annotatedbib

Relational Demography in the Workplace and Health: An Analysis of Gender and the Subordinate–Superordinate Role-Set*

SCOTT SCHIEMAN TARALYN MCMULLEN University of Toronto

Journal of Health and Social Behavior 2008, Vol 49 (September): 286–300

Using data from a 2005 national survey of working adults in the United States, we examine the effects of the gender composition of the superordinate–subordi- nate role-set on mental and physical health measures. Subordinates’ and super- ordinates’ genders are important determinants. Men who work in gender-mixed superordinate contexts (i.e., with one male and one female superior) report low- er levels of distress and physical symptoms than men who work with one male superior. Women who work with one male superior report less distress and few- er physical symptoms compared to women who work with one female superior or in gender-mixed superordinate contexts. With a few exceptions, these obser- vations generally hold net of occupation, job sector, and an array of work-re- lated conditions. We discuss the implications of these findings in light of pre- dictions derived from the similarity-attraction and role congruity theories. We also outline ways that theoretical development in relational demography can be refined by a more specific focus on the demographic characteristics—especial- ly gender—of the superordinate–subordinate role-set.

286

Diversity in the workgroup has emerged as a central issue of the modern workplace (Hodson 2002). We draw upon the “relational demogra- phy” literature to inform our analysis of the ef- fects of demographic composition on relation- al dynamics at work, especially the effects of similarities and differences in the gender char- acteristics of superordinates and subordinates (Tsui and O’Reilly 1989). Our rationale for linking these processes to health derives ex- plicitly from two sources. First, in his classic

paper on organizational demography, Pfeffer (1983) identified the potential health effects of relational demography, asserting that “the rela- tive proportions of [groups] condition the form and nature of social interaction and group processes that in turn affect workers’ psycho- logical well-being, attitudes, and even job per- formance” (p. 303). Here, we connect that no- tion to a core prediction of stress process theo- ry: Chronic stressors in roles can undermine well-being (Pearlin 1999). As we discuss be- low, research in relational demography alludes to these stress-related elements but has not tested stress process propositions with differ- ent health outcomes. Moreover, little is known about subordinates’ experiences of these pro- cesses within demographically diverse role- sets and in a broad cross-section of job sectors and occupations (Cortina et al. 2001; Perry, Kulik, and Zhou 1999; Tsui and O’Reilly 1989; Vecchio 1993; Vecchio and Bullis 2001).

* Grant R01 OH008141 (Scott Schieman, Principal Investigator) from the National Institute of Occupa- tional Safety and Health at the Centers for Disease Control supports this study. An earlier version of this manuscript was presented at the Tenth International Conference on Social Stress Research, Portsmouth, NH, 2006. Address correspondence to Scott Schieman, Department of Sociology, 725 Spadina Ave., University of Toronto, Toronto, ON M5S 2J4 Canada (e-mail: [email protected]).

THEORETICAL FRAMEWORK

Relational Demography in the Workplace

Relational demography involves the com- parative demographic characteristics of indi- viduals involved in interaction within dyads or groups at work (Tsui, Egan, and O’Reilly 1992). The increased shift toward team-orient- ed forms of organization across sectors under- scores the importance of relationships in the role-set (Vallas 2003). Among the many com- plexities of the workplace role-set, the vertical relations between supervisors and subordinates present fertile opportunities for stress (Kasl 1998). For example, Keashley and colleagues (1994) found that superordinates commit more acts of nonsexual, nonphysical abuse (57%) than coworkers (37%) or subordinates (5%). Likewise, a survey in the United Kingdom found that managers comprise 75 percent of all perpetrators of workplace “bullying” (Hoel, Cooper, and Faragher 2001). Superordinate– subordinate relations are important because they are frequent, often intense, and conse- quential interactions (Duffy and Ferrier 2003), and because “superiors are the primary referent for def ining performance expectations and standards for subordinates” (Tsui and O’Reilly 1989:405). While most research has focused on the effects of relational demography on task behaviors, evaluations, performance expecta- tions, and turnover intentions, we extend prior research by examining gender similarities and differences in the superordinate–subordinate role-set and their implications for subordi- nates’ health.

The Similarity and Dissimilarity Hypotheses

The demographic similarity or dissimilarity between superordinates and subordinates may influence individuals’ attitudes and behaviors and, in turn, individual-level outcomes (Perry et al. 1999). We blend ideas from similarity-at- traction, expectation states, and role congruity theories to identify two ways that gender may be relevant: (1) the extent to which the gender of superiors influences subordinates’ health; and (2) the ways that the gender of subordi- nates modifies that association. The similarity- attraction framework proposes that demo- graphic similarities increase attraction and convergence of attitudes, priorities, and expec- tations among workers. These processes, in turn, are believed to foster cohesion, decision- making, creative solutions and innovations,

and fewer conflicts among workers (O’Reilly, Williams, and Barsade 1998; Tsui and O’Reilly 1989). Collectively, these ideas provide a ratio- nale for the similarity hypothesis: Gender sim- ilarities between superordinates and subordi- nates should be associated with more favorable health among subordinates.

If demographic similarities have positive implications, it is plausible that demographic differences breed discontent. Therefore, the dissimilarity hypothesis predicts that gender dissimilarities between subordinates and supe- riors should be associated with poorer health among subordinates. Some scholars have ar- gued that workplace diversity undermines the identification with others that, in turn, dimin- ishes participation and organizational commit- ment (Chatman et al. 1998; Hodson 2002). Subordinates may be vulnerable to being the targets of the critical scrutiny and evaluation of dissimilar superiors (Duffy and Ferrier 2003). Dissimilarities associated with gender involve visible differences that have deep cognitive and social connections to stereotypes and preju- dice. Members of diverse work groups must of- ten confront opposing frames of reference that require them to manage their negative stereo- types about others (Pelled, Eisenhardt, and Xin 1999). Collectively, these processes may threaten workplace cohesion by increasing conflict with superordinates, role ambiguity, and job dissatisfaction (Tsui et al. 1992; Tsui and O’Reilly 1989).

Findings from prior studies generally favor the view that dissimilarities or demographic di- versity in the workplace are associated with deleterious personal and social outcomes (Hodson 2002; Vecchio and Bullis 2001). For example, empirical tests of the similarity-at- traction thesis document that demographic dis- similarity has negative effects on outcomes such as workers’ attitudes, job performance, and turnover (Crompton and Harris 1998; Jehn, Northcraft, and Neale 1999). To date, however, most theory and research has focused on the overall demographic composition of the workplace. While these studies provide a use- ful framework for the effects of demographic composition on processes and outcomes, the focus on overall gender composition of the workplace may misrepresent the effects of gen- der in specific components of the role-set.

