Administrative—Professional/Management—Technical/Management—Administrative/Management—Other[pleasespecify]11.Isyourcurrentpositioninthesametypeofdepartmentasyournewposition[asindicatedinQuestion#8]?Ifno,pleaseprovidethetypeofdepartment[e.g.,accounting].0YesNoPLEASETURNTOTHENEXTPAGE.819.Whatwillbeyourannualgrosssalary(includingcashbonusesiforyournewposition?(Pleasechecktheappropriaterange.J—Under$20,000—$20,000to29,999—$30,000to39,999—$40,000to49,999—$50,000to74,999—$75,000to99,999—$100,000to124,999$125,000to149,999—$150,000to174,999—$175,000to199,999—$200,000andover20.Onwhatdatedidyourpromotiontakeeffect?-21.Didyourpromotionrequireyouto...(Pleasecheckone.J—remaininthesamebuilding,onthesamefloorasyourpreviousjob—movetoadifferentfloorwhileremaininginthesamebuildingasyourpreviousjobIfyoucheckedthatyoumovedtoadifferentfloor,howmanyfloorsdidyoumove?___________floorsmovetoacompletelydifferentbuildingIfyoucheckedthatyoumovedtoadifferentbuildingsite,whatisthedistance[inkilometresjbetweenyournewandoldworkrocations?kilometresPLEASETURNTOTHENEXTPAGE.712.Whenisyourbirthday?17.Pleaseindicateyourhighestlevelofformaleducation.jMonth/DatelYearl13.Whatisyoursex?______Male______Female—Somehighschool—HighschoolgraduationSomecollege/universityCollegediploma—Universitydegree—SomegraduatestudyAdvanceddegree(pleasespecify:e.g.,Ph.D.,M.D.,LL.B.,M.B.A.,M.S.,etcj16.Whatisyourrelationshipstatus?—Marriedorlivingwithalong-termpartner—Notmarriedorlivingwithalong-termpartner—Under$20,000$20,000to29,999—$30,000to39,999—$40,000to49,999—$50,000to74,999—$75,000to99,999—$100,000to124.999—$125,000to149,999—$150,000to174,999—$175,000to199,999—$200,000andover- F’.)—Otherlpleasespecifyl14a.Howmanychildrendoyouhave?b.Whatistherangeofyourchildren’sages?From_______yearsto_______years;orIhaveonlyonechild,whois______yearsold15.Whichethnicgroupbestdescribesyourself?(Pleasecheckappropriateboxl:—AboriginalAsianBlackCaucasian—Hispanic—Other(pleasespecifyl18.Whatisyourannualgrosssalary(includingcashbonusesjforyourpositionprevioustoyournewposition?(Pleasechecktheappropriaterange.lPLEASETURNTOTHENEXTPAGE.Pleaseprovidebelowyouremployer’sname,thePARTICIPANTNUMBERlocationsite,andthenameofthefunctionaldepartmentforyourcurrentposition.Alsonotethataparticipantnumber[andnotyourname]isbeingusedonthisquestionnaire.Thisparticipantnumberwillensuretheconfidentialityandanonymityofyourresponsesasthispagewillberemovedfromthequestionnaireimmediatelyaftertheparticipantnumberisplacedonthefrontcover.Employer’sName_____________________________________—GeographicLocationofCurrentJob_____________________________________FunctionalDepartment________________________________Thatcompletesthequestionnaire.Ifyouhaveanythingyouwishtoaddinconnectionwithyourrelationshipsatyourworkplace,pleaseusetheblankpageattachedtothebackofthisquestionnaire.Ifyouwishtoreceiveacopyoftheaggregateresults,pleasecheckthebox:EJThankyouforyourtime.EXHIBIT 3Follow-up Promotion Sample Questionnaire144SURVEYOFINDIVIDUALWORKNETWORKSYoumayrecallseveralmonthsagofillingoutaquestionnaireexaminingtheworknetworksofmanagersandsupervisors.First,Ithankyoufortakingthetimetofill outthatquestionnaire.Second,inthatfirstquestionnaire,ImentionedthatIwouldbesendingyouafollow-upquestionnaireneartheendoftheyear.Thisquestionnaireisashorterversionoftheoriginal,andgiventhelowresponseratetothefirstquestionnaire,Iwouldgreatlyappreciateitifyouwouldtakethetimetofilloutthissecondandfinalquestionnaire.Thesolepurposeofthisquestionnaireistostudytheimpactajobtransition[e.g., apromotionihasonanindividual’sworknetworks.Again,noindivIdualresultsfromthisquestionnairewilleverRichardW.Stackmanbereportedtoanyone.Yourparticipationis,ofcourse,Ph.D.Candidatecompletelyvoluntary.PleasetrytoansweriLoftheFacultyofCommerceandBusinessAdministrationquestionsasked,andremembermypromiseofcompleteTheUniversityofBritishColumbiaconfidentiality.2053MainMallVancouver,B.C.V6T1Z2IhopethatyouwilltakethetimetocompletethisTelephone:[6041822-8369questionnaireandreturnittome,usingthestampedFax:[6041822-8517envelopethatisenclosed.Theinstrumenttakesbetween15to25minutestocomplete.Advisor:Dr.CraigC.PinderThankyouInadvanceforyourhelp.Telephone:[6041822-8374PARTONE:TASK-RELATEDTIESQUESTIONNAIREBeforebreakingthesealonthispartofthequestionnaire,pleaselistupto16indivIdualswithinyourcompanywhomyouconsidertobeinstrumentalincompletingthetasksassociatedwithyourcurrentiob.Task-relatedrelationshipsincludethoseinyourorganizationwhoaidyouand/orarenecessaryforyoutoperformthetasksassociatedwithyourjob.Inotherwords,yourelyontheseindividualstohelpyoudoyourjoband/ortheseindividualsrelyonyoutohelpthemdotheirjobs.Moreover,theseindividualsprovideyouwithjob-relatedresources[e.g.,information,expertise,professionaladvice,and/ormaterialresourcesi.0)Pleaseexcludeyourmanager[sl/supervisorislaswellasallofyoursubordinatesfromthislist.Pleasebreakthesealonthispartofquestionnaireuponcompletingyourlistandprovidetherequestedinformationinrelationtoeachindividualyouhavelisted.PLEASEREADTHEINSTRUCTIONSBEFOREBREAKINGThESEALSex(MlMale(F)FemaleHowlongDoesthisWhatIsthetitleofthisIsthisIfthispersonworksinaConsideringtheConsideringonlythepeopleyou(inyears)individualworkindividual’sCurrentpositionpersonIndifferentfunctionalhierarchyofthehavelistedinthefirstcolumn.haveyouattheIntheorganization?thesamedepartmentthanyou,organization,whoelsewouldyouthisknownthisfunctionalpleasenoteinwhichdoesthisindividualassociatewithinorderindIvidual?(11Samesitedepartmentfunctionaldepartmentindividualhaveaforyoutocompletethetasks(21Differentcite,asyou?thisindividualdoespositionthatIsassociatedwithyourjob?S5fliCitywork.(e.g.,sales,131Differentsite.(YlYesclaims,personnel](11atleasttwoForexample,ifyouandJaneSmithr6fferentCityIN)NolevelsabovealsoworkedfrequentlywithyoursIndIviduals#5.7and12.thenyouasyou?(21onelevelabovewouldwritebthenumbers5,7andyours12inthespaceprovided.(31atthesamelevelasyours(41onelevelbelowyours151atleasttwolevelsbelowyours?Example:JaneSmithF4-1/21Sr.AccountantVAccounting25,7,122. 3. 4. 5. 6. 7. 8. 9.,pleaselist upto15individualswithinyouroruanizatlonwhomyouconsidertobe‘friends”ofyours.Friendsaredefinedasthoseindividuals.withwhomyoufrequentlyorofteninteractforpersonalsatisfactionandenjoymentratherthanjustforthefulfilmentofaparticulartaskorgoal.Thoselistedwouldinclude:[11peopleyouseesociallyoutsideofwork,and121thosepeopleyouspendtimewithsociallyatworkle.g.,atlunchandcoffeebreaksjbutdonotseeoutsideofwork.Therefore,youaretoincludeanyonewhoyouconsidertobeafriend--evenifyoulistedthemasaninstrumentalrelationshioonthepreviousquestion.Also,ifyouconsideranysuperiorsand/orsubordinatesasfriends,pleasebesuretolistthemonthisquestionnaire.Pleasebreakthesealuponcompletingyourlistoffriendsandprovidetherequestedinformationinrelationtoeveryindividualyouhavelisted.PLEASEREADTHEINSTRUCTIONSBEFOREBREAKINGTHESEALSex(MiMale(FlFemaleHowlongIsthisanDoesthisWhatisthetitleofIsthisIfthispersonConsideringtheIsthisanHowoftenConsideringonly(inyearslIndividualindividualthisindividual’spersoninworksinahierarchyoftheindividualareyouInthepeopleyouhaveyouyouseeworkatthecurrentpositioninthesamedifferentfunctionalorganization,youcontactwithhavelistedIntheknownthissociallytheorganization?functionaldepartmentthandoesthismanage!thisperson--firstcolumn,withindividual?outsideof111Samesitedept.asyou,pleasenoteinIndividualhaveasupervise?eitherInwhomelsewouldwork?(21Differentyou?whichfunctionalpositionthatIspersonorbyyouandthissite,departmentthisLVIYesphone?individualsocialize(yjy5‘amecityIYIYesindividualdoes(11atleasttwo(NINotogether?(NINo131DifferentININowork.(e.g.,sales,levelsabove111Dailysite.differentclaims,personnellyours(21ThreetimesForexample,ifyou.121onelevelaboveperweekandJaneSmithalsocityyours131Oncapersocializedfrequently(31atthesameweakwithindIviduals#5,asyou?levelasyours(41Twicea7and12,thenyou(41onelevelbelowmonthwouldwriteintheyours(51Lessthannumbers5,7and12(51atleasttwoonceperInthespacelevelsbelowmonthprovided.yours?- (0Example:JaneSmithF4.112Y1Sr.AccountantVAccounting2N25.7,122. 3. 4. 5. 6. 7. 8. 9.’sages?________________________________________From_______yearsto_______years;orIhaveonlyonechild,whois_______yearsold2a.Pleaseprovidethenumberofindividualsyoudirectlysupervise.6.Pleaseindicateyourhighestlevelofformaleducation.01 CS—Somehighschoolb.Approximatelywhatpercentageoftheindividualsyou—HighschoolgraduationSdirectlysuperviseare...?—Somecollege/university—Collegediploma_____%Men_____%Women—Universitydegree—Somegraduatestudy—Advanceddegree(pleasespecify:e.g.,Ph.D.,M.D.,LL.B.,M.B.A.,M.S.,etc.13.Howmanyyearshaveyouworkedforyourpresentemployer?(Example:3-1/2years)________________________years—OtherIpleasespecifyl4.Whatisyoursex?MaleFemalePLEASETURNTOTHENEXTPAGE.47.Whatisyourcurrent,annualgrosssalarylincludingcashbonusesi?IPleasechecktheappropriaterange.J—Under$20,000—$20,000to29,999—$30,000to39,999—$40,000to49,999—$50,000to74,999—$75,000to99,999—$100,000to124,999—$125,000to149,999—$150,000to174,999—$175,000to199,999—$200,000andover8.Onwhatdaydidyourpromotiontakeeffect?- C,’9.Didyourpromotionrequireyouto...IPleasecheckone.J—remainInthesamebuilding,onthesamefloorasyourpreviousjob—movetoadifferentfloorwhiteremaininginthesamebuildingasyourpreviousjobIfyoucheckedthatyoumovedtoadifferentfloor,howmanyfloorsdidyoumove?floors—movetoacompletelydifferentbuildingIfyoucheckedthatyoumovedtoadifferentbuildingsite,whatisthedistanceunkilometresibetweenyournewandoldworklocations?