RELATIONAL DEMOGRAPHY AND HEALTH 287

Variants of the Similarity and Dissimilarity Hypotheses

While hypotheses about the effects of simi- larities and dissimilarities in the subordi- nate–superordinate role-set represent a plausi- ble starting point, we argue that these views oversimplify or neglect the potential health im- plications of deeper (and sometimes subtler) psychological, sociological, and cultural processes about the roles of women and men in positions of power. Our skepticism regarding the limits of the similarity-attraction thesis is rooted in several other theoretical views that make predictions based on relational norms and collective representations of legitimacy: expectation states and role congruity theories. We draw upon ideas from these theories to in- form several elaborations of the similarity and dissimilarity hypotheses.1

With its focus on social status characteris- tics, expectation states theory seeks to explain why some individuals have greater access to participation, influence, and favorable ap- praisals in group dynamics (Correll and Ridgeway 2003). It underscores that the per- ceptions that individuals have of their superiors is potentially influential for a range of person- al and social outcomes, including well-being. Gender is among the most influential status characteristics because of its effects on perfor- mance expectations and evaluations, and per- ceptions of attributes that are salient in con- texts involving power and prestige dynamics (Driskell and Mullen 1990; Freese and Cohen 1973). Gender is a diffuse status characteristic because individuals with advantaged charac- teristics (i.e., “being male”) are often deemed as more competent, valued, appropriate for leadership roles, and legitimately in possession of power in hierarchical settings (Ridgeway 2001). Taken together, these ideas imply that male superiors should be especially advan- taged with respect to workplace participation, influence, and legitimacy; by contrast, female superiors should tend to be disadvantaged. In turn, working with male or female superiors likely has different implications for the health and well-being of subordinates.

Here, we integrate ideas from role congruity theory to elaborate further on the predictions of the similarity-attraction paradigm. According to Eagly and Karau (2002), “a potential for prejudice exists when social perceivers hold a stereotype about a social group that is incon- gruent with the attributes that are thought to be

required for success in certain classes of social roles” (p. 574). Research on gender and lead- ership identifies the incongruity that women encounter when they are in positions that re- quire them to simultaneously deploy “leader” and “feminine” qualities in social roles (Eagly and Karau 2002; Moore, Grunberg, and Greenberg 2005). This “double bind” tension emerges because leadership roles involve agential or task-related attributes while “the fe- male role” is stereotypically associated with in- terpersonal and cooperative traits that are in- compatible with leadership. This leader-female incongruity is associated with workers’ more unfavorable appraisals about women’s poten- tial for leadership, evaluations of women’s ac- tual leadership behavior, and women leaders’ level of trustworthiness (Eagly, Johannesen- Schmidt, and van Engen 2003; Scott and Brown 2006; Valentine, Godkin, and Turner 2002).

Ely (1994) has identified the negative stereo- types that contribute to the deleterious effects of role incongruity processes: “women are in- secure, over controlling, and unable to engage in team play; their relationships are therefore competitive and difficult” (p. 203). Likewise, Wacjman (1998) observes that senior women may experience problems when they “manage like men.” Evidence confirms these unfavor- able perceptions by documenting that men and women tend to view women managers as lack- ing the necessary attributes for competency in the higher echelons of the workplace (Brenner, Tomkiewicz, and Schein 1989; Eagly, Makhijani, and Klonsky 1992; Wajcman 1996). Several studies document that individu- als who work with female superiors encounter the most problematic relations and unfavorable personal and social outcomes (Valentine et al. 2002; Vecchio and Bullis 2001). Taken togeth- er, these role incongruent ideas predict that the health advantages of similarity in the female subordinate-female superior dyad may be at- tenuated; likewise, the health disadvantages of dissimilarity in the male subordinate-female superior dyad may be exacerbated. By contrast, role congruent dynamics and the greater legit- imacy in relational norms for male leaders pre- dict that the health disadvantages of dissimi- larity in the female subordinate-male superior dyad may be attenuated.

An alternative argument presents a more positive view of women’s leadership and its im- plications for subordinates. In response to the

288 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

double bind, for example, women leaders may adopt a transformational style of leadership that entails empowerment and mentoring of subordinates and provides a more communal, democratic, and team-focused orientation (Eagly et al. 2003; Valentine and Godkin 2000). These ideas suggest that both women and men subordinates who work with a female superior should report better health. Integrating ideas from the similarity hypothe- sis, however, implies that women subordinates should experience the most health advantages by working with a female superior. Ridgeway and Smith-Lovin (1999) observe that gender “is a background identity that modifies other identities that are often more salient in the set- ting than it is” (p. 193). Thus, similarities in gender may offset problems associated with dissimilarities or role incongruity. Senior women may represent a positive force for low- er status women by modifying the workplace culture in ways that reduce competition and ag- gressiveness, by acting as role models who provide support and inspiration for lower status women, and by fostering more “women-friend- ly” policies (Burke and McKeen 1996; Ely 1994).

Gender-Mixed Superordinate Contexts

While most theory and evidence provides reasons why working with male or female su- periors might influence personal and social outcomes, surprisingly little is known about working in gender-mixed superordinate con- texts—that is, having both a male superior and a female superior. From the dissimilarity view, we might expect negative implications for both women and men subordinates. However, that view ignores the social and cultural complexi- ties embedded in relational norms about gen- der, leadership, and power. Unfortunately, there is a dearth of theory and evidence about the implications of gender-mixed superordi- nate contexts. Most studies focus on gender- mixed workgroups or the entire workplace. To our knowledge, none examine levels of health among subordinates who specifically work with only one male and one female superior. We attempt to address this deficit by delineat- ing competing views about the potential posi- tive and negative implications of gender-mixed superordinate contexts.

On the positive side, the role incongruent thesis underscores the power of the communal stereotype: Women superiors are more likely to

provide supportive environments, especially for subordinate women. That predicts favorable health status for women subordinates in gen- der-mixed scenarios. Moreover, the similarity hypothesis predicts that gender similarity in the female subordinate-female superior context should trump the effects of work status dissim- ilarities and offset any negative implications of the additional presence of a male superior. However, unfavorable countercurrents may prevail if women in gender-mixed scenarios “manage like men” in ways that disappoint subordinate women (Wajcman 1998). The bulk of relevant theory and evidence provides a compelling case for the negative implications of gender-mixed superordinate contexts, with most of it directed toward superordinate women. For example, Wharton and Baron (1991) describe an “intragender competition effect” in which women struggle to attain sup- portive benefits from other women but also en- counter competitive undercurrents. Likewise, in what has been called the “queen bee syn- drome,” female superiors in male-dominated contexts receive psychosocial rewards from higher status men when they denigrate other women (Cooper 1997; Staines and Tavris 1974). Here, it is plausible that many female superiors experience gender-mixed superordi- nate contexts as male-dominated. Some schol- ars contend that female superiors tend to be highly competitive with other women, and, even more insidiously, male superiors in these contexts may be more likely to pit women against each other (Ely 1994). Likewise, as Kanter (1977) observed, men may instigate tensions between superordinate and subordi- nate women by fostering and exaggerating competitive relations. Collectively, these ideas predict that women in gender-mixed superordi- nate contexts should experience the most unfa- vorable levels of health. Unfortunately, there is little prior theory or evidence to guide specula- tion about how gender-mixed superordinate contexts will influence subordinate men’s health. Therefore, while it seems plausible to suspect that men in gender-mixed scenarios will benefit, we do not attempt to specify the direction of that effect in advance.