_______kilometresPLEASETURNTOTHENEXTPAGE.5Pleaseprovidebelowyouremployer’sname,thelocationPARTICIPANTNUMBERsite,andthenameofthefunctionaldepartment foryourcurrentposition.Alsonotethataparticipantnumber[andnotyournamelisbeingusedonthisquestionnaire.Thisparticipantnumberwillensuretheconfidentialityandanonymityofyourresponsesasthispagewillberemovedfromthequestionnaireimmediatelyaftertheparticipantnumberisplacedonthefrontcover.Employer’sName_____________________________________GeographicLocationofCurrentJob_______________________________________FunctionalDepartment________________________________Thatcompletesthequestionnaire.Ifyouhaveanythingyouwishtoaddinconnectionwithyourrelationshipsatyourworkplace,pleaseusetheblankpageattachedtothebackofthisquestionnaire.Thankyouforyourtime.EXHIBIT 4Reliability Check Questionnaire153SURVEYOFINDMDUALWORKNETWORKSYoumayrecalloveramonthagofillingoutaquestionnaireexaminingtheworknetworksofmanagersandsupervisors.Ithankyoufortakingthetimetofill outthatquestionnaire.Inresearchofthissort,itisessentialtoknowwhetherornotthequestionnaireyieldsconsistentresults.Consequently,Iamaskingforyourhelponefinaltime.Iwouldgreatlyappreciateitifyouwouldpleasetakethetimetofilloutthisquestionnaire--whichisashorterversionoftheoriginal.Thesolepurposeofthisquestionnaireistocheckwhetherthe“original”surveyprovidesconsistentdata.NoindividualoraguregateresultsfromthisquestionnaIrewilleverbereportedtoanyone.Again,yourparticipationis,ofcourse,completelyvoluntary.Pleasetrytoansweriiofthequestionsasked,andremembermypromiseofcompleteconfidentiality.Ihopethatyouwilltakethetimetocompletethisquestionnaireandreturnittome,usingthestampedenvelopethatisenclosed.Theinstrumenttakesbetween15to25minutestocomplete.RichardW.StackmanPh.D.CandidateFacultyofCommerceandBusinessAdministrationTheUniversityofBritishColumbia2053MainMallVancouver,B.C.V6T1Z2Telephone:[604)822-8504Fax:[6041822-8517Advisor:Dr.CraigC.PinderTelephone:[604) 822-8374(3IThankyouinadvanceforyourhelp.PARTONE:TASK-RELATEDTIESQUESTIONNAIREBeforebreakingthesealonthispartofthequestionnaire,pleaselist upto15individualswithinyourcompanywhomyouconsidertobeusefulinperformingthetasksrequiredofyourcurrentjob.Task-relatedrelationshipsincludethoseinyourorganizationwhoaidyouand/orarenecessaryforyoutoperformthetasksrequiredinyourjob.Inotherwords,yourelyontheseindividualstohelpyoudoyourjob.[Pleasenotethatyoumayinteractwiththeseindividualsonadailybasisorasinfrequentlyasonceaweekormonth.]Moreover,theseindividualsprovideyouwithjob-relatedresources[e.g.,information,expertise,professionaladvice,and/ormaterialresources].Pleaseexcludeyourmanaoerlsl/supervisorlslaswellasallof(51yoursubordinatesfromthislist.Pleasebreakthesealonthispartofquestionnaireuponcompletingyourlistandprovidetherequestedinformationinrelationtoeachindividualyouhavelisted.- 01 C)Example:JaneSmithF4-1/21Sr.AccountantVAccounting25.7.122. 3. 4. 5. 6. 7. 8. 9.’scurrentpositionpersonindifferentfunctionalhierarchyofthehavelistedinthefirstcolumn.haveyouattheintheorganization?thesamedepartmentthanyou,organization,whoelsewouldyouthisknownthisfunctionalpleasenoteinwhichdoesthIsindividualassociatewithinorderindividual?(11Samesitedepartmentfunctionaldepartmentindividualhaveaforyoutocompletethetasks(21Differentsite.ssyou?thisindividualdoespositionthatisassociatedwithyourjob’55’flCItYwork.(e.g.,sales.(31Differentsite.(VIYesclaims,personnell(11atleasttwoForexample,ifyouandJaneSmithdifferentCityININolevelssbovealsoworkedfrequentlywithyoursIndividuals#5,7and12.thenyouasyou?(2joneleveiabovewouidwñtelnthenumbats5,7endyours12inthespaceprovided.131atthesamelevelasyours(41onelevelbelowyours(51atleasttwolevelsbelowyours?2PARTTWO:FRIENDSHIPTIESQUESTIONNAIREBeforebreakingthesealonthispartof thequestionnaire,pleaselist upto15individualswithinyourorganizationwhomyouconsidertobe*friendsofyours.Friendsaredefinedasthoseindividualswithwhomyoufrequentlyorofteninteractforpersonalsatisfactionandenjoymentratherthanjustforthefulfilment ofaparticulartaskorgoal.Thoselistedwouldinclude:111peopleyouseesociallyoutsideofwork,and[21thosepeopleyouspendtimewithsociallyatwork[e.g.,atlunchandcoffeebreaksjbutdonotseeoutsideofwork.Therefore,youaretoincludeanyonewhomyouconsidertobeafriend--evenifyoulistedthemasatask-relatedrelationshiponthepreviousquestion.Also,ifyouconsideranysuoeriorsand/orsubordinatesasfriends,Pleasebesuretolistthem.Pleasebreakthesealuponcompletingyourlist offriendsandprovidetherequestedinformationinrelationtoeveryindividualyouhavelisted.PLEASEREADTHEINSTRUCTIONSBEFOREBREAKINGTHESEALSex(MlMaleIFIFemaleHowlongIsthisenDoesthisWhatisthetitleofIsthisIfthispersonConsideringtheIsthisanHowoftenConsideringonly(inyearsiindividualindivIdualthIsindivIdual’spersonInworksinahierarchyoftheindividualareyouinthepeopleyouhaveyouyouseeworkatthecurrentpositioninthesamedifferentfunctionalorganization,youcontactwithhavelistedIntheknownthissociallytheorganization?functionaldepartmentthandoesthismanage!thisperson--firstcolumn,withindividual?outsideof(11Samesitedept.asyou,pleasenoteinindividualhaveasupervise?eitherInwhomelsewouldwork?(21Differentyou?whichfunctionalpositionthatispersonorbyyouthissite,departmentthis(VIYesphone?individualsocialize(VIYeasamecity(flYesindividualdoes(11atleasttwo(NINotogether?(NINo(31Different(NINowork.Ie.g..sales,levelsabove(11Dailyit’eclaims,porsonnellyours(21ThreetimesForexample,ifyeuite121onelevelaboveperweekandJaneSmithalsocyyours(31Oncepersocializedfrequently131atthesameweekwithindividuals#6.asyoulevelasyours(41TwIcea7and12.thenyou(41onelevelbelowmonthwouldwriteintheyours(51Lessthannumbers5,1and12151atleasttwoonceperintheapacelevelsbelowmonthprovided.yours?01 oExample:JaneSmithF4-112V1Sr.AccountantYAccounting2N25,7,122. 3. 4. 5. 6. 7. 8. 9. thenamelsiandposition[slofanyoneinyour3.Pleaselist thename[ornamesiofyourimmediateorganizationwhohasbeenamentortoyou.Amanager/supervisor.mentorissomeonewhohastakenapersonalinterestinyourcareerandhasguidedorsponsoredyou.This__________________________________________personmayhaveservedasacareerrolemodelandactivelyadvised,guided, andlorpromotedyour_____________________________________________careertraining.NamePosition4.Whatisyoursex?CD_______MaleFemale2.Pleaselist thenamelsiandposition[sJofanyoneinyour5.Howmanyyearshaveyouworkedforyourpresentorganizationwhohasbeenusefulorhelpfulinyouremployer?[Example:3-1/2years]learningtoperformthetasksanddutiesassociatedwithyourcurrentposition._______yearsNamePosition_______________________________________________6a.Howmanychildrendoyouhave?b.Whatistherangeofyourchildren’sages?From_______yearsto_______years;orIhaveonlyonechild,whois______yearsold.PLEASETURNTOTHENEXTPAGE.47.Pleaseindicateyouhighestlevelofformaleducation.[Pleasecheckoneonly.]—Somehighschool—Highschoolgraduation—Somecollege/university—CollegediplomaUniversitydegreeSomegraduatestudy—Advanceddegree[Pleasespecify:e.g.,Ph.D.,M.Sc.,M.B.A.]Pleaseprovidebelowyouremployer’sname,thelocationsite,andthenameofthefunctionaldepartmentforyourcurrentposition.AlsonotethataparticipantnumberLandnotyourname]isbeingusedonthisquestionnaire.Thisparticipantnumberwillensuretheconfidentialityandanonymityofyourresponses.Employer’sNameOther[Pleasespecify]GeographicLocationofCurrentJob0, 0FunctionalDepartment________________Thatcompletesthequestionnaire.Thankyouforyourtime!PARTICIPANTNUMBERAppendix B:NON-TRANSITION DATA:Variable Means, Standard Deviations, and CorrelationsPROMOTION DATA:Variable Means and Standard Deviations161TABLE B.1Non-Transition Respondent Sample Variable Means and Standard Deviations:Instrumental Networks (n = 242)StandardVariables Means DeviationsNetwork Size and Gender Mix:Number of Individuals Listed 10.63 4.14Number of Males 6.61 3.44Number of Females 4.02 2.47Range:Number of Individuals Listedper LocationSame Site 5.90 4.82Same City/Different Site 2.56 3.52Different City/Different Site 2.18 3.41Number of Individuals in 4.27 3.66Same FunctionNumber of Different Functions 3.81 2.78Hierarchical Rank -0.07 0.60Hierarchical Range 0.87 0.41Density 0.26 0.24162TABLE B.2Non-Transition Respondent Sample Variable Means and Standard Deviations:Expressive Networks (n = 242)StandardVariables Means DeviationsNetwork Size and Gender Mix:Number of Individuals Listed 8.55 4.40Number of Males 5.01 3.55Number of Females 3.54 2.93Number of Overlapping Ties 2.50 2.06Number of Overlapping Female Ties 0.89 1.17# of Individuals Seen Outside of Work 4.40 3.74# of Females Seen Outside of Work 1 .84 2.49Range:Number of Individuals Listedper LocationSame Site 5.36 4.04Same City/Different Site 1 .98 2.90Different City/Different Site 1 .35 2.20Number of Individuals in 5.03 3.97Same FunctionNumber of Different Functions 2.39 2.11Hierarchical Rank -0.35 0.64Hierarchical Range 0.96 0.41Number of Supervisors 0.32 0.48Number of Subordinates 1 .34 1 .70Density 0.21 0.21Frequency of Contact 3.38 0.94163TABLE B.3aBank Respondent Sample Variable Means and Standard Deviations:Instrumental Networks (n = 83)GenderDifferenceNon-Variables Bank Male Female Parametric1Means/(SD) Means/(SD) Means/(SD) p-valuesNetwork Size & Gender Mix:Number of Individuals Listed 9.57 (4.54) 10.87 (4.05) 8.41 (4.67) .016Number of Males 5.39 (3.53) 6.56 (3.08) 4.34 (3.60) .003Number of Females 4.18 (2.47) 4.31 (2.33) 4.07 (2.61) .737Range:Number of Individuals Listedper LocationSame Site 3.43 (4.20) 3.87 (4.77) 3.02 (3.62) .614Same City/Different Site 2.83 (3.54) 2.74 (3.89) 2.91 (3.24) .318Different City/Different Site 3.43 (3.97) 4.26 (4.06) 2.67 (3.78) .024Number of Individuals in 3.39 (3.83) 3.36 (3.41) 3.42 (4.22) .665Same FunctionNumber of Different 3.93 (2.96) 4.51 (2.93) 3.40 (2.92) .064FunctionsHierarchical Rank 0.03 (0.66) -0.01 (0.70) 0.06 (0.62) .922Hierarchical Range 0.94 (0.39) 1.01 (0.36) 0.87 (0.41) .169Density 0.24 (0.23) 0.26 (0.23) 0.22 (0.23) .3861Mann-Whitney U non-parametric test.164TABLE B.3bForestry Respondent Sample Variable Means and Standard Deviations:Instrumental Networks (n = 60)GenderDifferenceNon-Variables Forestry Male Female Parametric1Means/(SD) Means/(SD) Means/(SD) p-valuesNetwork Size & Gender Mix:Number of Individuals Listed 11.32 (4.05) 10.81 (4.19) 12.50 (3.52) .174Number of Males 8.18 (3.37) 8.07 (3.60) 8.44 (2.81) .691Number of Females 3.13 (2.15) 2.74 (2.11) 4.06 (2.01) .010Range:Number of Individuals Listedper LocationSame Site 7.20 (3.86) 6.93 (3.97) 7.83 (3.62) .365Same City/Different Site 0.93 (1 .70) 0.64 (1 .08) 1 .61 (2.55) .447Different City/Different Site 3.18 (3.27) 3.24 (3.43) 3.06 (2.94) .870Number of Individuals in 3.60 (2.84) 3.69 (3.02) 3.39 (2.45) .929Same FunctionNumber of Different 3.88 (2.51) 3.50 (2.43) 4.78 (2.53) .061FunctionsHierarchical Rank 0.06 (0.59) 0.02 (0.60) 0.14 (0.59) .493Hierarchical Range 0.95 (0.48) 0.88 (0.50) 1.11 (0.39) .110Density 0.29 (0.25) 0.35 (0.28) 0.17 (0.11) .0161Mann-Whitney U non-parametric test.165TABLE B.3cInsurance Respondent Sample Variable Means and Standard Deviations:Instrumental Networks (n = 99)GenderDifferenceNon-Variables Insurance Male Female Parametric1Means/(SD) Means/(SD) Means/(SD) p-valuesNetwork Size & Gender Mix:Number of Individuals Listed 11.10(3.67) 10.94 (3.66) 11.44 (3.74) .578Number of Males 6.69 (3.02) 7.12 (3.01) 5.78 (2.87) .048Number of Females 4.41 (2.54) 3.82 (2.22) 5.66 (2.75) .001Range:Number of Individuals Listedper LocationSame Site 7.18 (5.05) 7.09 (4.84) 7.39 (5.55) .912Same City/Different Site 3.34 (4.00) 3.52 (4.21) 2.94 (3.54) .493Different City/Different Site 0.53 (2.06) 0.31 (1 .10) 1 .00 (3.28) .421Number of Individuals in 5.40 (3.68) 5.70 (3.87) 4.78 (3.21) .208Same FunctionNumber of Different 3.68 (2.80) 3.40 (2.87) 4.25 (2.59) .061FunctionsHierarchical Rank -0.22 (0.54) -0.18 (0.51) -0.30 (0.59) .441Hierarchical Range 0.77 (0.36) 0.75 (0.35) 0.80 (0.40) .579Density 0.26 (0.23) 0.27 (0.25) 0.23 (0.19) .639U non-parametric test.166TABLE B.4aBank Respondent Sample Variable Means and Standard Deviations:Expressive Networks (n = 83)GenderDifferenceNon-Variables Bank Male Female Parametric1Means/(SD) Means/(SD) Means/(SD) p-valuesNetwork Size & Gender Mix:Number of Individuals Listed 8.08 (4.29) 7.67 (4.47) 8.45 (4.13) .530Number of Males 3.95 (2.86) 4.92 (3.01) 3.11 (2.46) .005Number of Females 4.23 (3.14) 2.95 (2.29) 5.34 (3.36) .001Number of Overlapping Ties 2.12 (2.01) 2.18 (2.27) 2.07 (1.77) .773Number of Overlapping 0.96 (1 .35) 0.82 (1 .34) 1 .09 (1 .36) .359Female Ties# of Individuals Seen 4.42 (3.85) 4.44 (3.54) 4.41 (4.13) .708Outside of Work# of Females Seen Outside 2.38 (2.97) 1 .36 (1 .88) 3.20 (3.42) .006of WorkRange:Number of Individuals Listedper LocationSame Site 3.66 (3.40) 4.00 (3.73) 3.39 (3.13) .549Same City/Different Site 2.51 (3.08) 1.67 (2.53) 3.20 (3.33) .020Different City/Different Site 2.23 (2.77) 2.64 (2.71) 1.89 (2.81) .120Number of Individuals in 4.03 (3.45) 3.39 (2.92) 4.55 (3.79) .172Same FunctionNumber of Different 2.61 (2.03) 2.89 (2.14) 2.39 (1.93) .303FunctionsHierarchical Rank -0.35 (0.68) -0.23 (0.73) -0.46 (0.62) .073Hierarchical Range 1 .07 (0.37) 1 .06 (0.42) 1 .07 (0.32) .786Number of Supervisors 0.31 (0.47) 0.42 (0.50) 0.23 (0.42) .071Number of Subordinates 1 .59 (1.77) 1 .67 (1 .72) 1 .52 (1 .82) .574Density 0.18 (0.19) 0.20 (0.21) 0.16 (0.18) .456Frequency of Contact 3.18 (0.95) 3.43 (0.92) 2.97 (0.94) .0481Mann-Whitney U non-parametric test.167TABLE B.4bForestry Respondent Sample Variable Means and Standard Deviations:Expressive Networks (n = 60)GenderDifferenceNon-Variables Forestry Male Female Parametric1Means/(SD) Means/(SD) Means/(SD) p-valuesNetwork Size & Gender Mix:Number of Individuals Listed 7.95 (4.59) 7.55 (4.83) 8.89 (3.95) .274Number of Males 5.35 (3.47) 5.55 (3.64) 4.89 (3.08) .560Number of Females 2.60 (2.09) 2.00 (1 .98) 4.00 (1 .68) .001Number of Overlapping Ties 3.08 (2.39) 3.00 (2.46) 3.28 (2.27) .541Number of Overlapping 0.85 (1.05) 0.62 (1.01) 1.39 (0.98) .002Female Ties# of Individuals Seen 4.22 (3.80) 4.24 (3.89) 4.17 (3.67) .948Outside of Work# of Female Seen Outside 1 .33 (1.67) 1 .07 (1 .58) 1 .94 (1 .76) .018of WorkRange:Number of Individuals Listedper LocationSame Site 6.17 (3.97) 6.10 (4.30) 6.33 (3.18) .645Same City/Different Site 0.53 (0.98) 0.43 (0.89) 0.78 (1.17) .145Different City/Different Site 1.30 (1.80) 1.07 (1.64) 1.83 (2.07) .141Number of Individuals in 4.02 (3.24) 4.36 (3.66) 3.22 (1 .77) .506Same FunctionNumber of Different 2.85 (2.37) 2.32 (2.11) 4.06 (2.53) .017FunctionsHierarchical Rank -0.17 (0.68) -0.16 (0.72) -0.17 (0.59) .973Hierarchical Range 1.03 (0.46) 0.98 (0.51) 1.15 (0.33) .384Number of Supervisors 0.33 (0.54) 0.33 (0.57) 0.33 (0.49) .809Number of Subordinates 0.75 (1.34) 0.71 (1.35) 0.83 (1.34) .424Density 0.24 (0.24) 0.28 (0.24) 0.17 (0.23) .020Frequency of Contact 3.76 (0.85) 3.92 (0.71) 3.43 (1.04) .0841Mann-Whitney U non-parametric test.168TABLE B.4cInsurance Respondent Sample Variable Means and Standard Deviations:Expressive Networks (n = 99)GenderDifferenceNon-Variables Insurance Male Female Parametric1Means/(SD) Means/(SD) Means/(SD) p-valuesNetwork Size & Gender Mix:Number of Individuals Listed 9.29 (4.32) 9.31 (4.46) 9.25 (4.08) .976Number of Males 5.69 (3.93) 6.85 (3.94) 3.25 (2.57) < .001Number of Females 3.61 (3.05) 2.46 (1 .97) 6.00 (3.52) < .001Number of Overlapping Ties 2.47 (1.83) 2.61 (1.90) 2.19 (1.65) .366Number of Overlapping 0.85 (1 .09) 0.66 (0.93) 1 .25 (1 .30) .020Female Ties# of Individuals Seen 4.50 (3.64) 4.06 (3.69) 5.41 (3.42) .037Outside of Work# of Female Seen Outside 1 .71 (2.42) 0.76 (1 .27) 3.69 (3.00) < .001of WorkRange:Number of Individuals Listedper LocationSame Site 6.26 (4.16) 6.58 (4.21) 5.59 (4.03) .289Same City/Different Site 2.43 (3.25) 2.32 (3.12) 2.66 (3.54) .704Different City/Different Site 0.67 (1 .56) 0.52 (1 .03) 1 .00 (2.30) .470Number of Individuals in 6.47 (4.33) 6.91 (4.38) 5.56 (4.16) .117Same FunctionNumber of Different 1 .95 (1 .93) 1 .71 (1.80) 2.44 (2.11) .057FunctionsHierarchical Rank -0.45 (0.57) -0.44 (0.58) -0.49 (0.58) .499Hierarchical Range 0.82 (0.36) 0.79 (0.34) 0.87 (0.38) .205Number of Supervisors 0.32 (0.47) 0.38 (0.49) 0.19 (0.40) .058Number of Subordinates 1 .50 (1 .76) 1 .62 (1.95) 1 .25 (1 .27) .682Density 0.22 (0.20) 0.24 (0.21) 0.17 (0.18) .071Frequency of Contact 3.33 (0.92) 3.63 (0.83) 2.72 (0.79) < .0011Mann-Whitney U non-parametric test.169TABLE B.5Description of Variable Labels for Tables B.6 through B.1 1Instrumental Dependent VariablesINUM Number of Individuals ListedIMAL Number of Males ListedIFEM Number of Females ListedISITE1 Number of Individuals Listed at Same SiteISITE2 Number of Individuals Listed at Different Site/Same CityISITE3 Number of Individuals Listed at Different Site/Different CityISAMFUN Number of Individuals Working in Same Function as RespondentIDFUNCT Number of Different Functions Listed for Individuals ListedIHRANK Hierarchical Rank of Individuals ListedIHRANGE Hierarchical Range of Individuals ListedIDENSA Density of TiesExpressive Dependent VariablesFNUM Number of Individuals ListedFMAL Number of Males ListedFFEM Number of Females ListedFSEEOUT Number of Individuals Listed Seen Outside of WorkFSEEFEM Number of Females Listed Seen Outside of WorkFOVERLP Number of Overlapping Instrumental and Expressive TiesFOVEFEM Number of Overlapping FemalesFSITE1 Number of Individuals Listed at Same SiteFSITE2 Number of Individuals Listed at Different Site/Same CityFSITE3 Number of Individuals Listed at Different Site/Different CityFSAMFUN Number of Individuals Working in Same Function as RespondentFDFUNCT Number of Different Functions Listed for Individuals ListedFHRANK Hierarchical Rank of Individuals ListedFHRANGE Hierarchical Range of Individuals ListedFSLJPV Number of Supervisors Listed as ExpressiveFSUBOR Number of Subordinates Listed as ExpressiveFDENSA Density of TiesFCONTA Frequency of Contact170TABLE B.5 (continued)Description of Variable Labels for Tables B.6 through B.1 1Predictor VariablesINUM Covariate: Number of Instrumental Individuals ListedFNUM Covariate: Number of Expressive Individuals ListedDGENDER Gender Dummy VariableDJCADMIN Job Category/Administrative Dummy VariableDJCMGR Job Category/Manager Dummy VariableDCHILDR Child-Rearing Responsibility Dummy VariableDCOBANK Company/Bank Dummy VariableDCOFORE Company/Forest Dummy VariableDJBL1 2 Job Level/Executive-Sr. Manager Dummy VariableDJBL3 Job Level/Middle Manager Dummy VariableDJBL4 Job Level/First-Line Supervisor Dummy VariableDRACE Race Dummy VariableXGENBAN Gender x Company/Bank Interaction TermXGENFOR Gender x Company/Forest Interaction TermXJCABAN Job Category/Administrative x Company/Bank Interaction TermXJCAFOR Job Category/Administrative x Company/Forest Interaction TermXJCMBAN Job Category/Manager x Company/Bank Interaction TermXJCMFOR Job Category/Manager x Company/Forest Interaction TermXCHIBAN Child-Rearing x Company/Bank Interaction TermXCHIFOR Child-Rearing x Company/Forest Interaction TermICOVBAN