Taking Other Potential Explanations into Account

The final section of our analyses examines the possibility that relational demography ef- fects are due to occupation and work condi-

RELATIONAL DEMOGRAPHY AND HEALTH 289

tions. In some instances, such as occupation, any observed effects are more likely due to confounding effects. For example, female su- periors may tend to cluster in lower status oc- cupations that, in turn, are more strongly asso- ciated with poorer health outcomes (Reskin and Bielby 2005; Smith 2002). In other cases there are clearer theoretical reasons for sus- pecting possible explanatory mechanisms. As we noted in our rationale for the hypotheses, dissimilarities and role incongruent conditions may correspond with unfavorable work condi- tions, discordant relations with superordinates, and job dissatisfaction. Specifically, incon- gruity between leader-female roles purported- ly increases workers’ negative appraisals of women leaders that should manifest in expo- sure to more unfavorable work conditions, poorer relationship quality, and greater dissat- isfaction. Under such conditions, status and role incongruity would likely be visible and, in turn, fuel subordinates’ negative attitudes and behaviors. Collectively, these ideas provide the rationale for our analyses of whether occupa- tion, work conditions, relationship quality with superordinates, and job dissatisfaction con- tribute to any disparities in health across su- perordinate–subordinate gender comparisons.

METHODS

Sample

The data derive from telephone interviews with 1,800 adults in the 50 U.S. states from February through August of 2005.2 Eligible participants were 18 years of age or older and participating in the paid labor force at the time of interview. Interviews were conducted in English, so participants had to be sufficiently fluent in order to complete the interview. We were able to successfully interview 70.8 per- cent of eligible respondents. The sample char- acteristics are similar to those of working adults in other national data sets such as the 2005 American Community Survey. Using American Community Survey data, we weight- ed analyses to achieve conformance with the population in terms of sex, age, race, marital status, and occupation. For the purposes of our analyses, we examine data from a subset of 1,537 workers who reported “yes” to the fol- lowing question: “In your current job, do you have a supervisor or manager? That is, some- one who manages, supervises, directs, or con- trols your work?” We included cases only if they had one male superior, one female superi-

or, or one male and one female superior (“gen- der-mixed”). We excluded individuals if they had two superiors of the same gender, more than two superordinates, or missing informa- tion on gender and superiors. These exclusions yielded 1,095 cases for analyses (444 men and 651 women). Of those 1,095 individuals, 521 have one male superior, 409 have one female superior, and 165 have one male and one fe- male superior.3

Measures

Psychological distress. We measured dis- tress by asking participants the number of days in the past seven days that they “felt that every- thing was an effort,” “felt sad,” “had trouble getting to sleep or staying asleep,” “had trouble keeping your mind on what you were doing,” “couldn’t get going,” “were unable to shake the blues,” “worried a lot about little things,” and “felt anxious or tense.” We averaged the items to create the distress index (� = .85).

Physical symptoms. We asked participants the number of days in the past seven days that they had “headaches,” “stomach pain or prob- lems like indigestion or heartburn,” “chest pain or rapid heart beat,” “neck or back pain,” “mus- cle aches, soreness, or stiffness,” and “felt tired or run down.” We averaged the items to create the physical symptoms index (� = .70).

Gender of superiors. We contrasted individ- uals who work with “one male superior” (i.e., this is the omitted reference group in regres- sion models) with those who work with “one female superior” and “gender-mixed superi- ors.” Each of these three groups was dummy- coded as 0 = no, 1 = yes.

Occupation. To assess occupation, we asked participants about the job title of the “main job at which you worked last week.” This question refers to the main place of employment; that is, the one at which participants spend the most time. We also asked about some of the main duties in order to more accurately code re- sponses. Using open-ended information, we coded responses into five categories in accor- dance with the Bureau of Labor Statistics codes. These include: “professional” (manage- rial and professional specialty occupations), “administrative” (technical, sales, and admin- istrative support occupations), “service” (ser- vice occupations), “craft” (precision produc- tion, craft, and repair occupations), and “labor” (operators or laborers). In regression analyses,

290 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

we use the “professional” category as the omit- ted reference group.

Job sector. We assessed participants’ job sector by contrasting (in regression analyses) the modal category of “private for-profit com- pany” with “government,” “non-profit organi- zation including tax-exempt or charitable orga- nizations,” and “self-employed/family busi- ness.”

Work hours. Work hours are measured as the number of hours worked in a typical week.

Job authority. We used four items to assess job authority: (1) “Do you influence or set the rate of pay received by others?” (2) “Do you have the authority to hire or fire others?” (3) “Do you supervise or manage anyone as part of your job?” and, if they reported “yes” to the last question, (4) “Do any of those individuals supervise or manage others?” We coded “no” responses as 0 and “yes” responses as 1. To create the index, we summed responses such that higher scores indicate more authority; items are similar to those in other studies (Elliott and Smith 2001).

Nonroutine work. To measure nonroutine work, we asked five items that blend themes associated with decision-making, creative problem-solving and learning opportunities, and engaging (not boring) work: (1) “How of- ten do you make decisions on what needs to be done?” (2) “How often do you have the chance to solve problems?” (3) “How often do you have the chance to learn new things?” (4) “How often does time feel like it is dragging at work?” and (5) “How often do you do the same things over and over again?” (4 and 5 are re- verse-coded). Response choices are “never” (coded 1), “rarely” (2), “sometimes” (3), and “frequently” (4). We averaged the items; high- er scores indicate higher levels of nonroutine work (� = .57).

Demanding work. We asked about demands in the workplace: “In the past 30 days, has any- one at work made too many demands on you?” Participants were able to choose any source (e.g., “supervisor, someone you supervise, cus- tomer/client, coworker, or someone else at work”) and then describe the frequency: “rarely” (coded 1), “sometimes” (2), and “fre- quently” (3). We coded individuals who report- ed no demands as 0. We also asked, “How of- ten do the demands of your job exceed those doable in an 8-hour workday?” Response choices are “never” (coded 0), “rarely” (1), “sometimes” (2), and “frequently” (3). We

standardized and averaged items to create the index.

Noxious work. We used three items to assess subordinates’ level of noxious workplace envi- ronments: “How often is your workplace .|.|.” (1) “noisy,” (2) “dirty or dusty,” and (3) “dan- gerous; that is you are at risk of illness or in- jury because of the work.” Response choices are never (coded 0), rarely (1), sometimes (2), and frequently (3). We averaged the items such that higher scores represent a higher level of exposure to noxious work (� = .63).

Job insecurity. One question asks partici- pants, “In the next 2 years, how likely is it that you will lose your job or be laid off?” Response choices are “not at all likely” (coded 0), “somewhat likely” (1), and “very likely” (2).

Personal income. Income is assessed with the question, “For the complete year of 2004, what was your total personal income, including income from all of your paid jobs, before tax- es?”

Job tenure. One item asks participants, “How many years have you worked at your cur- rent job?” Response choices are coded in years.