Instrumental Covariate x Company/Bank Interaction TermICOVFOR Instrumental Covariate x Company/Forest Interaction TermFCOVBAN Expressive Covariate x Company/Bank Interaction TermFCOVFOR Expressive Covariate x Company/Forest Interaction Term171TABLEB.6Two-tailedSpearmanRankCorrelations:InstrumentalNetworkDependentVariablesrnumimalfernistelisite2ste3isamfunidfunctihrankihranjeirnal794***ifern553***.001isitel.541***.446***.286***isite2.099.097.1O8’•359***isite3.223***.231***.032.273***.183**isamfun.323***.247***.224***.203**.205**.152*idfunct.563***434***.384***.278***-.000.258***.329***ihrank.162*.030.338***.173**.023.146*165*-.066ihrange.198**.101.212**-.006.104.193**-.125’.287***-.075idensa-.063-.037-.033.017-.110-.056.207**.150*-.065.171*p<.10;*p<.05;**p<01;***p<001TABLEB.7Two-tailedSpearmanRankCorrelations:ExpressiveNetworkDependentVariables!n!irnfmalffemfseeoutlseefemfoverlpfovefemfsitelfsite2fsite3fmal.748***ffem.601***-.0025fseeout.631***447***359***fseefem.392***-.042.665***579***foverlp.531***.511***.228***.249***.021fovefem345***.048.513***.163*.262***.626***fsitel.621***.540***.318***.319***.152*.427***.219**fsite2.323***193**.224**.271***.205**.076.079.248***fsite3.302***.233***.179**.272***.167*.104.086.141*.029CA)fsamfun.664***549***.296***.504***.263***.361***.184**.408***.278***.135**fdfunct.392***.226***.392***.163*.172**.272***.218**345***.033.128*fhrank178**.012.242***.186**239***.089024-089135*-055fhrange.079-.078.215**.088.151*-.103-.036.020.026230***fsupv.203**.305***-.003.112k’-.003.076-.011.266***-.118.104fsubor.464***.290***.310***.369***.246***.092.063.241***.262***.209**fdensa.035.107-.”.015-.099fconta-.128”.054.215**-.065.147*.088.024.361***371***.329***p<.10;*p<.OS;**p<.Ol;***p<001TABLEB.7(continued)Two-tailedSpearmanRankCorrelations:ExpressiveNetworkDependentVariablesfsamfunfdfunctfhrankthranqefsupvsuborfdensafdfunctfhrank.177**-.018fhrange-.040.160*•345***fsupv.247***.016.240***-.037fsubor.451***.038.364***.308***.031fdensa.229**.135*.035-.116’.154*.053tconta.091.222**.043-.061.203**.034•333**p<.10;*p<.05;**p<.01;***p<.001TABLEB.8Two-tailedSpearmanRankCorrelations:ExpressiveNetworkandInstrumentalNetworkDependentVariablesmumimalifemstelisite2isite3isamfunidfunctihrankihranqeidensafnum.303***.219**.187**.126k.177**-.023.176**.200**-.033.014025fmal.217**347***-.118.134*.126-.021.147*.113k’.007-.003.050ffem185**-080.402***087085-.004060206**-003021010fseeout.203**.102.189**.043.146*.047.166*.160*.008.024.023fseetem.079-.103.272***.*.061-.008-.013foverlp.329***•334***.11O’.301***.059-.050.194**.208**-.061.047.191**fovefem.243***060359***.218**.070-.067.172**.106-.050.063.090fsitel335***.292***.122.536***197***135*.077.244***.007-.054-.031fsite2-.076-.074.038.366***.705***9**.126-.056-.013.017-.015(3,tsite3.019.028-.025.310***-.038•557***-.**.061fsamfun.120k.115k.030-.064.263***-.089.440***-.127k-.023.129*.038fdfunct.252***.147*.216**.298***.156**.105.336***.488***.026.130*-.017fhrank-.078.038.170**.025-.117.074-.019-.050.356***-.022.027fhrange.110k.050.148*-.074-.003.287***-.127.217**-.057349***.146*fsupv.041.085-.*-.032-.062fsubor.**.188**.057-.044.150*-.058.071.141*fdensa.051.117-.088-.***.155*-.<.10;*p<.05;**p<.01;***p<.001fnum.303***dgender-.066djcadmin.026.085djcmgr-.022.010dchildr-.116-.042dcobank.166*-.073dcofore.105-.082djbIl2.172**.026djbl3-.013.001djbl4-.115.038drace-.068-.088xgenban.184**.012xgenfor.098.076xjcaban.072.049xjcafor-.020.006TABLEB.9Two-tailedSpearmanRankCorrelations:PredictorVariablesmumfnumdqenderdjcadmindjcmqrdchildrdcobankdcoforedibIl2.046.167**-0)-.119.665***.209**-042-.049.210**-.026.127*.060-.104.001.232***.009.415***-.074-.*.091.071-.097.077-.038.129*-.085•353***.**359***.603***.096.091-.109.120k-.014.039.053-.099.194**-.014.010-.110.053-.012-.054.048.142*-.079.161*.057.019.319***-.046.009.169**.248***.077-.025.476***.207**-.072.080-.064.348***.410***.015.208**447***-.017.001“p<.10;*p<.05;**p<.01;***p<.001TABLEB.9(continued)Two-tailedSpearmanRankCorrelations:PredictorVariablesmumfnumdqenderd1cadmindicmqrdchildrdcobankdcoforedibll2xjcmban-.076-.032.092.189**.418***.046.133*.121”-.031-.047xjcmfor.087-.006.072.157*.051-.032-.067.113”.133*.072xchiban-.051-.031-.090-.013.028.121”.114”-.056.070-.013xchifor-.016.013-.053-.035-.040-.120”-.050.026-.039.092icovban-.021-.019.188**-.014-.058.004-.055-.005.066-.073icovfor.296***-.019.140*.033.049-.021.037.169**-.024-.111”fcovban-.008.157*-.020.031-.030-.024-.084.128”.033.039fcovfor.000.262***.038-.*-.089.081.033“p<.10;*p<.05;**p<.Ol;***p<.001TABLEB.9(continued)Two-tailedSpearmanRankCorrelations:PredictorVariablesç(jdracexcienbanxgenforxicabanxjcaforxjcmbanxjcmforxchibanxchiloricovbanicovfordrace.064xgenban.015.037xgenfor.092-.041.417***xjcaban-.037.144*.100.006xjcafor.070-.061.002.253***.285***xjcmban-.020.133*.115-.101.648***.029xjcmfor-.055.004-.001-.106.170**•543***.415***xchiban-.031-.063.291***.135*-.090.068-.027.023xchifor-.040.027.128*187**.075-.068.039-.078.423***icovban-.029.135*.151*.047.031-.022-.011-.030-.116k.088icovfor-.047-.095.090.005-.056-.*.417***fcovban-.018.043.049-.097.120-.116.003.023-.068.035.293***.165*fcovfor-.084-.053-.042.080-.104.104.003-.005.041-.028.213**.314***“p<.10;*p<.05;**p<.01;***p<.001TABLEB.9(continued)Two-tailedSpearmanRankCorrelations:PredictorVariablesfcovbanfcovfor“p<.10;*p<.05;**p<.01;***p<.001‘—.1 CoTABLEB.10Two-tailedSpearmanRankCorrelations:InstrumentalNetworkDependentVariablesandPredictorVariablesimalitemisitelisite2isite3isamfunidfunctihrankihranieidensadgender.218**.192**-.087.061-.001-.*mum794***553***.541***.099.223***.323***.563***.162*.198**-.063djcadmin.132*.207**-.027.100.003-.***.057394***.087.316***.226***013.103127*-076dcofore.264***.214**.193**.266***.263***-.*.076djbll2.241***-.036.167*-.019.104.139*.204**.214**.155*-.043djbl3-.016-.038.196**-.003.118k-.030-.004.075.099-.0490djbl4.218**.155*.019.123.275***.108k’-.076.001.178**.086drace-.062-.053.006.015-.096.014.139*.020-.152*-.048xgenban.134*.133*-.070.066.137*-.010.183**.006-.119-.006xgenfor.*-.059.240**-.082.198**.149*-.093-.080-.080.028p<.10;*p<.05;**p<.01;***p<.001TABLEB.10(continued)Two-tailedSpearmanRankCorrelations:Instrumentalimalifemisitelisite2-.064.018-.065.103-.096-.048-.117k-.**.254**.163*-.023NetworkDependentVariablesandPredictorVariablesisite3isamfunidfunctihrankihrange-.096.043-.051.047-.115.128*-.***-.118k.255***.066xjcaforxjcmbanxjcmforxchibanxchiforicovbanicovfor-idensa-.031.111-.**-.119kp<.10;*p<.05;**p<.01;***p<.001TABLEB.11Two-tailedSpearmanRankCorrelations:ExpressiveNetworkDependentVariablesandPredictorVariablesfmalffemfseeoutfseetemwoverlpwovefemfsitelfsite2tsite3fsamfunfdfunctfhrankdgender.331***.461***.069•447***-.043.220**-.121.175**.057-.070.131*-.096fnum.748***.601***.631***.392***.531***345***.621***.323***.302***.664***.392***.178**djcadmin-.120.241***-.022.114’-.’.098.108k’.074-.111’.141*dchildr.027-.066-.055-.154*-.020-.036.008-.090.075-.016.002.004dcobank201**148*-.005.135*.153*.011326***.176**252***182**094-030dcofore065164*-.035-.086.138*.007.131*.297***.048131*.104174**djbll2.138*-.085-.059-.100.214**-.012.074-.044.004.137*.248***.174**djbl3-.*-.062.129*.066.196**.040-.063-.003F%)djbl4-.*.126.129*.033drace-.103.007-.081-.046-.123-.117k’-.041-.007-.088-.022-.014-.050xgenban-.037.031-.072-.011.011-.050-.052.112k-.121’.107-.134-.080xgenfor.167**-.033-.028-.”-.076-.024.146*-.119k’.048.183**-.093-.088.070-.036-.070“p<.10;*p<.05;**p<.01;***p<.001p<.10;*p<.05;**p<.01;***p<.001TABLEB.11(continued)Two-tailedSpearmanRankCorrelations:ExpressiveNetworkDependentVariablesandPredictorVariablesfmalffemfseeoutfseefemwoverlpwovefemfsiteltsite2fsite3fsamfunfdfunctfhrankxjcafor.105-.096.005-.023-.117-.114-.027.081-.038.043-.096-.016xjcmban-.”-.087.052.006xchitor-.*-.025-.059-.007-.143-.056.177**-.126”.013.157*-.076-.063fcovfor.156*133*-.114”-.028-.096-.035-.056-.083-.037.227***.084.071C)TABLEB.11(continued)Two-tailedSpearmanRankCorrelations:ExpressiveNetworkDependentVariablesandPredictorVariables-fhrangefsupvfsuborfdensafcontadgender.144*.137*.013.212**.363***fnum.079.203**.464***.035-.128kdjcadmin-.047-.073.058-.049-.029djcmgr.054.066.150*.001-.105dchildr.006.013.023-.010.043dcobank.214**-.002.114k.145*.165*dcofore.119-.004.235***.075.234***djbll2.087-.075.062-.035-.036djbl3.145*-.037.119-.101-.125djbl4-.155*.070-.037.056.024drace.034-.037-.022-.016-.060xgenban-.014-.063-.030.032.017xgenfor-.054.111-.121”-.043.020xjcaban-.020-.“p<.10;*p<.05;**p<.01;***p<.001TABLEB.11(continued)Two-tailedSpearmanRankCorrelations:ExpressiveNetworkDependentVariablesandPredictorVariablesthranqefsupvfsuborfdensafcontaxjcafor-.190.205**-.054.038-.049xjcmban-.069.046-.106-.048.036xjcmfor.159*.185**.157*.002-.005xchiban.043.141*-.008.027-.051xchitor-.045-.092.040-.002.032tcovban-.023-.109k.044-.002-.059fcovfor.054.042.212**.015.066p<.10;*p<.05;**p<.01;***p<.00101TABLE B.12Promotion Study Variable Means and Standard Deviations:Instrumental Networks (n = 33)Time 1 Time 2Variables Means/(SD) Means/(SD)Network Size and Gender Mix:Number of Individuals Listed 7.45 (4.70) 8.33 (4.63)Number of Males 3.36 (3.15) 4.39 (3.54)Number of Females 4.09 (3.07) 3.94 (2.42)Range:Number of Individuals Listedper LocationSame Site 2.61 (3.59) 2.33 (3.20)Same City/Different Site 3.32 (3.40) 3.76 (4.39)Different City/Different Site 2.00 (2.63) 2.24 (3.53)Number of Individuals in 3.42 (3.35) 3.21 (3.18)Same FunctionNumber of Different Functions 2.87 (2.30) 2.91 (2.32)Hierarchical Rank 0.07 (0.88) -0.03 (0.82)Hierarchical Range 1 .03 (0.45) 1 .00 (0.45)Density 0.35 (0.25) 0.33 (0.32)Time 1: First promotion questionnaireTime 2: Follow-up promotion questionnaire186TABLE B.13Promotion Study Variable Means and Standard Deviations:Expressive Networks (n = 33)Time 1 Time 2Variables Means/(SD) Means/(SD)Network Size and Gender Mix:Number of Individuals Listed 8.82 (4.34) 8.52 (4.53)Number of Males 3.03 (2.44) 2.91 (2.60)Number of Females 5.79 (3.45) 5.61 (3.54)Number of Overlapping Ties 2.03 (1 .83) 2.64 (1 .87)Number of Overlapping Female Ties 1 .30 (1 .45) 1 .73 (1 .35)# of Individuals Seen Outside of Work 5.84 (4.10) 4.70 (3.56)# of Females Seen Outside of Work 4.28 (3.27) 3.15 (2.92)Range:Number of Individuals Listedper LocationSame Site 2.39 (2.73) 2.76 (2.68)Same City/Different Site 4.27 (4.25) 3.52 (3.65)Different City/Different Site 2.15 (2.73) 2.24 (3.46)Number of Individuals in 4.73 (3.65) 4.55 (3.76)Same FunctionNumber of Different Functions 2.45 (2.35) 2.12 (1.65)Hierarchical Rank -0.37 (0.71) -0.45 (0.60)Hierarchical Range 1.10 (0.37) 0.98 (0.37)Number of Supervisors 0.27 (0.52) 0.30 (0.47)Number of Subordinates 1 .37 (1 .72) 1 .55 (1 .79)Density 0.22 (0.20) 0.27 (0.28)Frequency of Contact 2.76 (0.85) 3.21 (0.98)Time 1: First promotion questionnaireTime 2: Follow-up promotion questionnaire187TABLE B.14Kilmogorov-Smirnov (Lilliefors) Test for Normality1:Instrumental Network and Expressive Network Dependent Variables (n = 242)Instrumental Network Expressive NetworkVariables p-values p-valuesNumber of Individuals Listed < .001 < .01Number of Males Listed < .001 < .001Number of Females Listed < .001 < .001Number of Individuals Seen < .001Outside of WorkNumber of Females Seen < .001Outside of WorkNumber of Overlapping Ties < .001Number of Overlapping Female Ties < .001Number of Individuals Listed at < .001 < .001Same Site/Same CityNumber of Individuals Listed at < .001 < .001Different Site/Same CityNumber of Individuals Listed at < .001 < .001Different Site/Different CityNumber of Individuals in < .001 < .001Same FunctionNumber of Different Functions < .001 < .001Hierarchical Rank .075 .057Hierarchical Range > .20 .193Number of Supervisors < .001Number of Subordinates < .001Density < .001 < .001Frequency of Contact > .201H0: Data normally distributed.188Appendix C:RESPONDENT SAMPLE DEMOGRAPHICSData were collected on 17 demographic variables. The first set, the “numerical”data, included: the age of the individual, the number of children each respondent has, theages of the youngest and oldest child, time working since 1 8, time with employer, time inposition, the number of promotions, number of previous positions, the total number ofsubordinates and the total number of female subordinates. The other set, the “categorical”data, was collected on the study participants’ education level, ethnicity, job category, joblevel, relationships status, and salary level.THE NON-TRANSITION STUDY CATEGORICAL DEMOGRAPHIC DATAThe cross-tabulation percentages1for the entire respondent sample, for all males,for all females, and for each company are provided in Table C.1. Significant categoricaldifferences between men and women and also among companies are listed in Table C.2.The Pearson Chi-Square test was used to test for the categorical differences.Significant males-female differences. There were significant job categorydifferences between men and women. The cross tabulation conveyed an interesting,though not surprising, phenomenon where 20.4% of the women were in administrative oradministrative/management roles, compared to only 8.8% for the men. Moreover, womenwere less likely to categorize their jobs as technical/management orprofessional/management in comparison to men, 20.4% to 38.5%, respectively. This trendwas especially pronounced at the insurance company where women comprised 29.0%[versus 7.5% for the men] of the administrative or administrative/management roles.Alternatively, 86% of the men were in professional or technical management positions atthe insurance company compared to 61 .3% of the women.‘The categories used for the Chi-Square tests were collapsed (from those provided in thequestionnaire) for the following variables: education, ethnicity, job category, job level, andsalary.189TABLE C.1Non-Transition Study Cross-Tabulation Percentages:Categorical Demographic Data (n = 242)Entire Male Female Bank Forest Insur.EDUCATIONSome High School or 19.0 16.9 22.3 22.9 10.0 21.2High School GradSome College/University, 63.6 67.6 57.4 65.1 63.3 62.6College Diploma, orUniversity DegreeSome Graduate School, 17.4 15.5 20.2 12.0 26.7 16.2Advance Degree, or OtherETHNICITYCaucasian 93.3 95.2 90.3 93.8 91 .7 93.9Other 6.7 4.8 9.7 6.2 8.3 6.1JOB CATEGORYManagement 42.7 39.9 47.3 62.7 18.3 40.8Administrative or 13.3 8.8 20.4 12.0 13.3 14.3Administrative ManagementProfessional/Management or 31.5 38.5 20.4 19.3 38.3 37.8Technical/ManagementProfessional, Technical, 12.4 12.7 11.7 6.0 30.0 7.1or OtherJOB LEVELExecutive or Sr. Management 17.4 19.6 13.8 24.1 23.3 8.1Middle Manager 37.2 34.5 41.5 45.8 30.0 34.3First-line Manager 38.0 36.5 40.4 27.7 21.7 56.6Other 7.4 9.5 4.3 2.4 25.0 1.0RELATIONSHIP STATUSMarried or Living with a 86.8 90.5 80.9 89.2 80.0 88.9Long-term PartnerNot Married or Living with 13.2 9.5 19.1 10.8 20.0 11.1a Long-term PartnerSALARYUnder $49,999 10.7 6.1 18.1 24.1 10.0 0.0$50,000 to 74,999 67.4 68.9 64.9 44.6 63.3 88.9$75,000 to 99,999 14.5 16.2 11.7 15.7 18.3 11.1$100,000 and over 7.4 8.8 5.3 15.7 8.3 0.0190TABLE C.2Between Gender and Among Company Non-Transition StudySignificant (p < .10) Differences: Categorial Demographic DataGender CompanyEducation .094EthnicityJob Category .006 .000Job Level .000Relationship .030Salary .022 .000191Men and women differed significantly as to their relationship status, where 90.5%of the men were married or living with a long-term partner compared to 80.9% of thewomen. This trend was particularly evident at the bank, where 97.4% of the men weremarried or living with a long-term partner versus 81 .8% of the women. Finally, there weresignificant salary differences between men, where a higher percentage of men earned$50,000 or more in comparison to the women.Significant between-comrany differences. Significant differences among the threecompanies were also found. The bank had more employees classify themselves asmanagement (62.7%), whereas 30% of the forestry employees saw themselves asprofessional, technical or administrative, with no management designation. Furthermore,when looking at the job level of the respondents, the bank respondents were more likely tobe middle managers (45.8%), whereas 56.5% of the insurance respondents were first-linesupervisors. Finally, the majority of respondents earned between $50,000 and $74,999 atboth the forestry (63.3%) and insurance (88.9%) respondent pools. For the bank, thisfigure was 44.6%. Over 24% of the bank employees earned under $49,999, compared to10% of the insurance respondents and 0% of the respondents from the forestry sample.THE NON-TRANSITION STUDY NUMERICAL DEMOGRAPHIC DATATables C.3 and C.4 provide the means and standard deviations on the numericaldemographic variables. Male and female means can be found in Table C.3 with thesignificant p-values for male-female differences. Table C.4 lists the entire response setmeans and standard deviations and the response set means and standard deviations foreach company. Table C.4 also notes on which variables the three companies differedsignificantly.Age was the only numerical demographic variable normally distributed, and malefemale age differences were tested using the two independent sample t-tests method.Otherwise, the Mann-Whitney U test for two independent samples was used for the othernumerical demographic variables as the alternate to the t-test and its assumptions and192TABLE C.3Non-Transition Study Male-Female Means and Standard Deviations:Numerical Demographic Data (n = 242)GenderDifferenceNon-Male Female ParametricVariables Meansl(SD) Means/(SD) p-values1Age 46.32/(8.21) 43.34/(6.77) .039Time Working since 18 24.18/(9.32) 20.241(6.66) .010Time with Employer 14.96/(8.04) 13.14/(6.75) .095Time in Position 4.95/(3.80) 3.711(2.81) .001# of Promotions 3.41/(3.23) 3.53/(3.09)# of Previous Positions 3.671(2.89) 3.87/(2.70)# of Subordinates 7.14/(10.33) 8.94/(1 0.38) .039# of Subordinates! 3.86/7.23) 6.911(8.77) < .001Female# of Children 1.98/(1.10) 1.12/(1.14) < .001Age of Oldest Child 16.70/(9.74) 16.60/(9.83)Age of Youngest Child 1 3.52/(8.77) 13.78/(8.95)‘Mann-Whitney U non-parametric test; p < .10 listed.193TABLE C.4Non-Transition Study Company Means and Standard Deviations:Numerical Demographic Data (n = 242)CompanyDifferenceNon-Overall Banking Forestry Insurance ParametricVariables Means/(SD) Means!(SD) Means/(SD) Means/(SD) p-values’Age 45.17! 43.32/ 44.58/ 47.05! .006(7.80) (6.90) (8.33) (7.83)Time Working 22.66/ 20.82/ 22.38/ 24.38! .037since 18 (8.60) (7.66) (9.56) (8.47)Time with Employer 14.25/ 16.87/ 13.13! 12.74/ .009(7.60) (8.09) (8.78) (5.65)Time in Position 4.46/ 3.55! 5.54/ 4.57/ .010(3.50) (2.32) (4.79) (3.22)# of Promotions 3.46/ 6.13/ 1.58/ 2.42! < .001(3.17) (3.57) (1.63) (1.84)# of Previous 3.75/ 5.95/ 2.18/ 2.89/ < .001Positions (2.81) (2.87) (2.04) (1.94)# of Subordinates 7.83! 7.86! 3.35/ 10.55/ < .001(10.37) (9.59) (5.79) (12.16)# of Subordinates! 5.04/ 5.89! 1 .59! 