Superordinate conflict. To assess subordi- nates’ experiences of conflict with superiors in the past 30 days, participants were asked the extent that superiors (1) “treated you unfairly,” (2) “got annoyed or angry with you,” (3) “teased or nagged you,” (4) “gossiped or talked about you behind your back,” (5) “blamed or criticized you for something that wasn’t your fault,” (6) “gave you unclear directions about work you need to do,” and (7) “did not do the work that needed to be done or did it in a slop- py or incompetent way.” Response choices are “never” (coded 0), “rarely” (1), “sometimes” (2), and “frequently” (3). We averaged the items to create the index (� = .71).

Superordinate support. To assess subordi- nates’ experiences of support from superiors in the past 30 days, participants were asked the extent that superiors did the following: (1) “lis- tened to your ideas or opinions,” (2) “thanked you for the work you do,” (3) “gave you posi- tive feedback, guidance, or advice,” and (4) “said or did something that made you feel pride in your work.” Response choices are “never” (coded 0), “rarely” (1), “sometimes” (2), and “frequently” (3). We averaged these items to create the superordinate support index (� = .79).

RELATIONAL DEMOGRAPHY AND HEALTH 291

Job dissatisfaction. One item asks partici- pants, “How satisf ied do you feel with your job?” Response choices are coded such that higher scores indicate higher levels of job dissatisfaction: “very much satisf ied” (cod- ed 1), “quite a bit” (2), “somewhat” (3), and “not at all” (4).

Basic control measures. All analyses con- trol for age, race, marital status and educa- tion (but are excluded from the regression ta- bles to conserve space). Age is measured in years. We contrast non-Hispanic white with African Americans and those of “other” race-ethnic background. We contrast those who are currently married (coded 1) with all others for our measure of marital status (cod- ed 0). We measure education as less than high school (coded 1), high school gradu- ate/GED (2), specialized (vocational) train- ing (3), some college but no degree earned (4), associate’s degree (2-year) (5), college

graduate (B.A. or B.S.) (6), and post gradu- ate with advanced degree (e.g., M.A., Ph.D.) (coded 7). Table 1 shows the descriptive sta- tistics for all of the measures.

Plan of Analyses

We use ordinary least squares (OLS) re- gression analyses to assess our hypothesized focal associations. We conducted analyses separately for men (Table 2) and women (Table 3) because of expectations about gen- der differences. In separate analyses (not shown), we pooled the sample and tested for gender interactions (i.e., gender × female su- perordinate). Coeff icients shown in bold type in Tables 2 and 3 indicate that we ob- served statistically signif icant gender inter- action effects (more details later). In the two tables that present regression results, model 1 regresses the focal dependent variable— psychological distress or physical symp-

292 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

TABLE 1. Descriptive Statistics for Study Variables

Men Women Total (N = 444) (N = 651) (N = 1,095)

Focal Dependent Measures —Psychological Distress 1.375*** 1.964 1.725 —Physical Symptoms 1.376*** 1.951 1.718 Focal Comparisons —Male Superordinate .709*** .316 .475 —Female Superordinate .175*** .508 .373 —Gender-Mixed Superordinates .114** .175 .150 Focal Control Measures —Professional Occupation .268 .305 .290 —Administrative Occupation .274*** .454 .381 —Service Occupation .139 .178 .162 —Craft Occupation .141*** .012 .064 —Labor Occupation .175*** .049 .100 —For-Profit Sector .664* .588 .619 —Government Sector .211 .238 .227 —Non-Profit Sector .076* .124 .105 —Self-Employed/Family Business .047 .049 .048 —Job Tenure 6.060 5.800 5.905 —Work Hours 45.916*** 39.550 42.132 —Job Authority 1.011*** .715 .835 —Nonroutine Work 2.912 2.852 2.876 —Demanding Work .018 –.009 .002 —Noxiousness Work 2.558*** 2.334 2.425 —Job Insecurity .265 .216 .236 —Personal Income 54.248*** 37.105 44.056 —Superordinate Conflict 1.819 1.699 1.748 —Superordinate Support 5.608 5.682 5.652 —Job Dissatisfaction 1.961 1.920 1.936 Basic Control Measures —Age 43.065 43.931 43.580 —Married .612** .522 .558 —White .772** .700 .729 —African American .139 .184 .166 —Education 5.263 5.299 5.284

Note: * p < .05; ** p < .01; *** p < .001 (mean difference between men and women; two-tailed test).

toms—on the focal comparisons between workers with a male superior and those with a female superior or with gender-mixed su- perordinates. In subsequent models, we ad- just for occupation, job sector, and work con- ditions (model 2) and the quality of relation- ships with superordinates and job dissatis- faction (model 3).

RESULTS

Findings among Men

The f irst three columns in Table 2 show the results for models with distress as the fo- cal dependent variable. Model 1 indicates that men with a female superior and those with a male superior do not differ in levels of

RELATIONAL DEMOGRAPHY AND HEALTH 293

TABLE 2. Distress or Physical Symptoms Regressed on Superordinate Contrasts and Focal Controls (Men Only; N = 444)

Psychological Distress Physical Symptoms

Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

Focal Comparisons —Female Superordinatea .050 –.013 –.143 –.037 –.026 –.114 — (.251) (.239) (.220) (.188) (.170) (.167) —Gender-Mixed Superordinatesa –.462* –.471* –.456* –.471* –.500** –.493** — (.222) (.237) (.225) (.185) (.186) (.173) Focal Control Measures —Administrativeb –.137 –.125 –.160 –.147 — (.228) (.208) (.191) (.179) —Serviceb –.360 –.230 –.588* –.496* — (.271) (.249) (.228) (.209) —Craftb –.136 –.182 –.323 –.353 — (.341) (.303) (.287) (.262) —Laborb –.078 .025 –.627* –.543* — (.301) (.282) (.251) (.237) —Governmentc –.313* –.232 –.132 –.076 — (.154) (.159) (.140) (.133) —Non-Profitc .019 .103 –.012 .043 — (.272) (.261) (.217) (.224) —Self-Employed/Family Businessc .011 .022 .034 .031 — (.264) (.291) (.321) (.306) —Job Tenure –.017 –.015 –.033* –.033* — (.019) (.018) (.017) (.016) —Work Hours –.010 –.008 –.002 –.001 — (.006) (.005) (.006) (.006) —Job Authority –.021 –.048 .022 .005 — (.060) (.056) (.057) (.055) —Nonroutine Work –.742*** –.311 –.608*** –.368* — (.146) (.170) (.123) (.147) —Demanding Work .312* .177 .157 .059 — (.139) (.123) (.104) (.099) —Noxiousness Work .097 –.020 .328*** .242** — (.081) (.077) (.076) (.077) —Job Insecurity .182 .039 .139 .053 — (.169) (.134) (.118) (.111) —Personal Income .002 .002 .001 .001 — (.003) (.003) (.002) (.002) —Superordinate Conflict .130** .099** — (.044) (.035) —Superordinate Support .004 .016 — (.021) (.017) —Job Dissatisfaction .439** .262* — (.164) (.127) Constant 1.467 3.982 1.904 1.563 2.969 1.698 R-Square .059 .165 .280 .046 .195 .255