6.43! < .001Female (7.98) (8.26) (3.51) (9.08)# of Children 1.64/ 1.73/ 1.58/ 1.61/(1 .19) (1 .22) (1 .17) (1 .19)Age of Oldest Child 16.67/ 16.37/ 15.11/ 17.88/(9.74) (9.27) (10.11) (9.92)Age of Youngest 13.59/ 13.83/ 12.39/ 14.13/Child (8.79) (8.49) (9.04) (8.98)‘Kruskal-Wallis one-way analysis of variance test; p < .10 listed.194requirements. The Kruskal-WalIis one-way analysis of variance was used for testscomparing company means.Male-female differences. The males and females differed significantly (at the .05level) on the following demographic variables: age, time employed since 1 8, time inposition, number of subordinates, number of subordinates that are female, and the numberof children. Except for the insurance company, the men were, on average, 3 years olderthan the females. In all three companies, men had held their positions longer than thefemales. Women had more subordinates, on average, especially at the bank, and notsurprisingly, women’s subordinates were more likely to be females. Finally, women acrossall three companies consistently had fewer children than did the male respondents.Between company differences. Except for number of children and ages of theyoungest and oldest child, the three companies had significant differences across theremaining demographic variables.Non-transition study summary. The previous discussion highlighted the male-femaleand between-company differences. The male-female demographic differences are the mostsalient to this thesis, and there were number of gender differences. However, thedifferences were not great, and the majority have no relevance to the research questions orhypotheses in Chapter Two. The lone exception is job category, which is being controlledfor in the research models, as outlined in Chapter Three.THE PROMOTION STUDY DEMOGRAPHIC DATAThe promotion study demographic data were collected in the first questionnaire.Forty-three individuals returned the first questionnaire, and of those, only 33 returned thefollow-up questionnaire. The demographics of the 43-person and 33-person respondentsamples were compared, and there were no significant numerical differences.2 Moreover,2The Wilcoxon matched-pairs signed-ranks test was used to test for numericaldemographic data differences. Chi-Square tests were not conducted on the categorial databecause of small cell sizes.195TABLE C.5Promotion Study Cross-Tabulation Percentages:Categorical Demographic Data (n = 33)Entire Male FemaleEDUCATIONSome High School or 36.3 15.4 50.0High School GradSome College/University, 51.6 84.6 30.0College Diploma, orUniversity DegreeSome Graduate School, 12.1 0.0 20.0Advance Degree, or OtherETHNICITYCaucasian 93.8 100.0 89.5Other 6.2 0.0 10.5JOB CATEGORYManagement 57.6 69.2 50.0Administrative or 30.2 7.7 45.0Administrative ManagementProfessional/Management or 6.1 1 5.4 0.0Technical/ManagementProfessional, Technical, 6.1 7.7 5.0or OtherJOB LEVELExecutive or Sr. Management 18.2 23.1 15.0Middle Manager 42.4 46.2 40.0First-line Manager 30.3 30.8 30.0Other 9.1 0.0 15.0RELATIONSHIP STATUSMarried or Living with a 69.7 76.9 65.0Long-term PartnerNot Married or Living with 30.3 23.1 35.0a Long-term PartnerSALARYUnder $49,999 45.5 30.8 55.0$50,000 to 74,999 48.5 69.2 35.0$75,000 to 99,999 0.0 0.0 0.0$100,000 and over 6.0 0.0 10.0196TABLE C.6Promotion Study Male-Female Means and Standard Deviations:Numerical Demographic Data (n = 33)Entire Male FemaleAge 39.921(6.87) 38.92/(7.44) 39.40/(6.67)Time Working since 18 1 7.70/(8.57) 1 7.46/(8.77) 17.85/(8.66)Time with Employer 13.98/(7.56) 1 4.23/(6.47) 13.81/(8.36)Time in Position 3.22/(2.89) 3.571(3.40) 2.99/(2.56)# of Promotions 5.48/(3.26) 6.541(3.55) 4.80/(2.95)# of Previous Positions 5.721(3.46) 6.1 5/(3.26) 5.45/(3.63)# of Subordinates 7.331(11.83) 4.62/(4.65) 9.101(14.62)# of Subordinates/Female 5.95/(9.97) 3.85/(4.36) 7.32/(12.26)# of Children 1 .45/(1 .23) 1 .38/(1 .26) 1 .50/(1 .24)Age of Oldest Child 12.92/(8.67) 12.17/(10.65) 13.41/(7.54)Age of Youngest Child 10.65/(6.95) 13.17/(8.13) 9.27/(6.20)197there were no significant numerical differences between the 1 3 males and 20 femalescomprising the 33-person respondent sample. The categorial data percentages andnumerical variable means and standard deviations for the 33 respondents are provided inTables C.5 and C.6, respectively.Finally, the 33 respondents were compared to the non-transition banking respondentsample. The two samples differed significantly on relationship status, salary, age, timeworking since 1 8, and time with employer. The promotion respondents were younger andonly 69.7% of them were married or living with a long-term partner. This was incomparison to the non-transition banking respondent sample, where 89.2% were married orliving with a long-term partner. Also, the salary for the promotion sample was lower, as21 .3% of the respondents made $20,000 to $39,999 compared to 1 .2% of the nontransition banking employees. The differences between the non-transition and promotionrespondent samples seem logical in that those receiving promotions would more than likelyhave lower hierarchical positions. The promotion respondent sample would more likely thannot include younger individuals trying to increase both their hierarchical positions andsalaries.198C m -U m z C m z -1-a (0 (0maw.m -IC C-) 0 m m -I, 0 m 2 -1 Cl)TABLE D.1Description of Predictor Variable Labels for Tables D.2 through D.30VariableLabel DescriitionC:Number Covariate Number of Individuals ListedI:Gender Independent Gender of RespondentD:Bank Moderator Company/BankD :Forest Moderator Company/ForestD:Admin Dummy Job Category/AdministrativeD:Manager Dummy Job Category/ManagerD:Children Dummy Child-Rearing ResponsibilityX:GenBan Interaction Gender x Company/BankX:GenFor Interaction Gender x Company/ForestX:AdmBan Interaction Job Category/Administrative x Company/BankX :AdmFor Interaction Job Category/Administrative x Company/ForestX :MgrBan Interaction Job Category/Manager x Company/BankX:MgrFor Interaction Job Category/Manager x Company/ForestX:ChiBan Interaction Child-Rearing Responsibility x Company/BankX:ChiFor Interaction Child-Rearing Responsibility x Company/ForestX:CovBan Interaction # of Individuals Listed x Company/BankX:CovFor Interaction # of Individuals Listed x Company/Forest200TABLE D.2Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals ListedBase hifi ModeratorI:Gender -0.715 -0.374 -0.255(0.556) (0.564) (0.559)D:Admin 0.341 0.571 -0.383(1.061) (1.071) (1.158)D:Manager -0.136 0.292 -0.599(0.818) (0.852) (0.960)D:Bank 1.455* -1.001(0.623) (0.636)D:Forest 0.276 0.607(0.700) (0.705)X:GenBan 3.173*(1.277)X:GenFor 1 .069(1.458)X:AdmBan -1.460(2.900)X:AdmFor 1.963(2.540)X:MgrBan -2.878(2.459)X:MgrFor 1 .696(1.983)(Constant) 10.962*** 10.911*** 11.737***(0.784) (0.905) (1.001)R-Squared 0.008 0.038 0.095‘p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.201TABLE D.3Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Males ListedBase Shift ModeratorC:Number 0.660*** 0.646*** 0.648***(0.030) (0.029) (0.031)l:Gender 1.O43*** O.898*** O.977***(0.258) (0.254) (0.261)D:Admin 2.O44*** 1.615** 1.737**(0.491) (0.482) (0.531)D:Manager O.858* -0.313 -0.480(0.378) (0.384) (0.443)D:Bank -0.149 -0.098(0.284) (0.293)D:Forest 1.252*** 1.230***(0.315) (0.331)X:GenBan 0.877(0.598)X:GenFor 0.643(0.675)X:AdmBan 2.376A(1.328)X:AdmFor -1.122(1.167)X:MgrBan -1.665(1.134)X:MgrFor -0.683(0.913)X:CovBan 0.021(0.070)X:CovFor 0.129(0.080)(Constant) 0.910 0.281 0.378(0.490) (0.518) (0.582)R-Squared 0.695 0.720 0.730Ap<.lO; *p<.O5 **p<.ol; ***p<001Standard errors are in parentheses.202TABLE D.4Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Females ListedJufi ModeratorC:Number O.340*** 0.354*** O.352***(0.030) (0.029) (0.031)l:Gender 1.043*** O.898** 0.978***(0.258) (0.254) (0.261)D:Admin 2.042*** 1.613** 1.739**(0.491) (0.482) (0.531)D:Manager 0.858* 0.313 0.483(0.378) (0.384) (0.443)D:Bank 0.149 0.097(0.284) (0.293)D:Forest 1.252*** 1.230***(0.315) (0.331)X:GenBan -0.876(0.598)X:GenFor -0.643(0.675)X:AdmBan 2.386(1.328)X:AdmFor 1.117(1.167)X:MgrBan 1.679(1 .134)X:MgrFor 0.680(0.913)X:CovBan -0.021(0.070)X:CovFor -0.129(0.080)(Constant) 0.910A -0.281 -0.382(0.490) (0.518) (0.582)R-Squared 0.409 0.456 0.476p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.203TABLE D.5Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Same SiteBase hifi ModeratorC:Number 0.650*** 0.596*** O.618***(0.063) (0.061) (0.063)l:Gender -0.271 0.305 0.140(0.542) (0.526) (1.103)D:Admin -1.201 -0.952 1.235A(1.031) (0.999) (1.103)D:Manager -0.773 -0.215 -0.613(0.794) (0.795) (0.919)D:Bank 2.914*** -3.021 ***(0.588) (0.608)D:Forest -0.156 -0.062(0.653) (0.687)X:GenBan 0.147(1.240)X:GenFor -0.209(1.401)X:AdmBan 2.972(2.756)X:AdmFor 2.858(2.421)X:MgrBan 1.137(2.353)X:MgrFor 1.745(1.896)X:CovBan 0.329*(0.144)X:CovFor -0.199(0.166)(Constant) -0.168 0.778 0.911(1.028) (1.073) (1.208)R-Squared 0.319 0.391 0.414A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.204TABLE D.6Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Different Site/Same CityBase Shift ModeratorC:Number O.154A O.169** O.171**(0.054) (0.053) (0.055)l:Gender 0.218 0.068 0.212(0.464) (0.461) (0.473)D:Admin 2.079* 1.375 2.016*(0.883) (0.876) (0.963)D:Manager 1.836** 0.985 1.605*(0.680) (0.696) (0.803)D:Bank -0.262 -0.234(0.515) (0.531)D:Forest 2.214*** 2.138***(0.572) (0.600)D:GenBan 1.863’(1.084)D:GenFor 1.642(1.224)X:AdmBan -1.834(2.407)X:AdmFor -2.625(2.115)X:MgrBan 0.011(2.055)X:MgrFor -1.838(1.656)X:CovBan 0.218A(0.126)X:CovFor -0.014(0.145)(Constant) -0.801 0.468 -0.244(0.881) (0.941) (1.055)R-Squared 0.064 0.125 0.159p<.lO; *p<•o5 **p<Ol; ***p<001Standard errors are in parentheses.205TABLE D.7Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Different Site/Different CityBase Shift ModeratorC:Number 0.184** 0.226*** O.207***(0.052) (0.048) (0.050)I:Gender 0.119 -0.327 -0.350(0.451) (0.416) (0.427)D:Admin -0.790 -0.319 -0.644(0.858) (0.789) (0.871)D:Manager -0.91 1 -0.611 -0.820(0.661) (0.627) (0.726)D:Bank 3.317*** 3.417***(0.464) (0.480)D:Forest 2.461*** 2.268***(0.515) (0.543)X:GenBan -1.772(0.980)X:GenFor -1.367(1.107)X:AdmBan -2.087(2.177)X:AdmFor -0.895(1.913)X:MgrBan -1.921(1.858)X:MgrFor -0.533(1.497)X:CovBan 0.073(0.114)X:CovFor O.232(0.131)(Constant) 0.961 -1.344 -0.859(0.856) (0.848) (1.954)R-Squared 0.058 0.242 0.266p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.206TABLE D.8Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Same FunctionJ1tI ModeratorC:Number 0.316*** 0.304*** 0.334***(0.054) (0.053) (0.055)J:Gender -0.398 -0.270 -0.116(0.462) (0.461) (0.466)D:Admin 0.095 -0.384 0.687(0.880) (0.875) (0.951)D:Manager 0.332 -0.134 0.857(0.678) (0.696) (0.792)D:Bank 1.495** 1.461**(0.515) (0.524)D:Forest 1.916** 1.824**(0.572) (0.592)D:GenBan 1 .926k’(1.069)D:GenFor 0.404(1.208)X:AdmBan 0.109(2.375)X:AdmFor -2.264(2.087)X:MgrBan -0.982(2.028)X:MgrFor 3.422*(1.634)X:CovBan -0.021(0.124)X:CovFor 0.296*(0.143)(Constant) 0.804 2.277* 0.830(0.878) (0.940) (1.041)R-Squared 0.135 0.185 0.239‘p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.207TABLE D.9Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Different Functions ListedBase Shift ModeratorC:Number 0.378*** O.391*** 0.389***(0.036) (0.037) (0.038)l:Gender 0.513 0.376 0.272(0.312) (0.317) (0.326)D:Admin 0.068 0.052 -0.117(0.594) (0.603) (0.664)D:Manager 0.352 0.267 1.132(0.458) (0.479) (0.553)D:Bank 0.764* 0.751*(0.355) (0.366)D:Forest 0.191 0.262(0.394) (0.414)X:GenBan -0.582(0.747)X:GenFor 0.240(0.844)X:AdmBan -1.044(1.660)X:AdmFor -0.989(1.459)X:MgrBan 1 .042(1.417)X:MgrFor 1.127(1.142)X:CovBan -0.082(0.087)X:CovFor -0.034(0.100)(Constant) -0.671 -0.999 -0.808(0.592) (0.647) (0.728)R-Squared 0.317 0.331 0.355A p < .10; * p < .05; ** p < .01; p < .001Standard errors are in parentheses.208TABLE D.1OInstrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Hierarchical Rank of Individuals ListedBase Shift ModeratorC:Number 0.O21* O.O2O* -O.O19(0.009) (0.009) (0.010)I:Gender 0.015 -0.005 -0.031(0.081) (0.082) (0.083)D:Admin -0.154 -0.092 -0.169(0.154) (0.155) (0.168)D:Manager -0.166 -0.108 -0.206(0.119) (0.123) (0.140)D:Bank 0.215* 0.171’(0.091) (0.093)D:Forest 0.254* O.196A(0.101) (0.105)X:GenBan 0.103(0.189)X:GenFor 0.160(0.214)X:AdmBan 0.453(0.420)X:AdmFor O.723(0.369)X:MgrBan 0.437(0.359)X:MgrFor 0.263(0.289)X:CovBan 0.003(0.022)X:CovFor 0.068**(0.025)(Constant) 0.30O 0.099 0.208(0.154) (0.167) (0.184)R-Squared 0.030 0.064 0.128A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.209TABLE D.11Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Hierarchical Range of Individuals ListedBase Shift ModeratorC:Number O.019** O.021** 0.020**(0.006) (0.006) (0.007)I:Gender 0.048 0.025 0.009(0.055) (0.055) (0.056)D:Admin 0.054 0.099 0.110(0.104) (0.104) (0.114)D:Manager 0.039 0.077 0.084(0.080) (0.082) (0.095)D:Bank 0.201** 0.211***(0.061) (0.063)D:Forest 0.199 0.197**(0.068) (0.071)X:GenBan -0.139(0.129)X:GenFor 0.149(0.145)X:AdmBan -0.162(0.286)X:AdmFor -0.344(0.25 1)X:MgrBan -0.029(0.244)X:MgrFor -0.087(0.196)X:CovBan -0.013(0.015)X:CovFor 0.014(0.017)(Constant) O.612*** 0.446*** 0.453***(0.104) (0.111) (0.125)R-Squared 0.041 0.096 0.129p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.210TABLE D.12Instrumental Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Densitybift ModeratorC:Number -0.01 1 * * 0.012* * 0.012* *(0.004) (0.004) (0.004)l:Gender O.O77* O.O7O* -0.053(0.032) (0.033) (0.033)D:Admin O.119A O.11OA -0.051(0.061) (0.062) (0.067)D:Manager -0.074 -0.061 0.009(0.047) (0.050) (0.056)D:Bank -0.024 -0.034(0.037) (0.037)D:Forest 0.019 0.020(0.041) (0.042)X:GenBan 0.022(0.076)X:GenFor -0.078(0.086)X:AdmBan 0.117(0.168)X:AdmFor -0.105(0.148)X:MgrBan 0.252A(0.144)X:MgrFor -0.016(0.116)X:CovBan 0.O15’(0.009)X:CovFor 0.017A(0.010)(Constant) O.481*** 0.478*** 0.407***(0.061) (0.067) (0.074)R-Squared 0.080 0.084 0.154p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.211TABLE D.13Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed.Shift Moderatorl:Gender 0.264 0.450 0.469(0.604) (0.620) (0.635)D:Admin 1.823 1.591 2.119A(1.132) (1.150) (1.282)D:Manager 1.105 0.842 1.234(0.875) (0.919) (1.065)D:Children -0.285 -0.197 -0.165(0.585) (0.587) (0.597)D:Bank 1.281A 1.405*(0.671) (0.711)D:Forest -1.129 -1.135(0.749) (0.780)X:GenBan 0.825(1.464)X:GenFor 1 .509(1.630)X:AdmBan 3.322(3.203)X:AdmFor 0.229(2.819)X:MgrBan 2.375(2.728)X:MgrFor -0.218(2.201)X:ChiBan 0.018(1.395)X:ChiFor 0.431(1.498)(Constant) 7.600*** 8.357*** 7.951***(0.942) (1.065) (1.195)R-Squared 0.015 0.033 0.044p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.212TABLE D.14Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Males ListedModeratorC:Number 0.627* * * 0.624* * * 0.617* * *(0.028) (0.028) (0.027)l:Gender 2.610*** 2.425*** -2.341 ***(0.263) (0.268) (0.256)D:Admin 1.499** 1.273* 0.997A(0.496) (0.498) (0.521)D:Manager -0.185 0.144 0.261(0.382) (0.397) (0.431)D:Children -0.261 -0.169 -0.121(0.255) (0.253) (0.241)D:Bank 0.549A 0.583*(0.292) (0.290)D:Forest 0.468 0.652*(0.325) (0.316)X:ChiBan 1.626**(0.562)X:ChiFor 0.758(0.603)X:GenBan 1.690**(0.590)X:GenFor 1.740**(0.660)X:AdmBan -1.774(1.306)X:AdmFor 0.177(1.136)X:MgrBan -1.316(1.107)X:MgrFor -1.413(0.886)X:CovBan 0.158*(0.063)X:CovFor 0.045(0.066)(Constant) 1.123* 0.826 0.553(0.463) (0.516) (0.528)R-Squared 0.715 0.726 0.766‘p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.213TABLE D.15Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Females ListedbIfi ModeratorC:Number 0.373*** O.376*** 0.383***(0.028) (0.028) (0.027)l:Gender 2.610*** 2.425*** 2.341***(0.263) (0.268) (0.256)D:Admin 1.497** 1.272* O.994A(0.496) (0.498) (0.521)D:Manager 0.182 -0.147 -0.265(0.382) (0.397) (0.431)D:Children 0.261 0.168 0.121(0.255) (0.253) (0.241)D:Bank 0.549k’ 0.583*(0.292) (0.290)D:Forest -0.468 0.652*(0.325) (0.316)X:ChiBan 1.626**(0.561)X:ChiFor -0.758(0.603)X:GenBan 1.689**(0.590)X:GenFor 1.738**(0.660)X:AdmBan 1.773(1.305)X:AdmFor -0.178(1.136)X:MgrBan 1.319(1.106)X:MgrFor 1.418(0.886)X:CovBan 0.158*(0.063)X:CovFor -0.045(0.066)(Constant) -1 .121 * -0.824 -0.549(0.463) (0.516) (0.528)R-Squared 0.580 0.596 0.655p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.214TABLE D.16Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Overlapping Expressive and Instrumental Tiesshift ModeratorC:Number 0.255*** 0.260*** 0.257***(0.026) (0.025) (0.025)l:Gender -0.379 -0.271 -0.374(0.239) (0.241) (0.242)D:Admin -0.367 -0.035 -0.194(0.451) (0.448) (0.493)D:Manager -0.118 0.303 0.133(0.348) (0.357) (0.408)D:Children -0.148 -0.088 -0.123(0.232) (0.228) (0.228)D:Bank 0.011 -0.015(0.263) (0.274)D:Forest 1.020** 1.121***(0.292) (0.299)X:ChiBan -0.414(0.531)X:ChiFor 0.405(0.570)X:GenBan -0.009(0.558)X:GenFor 0.334(0.624)X:AdmBan 0.554(1.235)X:AdmFor -0.087(1.075)X:MgrBan 0.696(1.046)X:MgrFor 1 .265(0.838)X:CovBan 0.057(0.060)X:CovFor 0.173**(0.063)(Constant) 0.687 -0.045 0.264(0.421) (0.465) (0.499)R-Squared 0.299 0.339 0.378A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.215TABLE D.17Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Overlapping Female Tieshiit ModeratorC:Number 0.090*** O.092*** 0.092***(0.016) (0.016) (0.016)l:Gender O.437** O.417** 0.373*(0.148) (0.153) (0.154)D:Admin 0.230 0.259 0.205(0.279) (0.285) (0.314)D:Manager 0.023 0.048 0.016(0.215) (0.228) (0.260)D:Children -0.060 -0.068 -0.102(0.144) (0.145) (0.145)D:Bank 0.150 0.167(0.167) (0.174)D:Forest 0.150 0.150(0.186) (0.190)X:ChiBan -0.488(0.338)X:ChiFor -0.145(0.363)X:GenBan -0.510(0.355)X:GenFor 0.111(0.397)X:AdmBan 0.667(0.787)X:AdmFor -0.370(0.685)X:MgrBan 0.596(0.667)X:MgrFor 0.529(0.534)X:CovBan 0.070A(0.038)X:CovFor 0.057(0.040)(Constant) -0.068 -0.186 -0.056(0.261) (0.296) (0.318)R-Squared 0.168 0.172 0.219A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.216TABLE D.18Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Seen Outside of WorkBase Shift ModeratorC:Number 0.554*** O.562*** O.565***(0.042) (0.043) (0.043)I:Gender 0.251 0.159 0.128(0.393) (0.405) (0.407)D:Admin -0.875 -0.795 -0.81 1(0.739) (0.754) (0.827)D:Manager -0.023 0.033 0.148(0.570) (0.601) (0.685)D:Children -0.492 -0.532 -0.560-(0.380) (0.383) (0.382)D:Bank 0.590 0.481(0.442) (0.460)D:Forest 0.484 0.426-(0.491) (0.502)X:ChiBan 1.126(0.891)X:ChiFor 1.754(0.957)X:GenBan -1.566k’(0.936)X:GenFor -1 .976k’(1.047)X:AdmBan 0.439(2.072)X:AdmFor -0.020(1.804)X:MgrBan 1.811(1.756)X:MgrFor 1.186(1.406)X:CovBan 0.015(0.100)X:CovFor 0.114(0.105)(Constant) -0.034 -0.423 -0.412(0.691) (0.781) (0.838)R-Squared 0.435 0.440 0.474“p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.217TABLE D.19Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Females Seen Outside of Workbiit ModeratorC:Number O.219*** O.225*** O.232***(0.030) (0.031) (0.030)l:Gender 1.958*** 1.817*** 1.765***(0.283) (0.291) (0.289)D:Admin 0.656 0.580 0.556(0.533) (0.541) (0.588)D:Manager 0.281 