* p < .05; ** p < .01; *** p < .001 (two-tailed test). a Compared to male superordinate. b Compared to professional occupations. c Compared to for-profit sector. Note: Unstandardized regression coefficients, with standard errors in parentheses. Coefficients in bold indicate a statis- tically significant difference between women and men.

distress. However, men who have gender- mixed superordinates report significantly less distress than men with a male superior. As shown in model 2, individuals who work in the for-profit sector report more distress than those in the government sector. Nonroutine work is associated negatively with distress, while de- manding work is associated positively with dis- tress. In model 3, conflict with superiors and job dissatisfaction are associated positively with distress. These adjustments, however, fail to account for the differences among superor- dinate comparisons observed in model 1. Finally, although not part of our focal associa- tions, several other patterns deserve brief men- tion. Men in the government sector report less distress because of their lower levels of inter- personal conflict compared to men in the pri- vate sector. Likewise, nonroutine work is asso- ciated with less distress because it also tends to be associated with less job dissatisfaction. Finally, demanding work is associated with more distress because of its positive associa- tion with job dissatisfaction and conflict with superiors.

The last three columns in Table 2 show the results for models with physical symptoms as the focal dependent variable. Model 1 indicates that men with a female superior and those with a male superior do not differ in levels of phys- ical symptoms. By contrast, however, men with gender-mixed superordinates report fewer physical symptoms. As shown in model 2, men in professional occupations report more physi- cal symptoms than men in service and labor jobs. Nonroutine work and job tenure are asso- ciated negatively with physical symptoms; noxious work is associated positively with symptoms. Separate analyses (not shown) indi- cate that nonroutine and noxious work function as suppressors of the differences between pro- fessionals and these occupation groups. Specifically, the differences in symptoms be- tween professionals versus service and labor occupations do not appear until we account for professionals’ higher levels of nonroutine work and lower levels of noxious work. In model 3, conflict with superiors and job dissatisfaction are associated positively with physical symp- toms. Taken together, however, these adjust- ments have little influence on the comparisons established in model 1.

In sum, men with one male superior do not differ from those with one female superior in levels of distress or physical symptoms. By

contrast, men who work with one male and one female superior (the “gender-mixed superordi- nate” scenario) report lower levels of distress and physical symptoms than men who work with one male superior—net of occupation, job sector, and an array of work-related conditions.

Findings among Women

As shown in model 1 of Table 3, women with a female superior and those in gender-mixed superordinate contexts report more distress compared to women who work with a male su- perior. In model 2, job tenure and nonroutine work are associated negatively with distress; demanding work and job insecurity are associ- ated positively with distress. Separate analyses (not shown) indicate that women with a female superior and those with gender-mixed superor- dinates report shorter tenures than women with a male superior. This contributes to the dispar- ities in distress across superordinate compar- isons shown in model 1 and reduces the gen- der-mixed coefficient to nonsignificance. In model 3, conflict with superiors is associated positively with distress; however, conflict does not vary across superordinate groups. Thus, its inclusion in model 3 has little influence on the differences that were observed in model 2. Also, among the peripheral findings, it is note- worthy that women with longer job tenures and nonroutine work report less distress because they experience less conflict with superiors.

The last three columns in Table 3 show the results for physical symptoms. Model 1 indi- cates that, compared to those who work with a male superior, women with one female superi- or or in gender-mixed superordinate contexts report more physical symptoms. In model 2, job tenure is associated negatively with physi- cal symptoms; demanding work, noxious work, and job insecurity are associated posi- tively with physical symptoms. Relative to women with a male superior, the shorter tenure among women with female superiors or in gen- der-mixed superordinate contexts contributes to their higher levels of physical symptoms. Likewise, in model 3, conflict with superiors is associated positively with physical symptoms. Taken together, however, these adjustments have little additional influence on the focal comparisons established in model 2. As we ob- served for distress, women with longer job tenures report fewer physical symptoms be- cause they experience less conflict with supe- riors.4

294 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Assessing Gender Differences

In additional analyses (not shown), we pooled women and men into one sample and tested for statistically significant gender inter- actions: gender × female subordinate and gen- der × gender-mixed subordinates. We identify statistically significant differences in bold font

in Tables 2 and 3. Specifically, we found a sig- nificant gender × gender-mixed superordinates coefficient, which indicates that the difference between women and men in gender-mixed su- perordinate contexts is significant for both dis- tress and physical symptoms. To reiterate, compared to those who work with a male su- perior, men with gender-mixed superiors re-

RELATIONAL DEMOGRAPHY AND HEALTH 295

TABLE 3. Distress or Physical Symptoms Regressed on Superordinate Contrasts and Focal Controls (Women Only; N = 651)

Psychological Distress Physical Symptoms

Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

Focal Comparisons — Female Superordinatea .388* .319* .336* .424** .390** .404** — (.157) (.148) (.146) (.139) (.138) (.138) — Gender-Mixed Superordinatesa .470* .397 .370 .373* .285 .259 — (.221) (.202) (.191) (.180) (.176) (.175) Focal Control Measures — Administrativeb .075 .109 .010 .035 — (.190) (.186) (.142) (.140) —Serviceb .397 .441 .345 .378 — (.233) (.229) (.212) (.208) —Craftb –.537 –.395 –.907 –.813 — (.711) (.672) (.508) (.490) —Laborb –.703 –.544 –.102 .016 — (.402) (.389) (.403) (.406) —Governmentc .004 .093 .090 .161 — (.175) (.177) (.145) (.142) — Non-Profitc –.097 –.084 –.060 –.049 — (.237) (.232) (.173) (.164) —Self-Employed/Family Businessc .261 .301 .271 .307 — (.258) (.247) (.352) (.352) —Job Tenure –.051** –.038 –.036* –.026 — (.020) (.019) (.018) (.018) —Work Hours .008 .007 .004 .004 — (.006) (.006) (.006) (.006) —Job Authority .052 .053 .033 .034 — (.065) (.064) (.055) (.055) —Nonroutine Work –.412* –.197 –.192 –.043 — (.161) (.191) (.139) (.141) —Demanding Work .598*** .466*** .366*** .257** — (.098) (.105) (.081) (.082) —Noxiousness Work .127 .115 .247* .239* — (.109) (.109) (.099) (.097) —Job Insecurity .473** .437** .406* .379* — (.173) (.167) (.160) (.157) —Personal Income .000 .000 .001 .001 — (.003) (.003) (.003) (.003) —Superordinate Conflict .095** .085** — (.033) (.026) —Superordinate Support –.022 –.016 — (.021) (.016) —Job Dissatisfaction .174 .105 — (.101) (.078) Constant 1.848 2.604 1.453 1.803 1.662 .843 R-Square .065 .209 .243 .065 .177 .208

* p < .05; ** p < .01; *** p < .001 (two-tailed test). a Compared to male superordinate. b Compared to professional occupations. c Compared to for-profit sector. Note: Unstandardized regression coefficients, with standard errors in parentheses. Coefficients in bold indicate a statis- tically significant difference between women and men.

port less distress and fewer physical symptoms. By contrast, women in gender-mixed superor- dinate contexts report higher levels of distress and more symptoms. We also found that the gender × female subordinate interaction is sta- tistically signif icant, but only for physical symptoms.5

DISCUSSION

In our analysis of relational demography in the workplace and health, two main contribu- tions emerge: (1) We document the relevance of superordinates’ gender for levels of psycho- logical distress and physical symptoms among subordinates; and (2) We describe the ways that subordinates’ gender functions as an effect modifier. First and foremost, our observations underscore the importance of considering the gender of both subordinates and superordi- nates. We discuss each set of findings below, describing their connections to our hypotheses and their implications for the similarity and dissimilarity hypotheses and the variants of those views. We also outline directions for fu- ture investigations that would further elaborate on our findings.

According to the similarity hypothesis, gen- der similarities among subordinates and their superiors should be associated with more fa- vorable levels of health among subordinates. By contrast, the dissimilarity hypothesis pre- dicts poorer health among subordinates with different genders than their superiors. Our ob- servations do not provide clear support for ei- ther view. For example, we found that men with a female superior share similar levels of dis- tress and physical symptoms as men with a male superior. Likewise, we also found that women with a female superior reported more distress and physical symptoms than women with a male superior. Collectively, these find- ings challenge both the similarity and dissimi- larity hypotheses and underscore the need for other theoretical ideas that might illuminate the patterns.

We draw upon the ideas of expectation states and role congruity theories to guide our elabo- ration and refinement of the similarity-dissim- ilarity perspective. Based on role congruent dynamics and the claims of legitimacy in rela- tional norms for male leaders, we predicted that the health disadvantages of dissimilarity in the female subordinate-male superior dyad would be attenuated. Our observations are con- sistent with this notion. The role incongruent

thesis also predicted that the health advantages of similarity in the female subordinate-female superior dyad should be attenuated. The find- ings seem to corroborate that claim: Women with a female superior report more distress and physical symptoms. Role congruity theory pro- vides insights that might help explain that pat- tern. In the “double bind,” women superiors en- counter conflicting roles that expect simultane- ous deployment of leader and feminine attrib- utes (Eagly and Karau 2002). Evidence shows that women leaders are often disadvantaged in more masculine role scenarios; that is, the greater the role incongruence the more unfa- vorably women leaders are rated on attributes such as effectiveness or competence (Eagly, Karau, and Makhijani 1995). Moreover, Ridgeway (2001) has articulated the tension between expectations surrounding traditional gender roles and exercising authority:

.|.|. when women assert authority over oth- ers they violate the essential hierarchical el- ement of gender status beliefs. This viola- tion may provoke negative reactions and re- sistance to their efforts. (P. 648)

Women superordinates may be aware of these incongruent expectations and compensate by acting more aggressively towards female sub- ordinates (Ely 1994). Collectively, the vast bulk of existing theory yields the controversial notion that working under female superordi- nates is especially disadvantageous for women. Much has been written about successful women shedding their feminine identity in or- der to navigate the challenges of higher status positions. Many scholars have speculated that these processes often represent signif icant sources of stress for subordinate women. Moreover, the expectations of social support may be especially relevant here. The similari- ty-attraction paradigm posits that the shared gender of female subordinates and superordi- nates should trump the effects of work status dissimilarities or deviations from relational norms (i.e., “leader = male”). Despite role in- congruities, the communal stereotype implies that women superordinates should provide es- pecially supportive environments for subordi- nate women. However, countercurrents may prevail if women “manage like men” in ways that disappoint subordinate women. Even more problematic is the notion that junior women are less likely to respect senior women because— as expectation states and role incongruity the- ories suggest—women are viewed less favor-

296 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

ably and with less legitimacy when they hold leadership positions. Although data limitations constrain our ability to document these dy- namics, the next step is to establish whether or not these processes—including perceived le- gitimacy, intragender competition, and expect- ed versus received support—contribute to stress and unfavorable health outcomes among women with female superiors.

While the comparisons of individuals work- ing with only one male or female superior are provocative, we also found complex patterns among those working with one male and one female (the “gender-mixed” superordinate context). Male subordinates in this context re- ported lower levels of distress and physical symptoms than their male counterparts with a male superior, net of occupation, job sector, and an array of work-related conditions. By contrast, female subordinates in gender-mixed superordinate contexts reported higher levels of distress and physical symptoms than women with a male superior. These patterns under- score the inadequacies of the similarity-dis- similarity thesis. Unexpectedly, however, we found that the inclusion of job tenure con- tributes to these differences. Women in gender- mixed superordinate contexts tend to report shorter job tenures, and shorter tenures are as- sociated with more distress and physical symp- toms. This finding undermines claims about more complex influences of social and cultur- al complexities of gender in leadership and power dynamics.

As we have noted, most theory and evidence on this topic focuses on overall demographic dissimilarities or diversity in the workplace. The few studies that have focused on subordi- nate–superordinate relations usually examine dyads in specific workplace organizations and have primarily assessed outcomes like work- ers’ attitudes, job performance, and turnover. The general conclusion of that research is that dissimilarities or demographic diversity in the workplace are associated with deleterious per- sonal and social outcomes. These ideas are consistent with the similarity-attraction thesis: Similarities are preferred among workers while dissimilarities breed discontent. Our findings challenge those ideas as oversimplifying the complexity of the role-set. The dynamics in subordinate–superordinate relationships pro- vide a unique set of insights into workers’ well- being. Yet, our research is limited: We do not have information about subordinates’ evalua-

tions and appraisals of both male and female superiors. Future research should assess subor- dinates’ perceptions of competence, legitima- cy, conflict, support, and satisfaction of each superior in gender-mixed contexts. Moreover, it would be useful to know more about the quality of the relationships between male and female superiors (from subordinates’ perspec- tive) and the relative position of each in the sta- tus hierarchy at work.

Several other limitations deserve brief men- tion. Although we do not explicitly discuss causality, our analyses are framed around an assumption that the gender of the superordi- nate–subordinate role-set influences health. It is plausible that causality runs in the opposite direction: Individuals with poorer health could be more likely to select into subordinate posi- tions. In our view, this direction is less tenable. Given our observations, selection bias would also result in individuals in poorer health being more likely to select into subordinate positions within female or gender-mixed superordinate contexts (since these conditions are associated with the poorest health among women). We cannot think of a compelling reason why indi- viduals in poor health would be less likely to select into subordinate work roles within men- only superordinate contexts. By contrast, there are clearer theoretical and empirical arguments for social-structural influences on health rather than the reverse. In addition, supplemental analyses found no health differences between individuals with and without superordinates; likewise, health differences between those who have subordinates and those who do not are in- signif icant. These patterns concur with Mirowsky and Ross’s (2003) observation that job authority is unrelated to health. Nonetheless, longitudinal data could more ac- curately determine the causal mechanisms in our focal associations.