0.141 0.287(0.411) (0.432) (0.487)D:Children -0.168 -0.235 -0.252(0.274) (0.275) (0.272)D:Bank 0.578k’ 0.487(0.317) (0.327)D:Forest 0.003 -0.224(0.353) (0.357)X:ChiBan -0.155(0.634)X:ChiFor 0.056(0.681)X:GenBan -1.188”(0.666)X:GenFor 2.092**(0.744)X:AdmBan 0.254(1.474)X:AdmFor -0.622(1.283)X:MgrBan 1.579(1.249)X:MgrFor 0.529(1.000)X:CovBan 0.123”(0.071)X:CovFor 0.005(0.075)(Constant) 0.999* -1 .042” -1 .076(0.498) (0.561) (0.595)R-Squared 0.338 0.349 0.400“p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.218TABLE D.20Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Same Sitetiifi ModeratorC:Number O.619*** 0.604*** 0.596***(0.044) (0.043) (0.042)I:Gender 1.364** O.924* O.915*(0.407) (0.405) (0.404)D:Admin 2.O48* * -1 .724* 2.O54*(0.766) (0.753) (0.822)D:Manager 2.147*** 1.6O7** 1.972**(0.590) (0.601) (0.680)D:Children -0.073 0.140 0.126(0.394) (0.383) (0.379)D:Bank 1.655*** 1.75O***(0.441) (0.457)D:Forest 0.324 0.482(0.491) (0.499)X:ChiBan -0.440(0.885)X:ChiFor -0.552(0.951)X:GenBan -0.334(0.930)X:GenFor -0.092(1.040)X:AdmBan 1.831(2.059)X:AdmFor 3.093(1.792)X:MgrBan 0.712(1.745)X:MgrFor 1.596(1.397)X:CovBan 0.199*(0.100)X:CovFor 0.189A(0.105)(Constant) 2.510** 2.394** 2.847***(0.71 5) (0.780) (0.832)R-Squared 0.482 0.522 0.555A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.219TABLE D.21Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Different Site/Same Cityhifi ModeratorC:Number O.221*** O.216*** O.218***(0.040) (0.039) (0.039)l:Gender 0.810* 0.599 0.615(0.368) (0.371) (0.372)D:Admin 1.23O’ 0.736 1.271(0.693) (0.690) (0.757)D:Manager 1.422** 0.776 1.267*(0.534) (0.550) (0.627)D:Children -0.255 -0.367 -0.336(0.356) (0.351) (0.350)D:Bank 0.233 0.288(0.404) (0.421)D:Forest 1.418** 1.492**(0.499) (0.459)X:ChiBan 0.042(0.816)X:ChiFor 0.682(0.876)X:GenBan 1 .005(0.857)X:GenFor 0.048(0.958)X:AdmBan -1.880(1.897)X:AdmFor -2.459(1.652)X:MgrBan -0.946(1.608)X:MgrFor -2.065(1.288)X:CovBan -0.085(0.092)X:CovFor 0.253**(0.096)(Constant) -1 .309* -0.308 -0.958(0.647) (0.715) (0.767)R-Squared 0.117 0.223 0.273‘p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.220TABLE D.22Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Different Site/Different CityBase Shift ModeratorC:Number O.142*** O.164*** O.166***(0.031) (0.029) (0.029)l:Gender 0.426 0.131 0.124(0.288) (0.280) (0.282)D:Admin 0.735 O.862A 0.756(0.542) (0.521) (0.573)D:Manager 0.922* 0.953* O.921(0.418) (0.415) (0.474)D:Children 0.500A 0.367 0.318(0.279) (0.265) (0.265)D:Bank 1.678*** 1.726***(0.305) (0.318)D:Forest 1.067** 1.014**(0.339) (0.348)X:ChiBan -0.041(0.618)X:ChiFor -0.336(0.663)X:GenBan 1.325*(0.649)X:GenFor -0.050(0.725)X:AdmBan 0.658(1.436)X:AdmFor -0.425(1.250)X:MgrBan 0.497(1.217)X:MgrFor 0.316(0.974)X:CovBan 0.167*(0.070)X:CovFor 0.103(0.073)(Constant) -1.074 -1.957 1.800**(0.506) (0.539) (0.580)R-Squared 0.119 0.226 0.265A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.221TABLE D.23Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Individuals Listed at Same Functionhift ModeratorC:Number O.608*** 0.585*** O.581***(0.043) (0.042) (0.041)l:Gender O.944* O.725A -0.791 *(0.401) (0.401) (0.396)D:Admin 0.789 0.491 0.780(0.754) (0.745) (0.806)D:Manager 0.911 0.655 0.863(0.581) (0.594) (0.667)D:Children -0.180 -0.087 -0.027(0.388) (0.379) (0.372)D:Bank 1.588*** 1.780***(0.436) (0.448)D:Forest 1.556** 1.548**(0.486) (1 .489)D:ChiBan -0.715(0.868)D:ChiFor 0.173(0.932)X:GenBan 1.978*(0.912)X:GenFor -0.377(1.020)X:AdmBan 0.756(2.018)X:AdmFor -0.299(1.757)X:MgrBan 1.871(1.710)X:MgrFor 0.711(1.370)X:CovBan 0.296**(0.098)X:CovFor 0.247*(0.102)(Constant) -0.488 0.740 0.442(0.704) (0.772) (0.816)R-Squared 0.477 0.514 0.563A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.222TABLE D.24Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Different Functions ListedBase Shift ModeratorC:Number O.198*** O.212*** 0.209***(0.029) (0.028) (0.027)l:Gender 0.432 0.347 0.356(0.265) (0.266) (0.261)D:Admin -0.047 0.214 0.018(0.499) (0.495) (0.531)D:Manager -0.515 -0.249 -0.432(0.385) (0.394) (0.439)D:Children 0.075 0.043 -0.030(0.257) (0.251) (0.245)D:Bank 0.856** O.978**(0.290) (0.295)D:Forest 1.147** 1.277***(0.322) (0.322)X:ChiBan 0.066(0.572)X:ChiFor -0.038(0.614)X:GenBan 1.408*(0.601)X:GenFor 0.577(0.672)X:AdmBan 0.077(1.329)X:AdmFor 0.172(1.157)X:MgrBan -0.808(1.127)X:MgrFor 0.199(0.902)X:CovBan 0.056(0.064)X:CovFor 0.221**(0.067)(Constant) O.886 0.003 0.303(0.466) (0.512) (0.537)R-Squared 0.196 0.248 0.324‘p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.223TABLE D.25Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Hierarchical Rank of Individuals ModeratorC:Number O.O27** O.O25* O.O24*(0.009) (0.009) (0.010)l:Gender -0.116 -0.122 -0.128(0.087) (0.090) (0.091)D:Admin 0.289A -0.242 -0.293(0.164) (0.167) (0.185)D:Manager 0.358* * O.308* 0.357*(0.127) (0.133) (0.153)D:Children -0.086 -0.088 -0.106(0.085) (0.085) (0.085)D:Bank 0.109 0.157(0.098) (0.103)D:Forest 0.185A 0.226*(0.109) (0.112)X:ChiBan -0.227(0.199)X:ChiFor 0.013(0.214)X:GenBan -0.181(0.209)X:GenFor 0.091(0.234)X:AdmBan -0.732(0.463)X:AdmFor 0.020(0.403)X:MgrBan -0.496(0.392)X:MgrFor -0.114(0.314)X:CovBan -0.033(0.022)X:CovFor 0.014(0.023)(Constant) 0.272A 0.131 0.175(0.153) (0.173) (0.187)R-Squared 0.078 0.091 0.131A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.224TABLE D.26Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Hierarchical Range of Individuals ListedhLft ModeratorC:Number 0.005 0.009 0.009(0.006) (0.006) (0.006)I:Gender 0.120* 0.089 0.077(0.056) (0.056) (0.056)D:Admin -0.031 0.019 -0.004(0.106) (0.104) (0.113)D:Manager 0.046 0.090 0.070(0.082) (0.083) (0.094)D:Children 0.028 0.014 -0.001(0.055) (0.053) (0.052)D:Bank O.241*** 0.285***(0.061) (0.063)D:Forest 0.249*** 0.284A(0.068) (0.069)X:ChiBan -0.061(0.122)X:ChiFor -0.080(0.131)X:GenBan -0.092(0.128)X:GenFor 0.075(0.144)X:AdmBan -0.389(0.284)X:AdmFor 0.622*(0.247)X:MgrBan 0.484*(0.241)X:MgrFor -0.088(0.193)X:CovBan -0.004(0.014)X:CovFor 0.008(0.014)(Constant) 0.819*** 0.623*** 0.641***(0.099) (0.180) (0.115)R-Squared 0.026 0.110 0.174“p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.225TABLE D.27Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Supervisors Listedjjjft ModeratorC:Number 0.025** 0.026*** 0.025***(0.007) (0.007) (0.007)I:Gender 0.143* -0.151 * -0.1 27A(0.065) (0.068) (0.067)D:Admin -0.065 -0.056 -0.030(0.123) (0.126) (0.136)D:Manager 0.009 0.017 0.036(0.095) (0.100) (0.112)D:Children 0.004 0.000 0.010(0.063) (0.064) (0.063)D:Bank 0.057 0.024(0.074) (0.075)D:Forest 0.051 0.041(0.082) (0.082)X:ChiBan 0.233(0.146)X:ChiFor -0.018(0.157)X:GenBan 0.034(0.154)X:GenFor 0.048(0.172)X:AdmBan 0.111(0.340)X:AdmFor 0.548A(0.296)X:MgrBan 0.107(0.288)X:MgrFor -0.164(0.231)X:CovBan -0.006(0.016)X:CovFor 0.032”(0.017)(Constant) 0.161 0.120 0.102(0.115) (0.130) (0.137)R-Squared 0.073 0.076 0.160“p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.226TABLE D.28Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Number of Subordinates Listedj2j[t ModeratorC:Number O.165*** O.166*** O.167***(0.022) (0.022) (0.022)I:Gender -0.161 -0.288 -0.317(0.208) (0.213) (0.211)D:Admin 0.916* O.743” O.714(0.392) (0.395) (0.430)D:Manager 0.889** 0.643* 0.579(0.302) (0.315) (0.356)D:Children -0.035 -0.099 -0.098(0.202) (0.201) (0.198)D:Bank 0.348 0.373(0.232) (0.239)D:Forest -0.388 -0.316(0.258) (0.261)X:ChiBan 0.296(0.463)X:ChiFor 0.688(0.497)X:GenBan 0.186(0.486)X:GenFor 0.644(0.544)X:Adm Ban 1 .084(1.076)X:AdmFor -0.692(0.937)X:MgrBan 0.935(0.912)X:MgrFor 1.148(0.731)X:CovBan 0.030(0.052)X:CovFor 0.107A(0.056)(Constant) 0.768* -0.514 -0.46 1(0.366) (0.410) (0.435)R-Squared 0.228 0.251 0.313A p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.227TABLE D.29Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:DensityBase Shift ModeratorC:Number -0.005 -0.005 -0.005(0.003) (0.003) (0.003)l:Gender O.O81** 0.072* O.O75*(0.029) (0.030) (0.031)D:Admin -0.044 -0.037 -0.098(0.055) (0.057) (0.063)D:Manager -0.024 -0.012 -0.072(0.043) (0.045) (0.052)D:Children -0.037 -0.033 -0.031(0.028) (0.029) (0.029)D:Bank -0.034 -0.027(0.033) (0.035)D:Forest 0.010 0.021(0.037) (0.038)X:ChiBan -0.010(0.068)X:ChiFor 0.017(0.073)X:GenBan 0.022(0.071)X:GenFor -0.028(0.080)X:AdmBan -0.112(0.158)X:AdmFor 0.180(0.138)X:MgrBan -0.119(0.134)X:MgrFor 0.153(0.107)X:CovBan 0.000(0.008)X:CovFor -0.002(0.008)(Constant) 0.327* * * 0.322* * * 0.372* * *(0.052) (0.058) (0.064)R-Squared 0.054 0.060 0.090‘p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.228TABLE D.30Expressive Network Regression Model AnalysisBeta Coefficients and R-Squared Comparisons on Dependent Variable:Frequency of Contacthifi ModeratorC:Number -0.018 -0.016 -0.018(0.013) (0.013) (0.013)I:Gender O.727*** O.687*** O.724***(0.121) (0.124) (0.126)D:Admin -0.234 -0.126 -0.226(0.227) (0.230) (0.255)D:Manager O.444* -O.3O5 -0.441 *(0.175) (0.183) (0.211)D:Children -0.111 -0.090 -0.085(0.117) (0.117) (0.118)D:Bank -0.017 -0.033(0.135) (0.142)D:Forest 0.325* 0.379*(0.150) (0.155)X:ChiBan -0.210(0.275)X:ChiFor 0.176(0.295)X:GenBan 0.467(0.289)X:GenFor 0.582A(0.323)X:AdmBan 0.089(0.639)X:AdmFor 0.554(0.557)X:MgrBan 0.302(0.542)X:MgrFor 0.601(0.434)X:CovBan 0.025(0.031)X:CovFor 0.032(0.032)(Constant) 4.237*** 4.005*** 4.159***(0.212) (0.238) (0.258)R-Squared 0.166 0.186 0.220p < .10; * p < .05; ** p < .01; *** p < .001Standard errors are in parentheses.229


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items