Selection may account for the finding that women with a male superior have better health than those with a female superior. Those with more distress and physical symptoms may se- lect into less-challenging positions. Structural arrangements that have tended to reserve better jobs for men may imply that jobs with male su- pervisors also tend to have other favorable con- ditions (e.g., prestige, pay). The healthy may be more able to strive for those positions, while someone in ill health may find it difficult to search for or to be hired in such a job. However, it is more puzzling why women in poorer

RELATIONAL DEMOGRAPHY AND HEALTH 297

health would select into gender-mixed supervi- sory conditions. This is an area of research that deserves much more detailed attention.

CONCLUSION

In sum, sociologists of mental health have long underscored the relevance of the quality of role conditions for stress and health (Pearlin 1999). It seems reasonable to extend that tradi- tion to study the relevance of relational de- mography in the workplace for individuals’ well-being. In the present work, we offer pre- liminary steps toward that effort, and we dis- cover that the gender composition of the su- perordinate–subordinate role-set matters dif- ferently for the health and well-being of women and men subordinates.

NOTES

1. We wish to emphasize here that we are syn- thesizing and applying ideas from expecta- tion states and role congruity theories to in- form the multiple ways that gender of sub- ordinate–superordinate role-sets might in- fluence health, beyond those predicted by the similarity-attraction thesis. It is impor- tant to underscore that we are not explicitly testing the numerous processes or mecha- nisms embedded in these theories. That ef- fort would require data well beyond those available to us. Nonetheless, we outline ways that the concepts and evidence associ- ated with expectation states and role con- gruity theories yield insights for our analyt- ical framework.

2. To obtain the sample, we used a list-assist- ed random digit dialing selection drawn proportionally from all 50 states from GENESYS Sampling Systems. The sam- pling approach employed the list + 1 method, which tends to yield a higher pro- portion of productive numbers (Lepkowski 1988). List-assisted random digit dialing is widely accepted now by most social survey research organizations as a cost-effective al- ternative to the pure random digit dialing methods originally developed by Waksberg (1978). List-assisted random digit dialing increases the probability of encountering residential numbers while minimizing the biases often associated with non-traditional random digit dialing techniques. For our study, GENESYS generated a sample from 50 states that was drawn in proportion to the distribution of households. The total sample

was based on the following four criteria: (1) telephone numbers associated with residen- tial households, (2) households agreeing to answer the screening questions, (3) success- fully screened households that have one or more adult members who are currently working, and (4) eligible households with a subsampled adult who agreed to participate in the interview.

3. In the sample of 1,537, 8 percent of those who reported having a superior had three superiors, and 6 percent reported more than three superiors. We decided to focus exclu- sively on individuals with one superior or gender-mixed superiors (one of each) in or- der to minimize the variation embedded in more complicated subordinate–subordinate role-sets. However, later we report compar- isons between gender-mixed contexts and those with two superiors of the same gender.

4. Some readers may wonder about women and men who work with two same-gender superiors. Separate analyses (not shown) in- dicate that women with two female or two male superiors do not differ significantly from women who work with one male su- perior in distress and physical symptoms. Likewise, men with two male superiors do not differ significantly from males with one male superior in distress and physical symp- toms. However, compared to men with one male superior, men with two female superi- ors report significantly lower levels of dis- tress. These comparisons were excluded from the focal associations reported in Tables 2 and 3 because of small cell sizes and elevated standard errors, which under- mine our confidence in the estimates. For example, while it is interesting that men working with two female superiors report less distress than those working with one male superior, there are only 11 men with two female superiors. Future investigations of the effects of having multiple same-gen- der superiors can provide valuable informa- tion and credible comparisons if a larger sample can be obtained.

5. Although it is beyond the scope of this arti- cle, we recognize that some readers may suspect that gender composition of cowork- ers, clients/customers, and subordinates al- so has relevance for workers’ health. In nu- merous additional analyses (not shown) of workers who reported having coworkers, subordinates, and customers/clients, we

298 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

compared workers in all-male contexts within each scenario to those with all-fe- male or gender-mixed compositions in each respective role-set. However, none of these contrasts revealed statistically significant disparities in levels of distress or physical symptoms (full analyses available from the authors upon request). These patterns fur- ther underscore our earlier assertions about the importance of relational demography within the superordinate–subordinate role- set.

REFERENCES Brenner, O.C., Joseph Tomkiewicz, and Virginia

Ellen Schein. 1989. “The Relationship between Sex Role Stereotypes and Requisite Manage- ment Characteristics Revisited.” Academy of Management Journal 32:662–69.

Burke, Ronald J. and Carol A. McKeen. 1996. “Do Women at the Top Make a Difference? Gender Proportions and the Experience of Managerial and Professional Women.” Human Relations 49:1093–1104.

Chatman, Jennifer A., Jeffrey T. Polzer, Sigal G. Barsade, and Margaret A. Neale. 1998. “Being Different Yet Feeling Similar: The Influence of Demographic Composition and Organizational Culture on Work Processes and Outcomes.” Administrative Science Quarterly 43:749–80.

Cooper, Virginia W. 1997. “Homophily or the Queen Bee Syndrome: Female Evaluation of Female Leadership.” Small Group Research 28:483–99.

Correll, Shelley J. and Cecilia L. Ridgeway. 2003. “Expectations States Theory.” Pp. 29–51 in Handbook of Social Psychology, Handbooks of Sociology and Social Research, edited by J. Delamater. New York, NY: Kluwer Academic/Plenum Publishers.

Cortina, Lilia, Vicki Magley, Jill H. Williams, and Regina D. Langhout. 2001. “Incivility in the Workplace: Incidence and Impact.” Journal of Occupational Health Psychology 6:64–80.

Crompton, Rosemary and Fiona Harris. 1998. “Gender Relations and Employment: The Impact of Occupation.” Work, Employment and Society 12:297–315.

Driskell, James E. and Brian Mullen. 1990. “Status, Expectations, and Behavior: A Meta-Analytic Review and Test of the Theory.” Personality and Psychology Bulletin 16:541–53.

Duffy, Michelle K. and Walter J. Ferrier. 2003. “Birds of a Feather .|.|.? How Supervisor–Subor- dinate Dissimilarity Moderates the Influence of Supervisor Behaviors on Workplace Attitudes.” Group and Organization Management 28:217–48.

Eagly, Alice H., Mary C. Johannesen-Schmidt, and

Marloes L. van Engen. 2003. “Transformational, Transactional, and Laissez-Faire Leadership Styles: A Meta-Analysis Comparing Women and Men.” Psychological Bulletin 129:569–91.

Eagly, Alice H. and Steven J. Karau. 2002. “Role Congruity Theory of Prejudice toward Female Leaders.” Psychological Review 109:573–98.

Eagly, Alice H., Steven J. Karau, and Mona G. Makhijani. 1995. “Gender and the Effectiveness of Leaders: A Meta-Analysis.” Psychological Bulletin 117:125–45.

Eagly, Alice H., Mona G. Makhijani, and Bruce G. Klonsky. 1992. “Gender and the Evaluation of Leaders: A Meta-Analysis.” Psychological Bul- letin 111:3–22.

Elliott, James R. and Ryan A. Smith. 2001 “Ethnic Matching of Supervisors to Subordinate Work Groups: Findings on “Bottom-up” Ascription and Social Closure.” Social Problems 48:258–76.

Ely, Robin J. 1994. “The Effects of Organizational Demographics and Social Identity on Relation- ships among Professional Women.” Administra- tive Science Quarterly 39:203–38.

Freese, Lee and Bernard P. Cohen. 1973. “Eliminat- ing Status Generalization.” Sociometry 36:177–93.

Hodson, Randy. 2002. “Demography or Respect?: Work Group Demography Versus Organizational Dynamics as Determinants of Meaning and Sat- isfaction at Work.” British Journal of Sociology 53:291–317.

Hoel, Cary L., Brian Cooper, and Helge Faragher. 2001. “The Experience of Bullying in Great Britain: The Impact of Organizational Status.” European Journal of Work and Organizational Psychology 10:443–65.

Jehn, Karen A., Gregory B. Northcraft, and Mar- garet A. Neale. 1999. “Why Differences Make a Difference: A Field Study of Diversity, Conflict, and Performance in Workgroups.” Administra- tive Science Quarterly 44:741–63.

Kanter, Rosabeth. 1977. Men and Women of the Corporation. New York: Basic Books.

Kasl, Stanislav V. 1998. “Measuring Job Stressors and Studying the Health Impact of the Work Environment: An Epidemiologic Commentary.” Journal of Occupational Health Psychology 3:390–401.

Lepkowski, James M. 1988. “Telephone Sampling Methods in the United States.” Pp. 73–98 in Telephone Survey Methodology, edited by R. M. Groves, P. P. Biemer, and L. E. Lyberg. New York, NY: John Wiley & Sons.

Mirowsky, John and Catherine E. Ross. 2003. Social Causes of Psychological Distress. New York, NY: Aldine de Gruyter.

Moore, Sarah, Leon Grunberg, and Edward S. Greenberg. 2005. “Are Female Supervisors Good for Employee Job Experiences, Health,

RELATIONAL DEMOGRAPHY AND HEALTH 299

Scott Schieman is professor of sociology at the University of Toronto. His research focuses on work, stress, and health processes. In addition, he is currently working on a book titled Divine Control, which explores beliefs about God’s involvement and causal relevance in everyday life.

Tara McMullen is a third year PhD candidate in the Department of Sociology at the University of Toronto. Her interests include gender and the intersection of health and mental health processes.

300 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

and Well-Being?” Women in Management Review 20:86–95.

O’Reilly, Charles A., Katherine Y. Williams, and Sigal Barsade. 1998. “Group Demography and Innovation: Does Diversity Help?” Pp. 183–207 in Research on Managing Groups and Teams, Vol. 1, Composition, edited by M. A. Neale, E. A. Mannix, and D. H. Gruenfeld. Stamford, CT: JAI Press.

Pearlin, Leonard I. 1999. “The Stress Process Revis- ited: Reflections on Concepts and Their Interre- lationships.” Pp. 395–415 in The Handbook of the Sociology of Mental Health, edited by C. S. Aneshensel and J. C. Phelan. New York: Kluwer Academic/Plenum.

Pelled, Lisa Hope, Kathleen M. Eisenhardt, and Katherine R. Xin. 1999. “Exploring the Black Box: An Analysis of Work Group Diversity, Con- flict, and Performance.” Administrative Science Quarterly 44:1–28.

Perry, Elissa L., Carol T. Kulik, and Jing Zhou. 1999. “A Closer Look at the Effects of Subordi- nate-Supervisor Age Differences.” Journal of Organizational Behavior 20:341–57.

Pfeffer, Jeffrey. 1983. “Organizational Demogra- phy.” Pp. 299–357 in Research in Organization- al Behavior, edited by L. L. Cummings and B. M. Staw. Greenwich, CT: JAI.

Reskin, Barbara F. and Denise B. Bielby. 2005. “A Sociological Perspective on Gender and Career Outcomes.” The Journal of Economic Perspec- tives 19:71–86.

Ridgeway, Cecilia L. 2001. “Gender, Status, and Leadership.” Journal of Social Issues 57: 637–55.

Ridgeway, Cecilia L. and Lynn Smith-Lovin. 1999. “The Gender System and Interaction.” Annual Review of Sociology 25:191–216.

Scott, Kristyn A. and Doug J. Brown. 2006. “Female First, Leader Second? Gender Bias in the Encod- ing of Leadership Behavior.” Organizational Behavior and Human Decision Processes 101:230–42.

Smith, Ryan A. 2002. “Race, Gender, and Authority in the Workplace: Theory and Research.” Annual Review of Sociology 28:509–42.

Staines, Graham and Carol Tavris. 1974. “The

Queen Bee Syndrome.” Psychology Today 7:55–60.

Tsui, Anne S., Terri D. Egan, and Charles A. O’Reil- ly. 1992. “Being Different: Relational Demogra- phy and Organizational Attachment.” Adminis- trative Science Quarterly 37:549–79.

Tsui, Anne S. and Charles A. O’Reilly. 1989. “Beyond Simple Demographic Effects: The Importance of Relational Demography in Supe- rior-Subordinate Dyads.” Academy of Manage- ment Journal 32:402–23.

Valentine, Sean and Lynn Godkin. 2000. “Supervi- sor Gender, Leadership Style, and Perceived Job Design.” Women in Management Review 15:117–29.

Valentine, Sean, Lynn Godkin, and James H. Turn- er. 2002. “Women’s Management, Perceived Job Responsibility, and Job Search Intention.” Women in Management Review 17:29–38.

Vallas, Steven P. 2003. “Why Teamwork Fails: Obstacles to Workplace Change in Four Manu- facturing Plants.” American Sociological Review 68:223–50.

Vecchio, Robert P. 1993. “The Impact of Differ- ences in Subordinate and Supervisor Age on Attitudes and Performance.” Psychology and Ageing 8:112–19.

Vecchio, Robert P. and R. Craig Bullis. 2001. “Mod- erators of the Influence of Supervisor-Subordi- nate Similarity on Subordinate Outcomes.” Jour- nal of Applied Psychology 86:884–96.

Wajcman, Judy. 1996. “Desperately Seeking Differ- ences: Is Management Style Gendered?” British Journal of Industrial Relations 34:333–49.

———. 1998. Managing Like a Man: Women and Men in Corporate Management. University Park: Pennsylvania State University Press.

Waksberg, Joseph. 1978. “Sampling Methods for Random Digit Dialing.” Journal of the American Statistical Association 73:40–46.

Wharton, Amy and James N. Baron. 1991. “Satis- faction? The Psychological Impact of Gender Segregation on Women at Work.” The Sociologi- cal Quarterly 32:365–87.

Wheaton, Blair. 2001. “The Role of Sociology in the Study of Mental Health .|.|. and the Role of Mental Health in Sociology.” Journal of Health and Social Behavior 42:221–34.