ANNOTATED BIBLIOGRAPHY

sagepub.com/journalsPermissions.nav DOI: 10.1177/0734371X0935\m1823 h\f\fp://roppa.sagepub.com Workforce Diversity in the New Mi\f\fennium: \brospects for Research David W. \bitts 1 and Lois Recascin\Mo Wise 2 Abstract Public organiza\fions in \fhe new millennium are \fasked wi\fh a myriad of human resource managemen\f challenges \fha\f s\fem from workforce diversi\fy, bu\f \fhe field of public adminis\fra\fion has no\f produced a body of research \fha\f adequa\fely assis\fs \fhem wi\fh \fhese s\fruggles. In 2000, Wise and Tschirhar\f called for “grea\fer con\fribu\fion from public adminis\fra\fion scholars \fo \fhe body of research focusing on how human diversi\fy can bes\f be managed \fo produce posi\five resul\fs.” They found \fha\f exis\fing research con\fribu\fed li\f\fle usable knowledge for diversi\fy managemen\f policies and programs.

The au\fhors examine whe\fher \fheir call for more rigorous and more prac\fice-orien\fed research has been heeded by iden\fifying ar\ficles on workforce diversi\fy published in a core se\f of public adminis\fra\fion journals since 2000. A broad overview of \fhe li\fera\fure on diversi\fy is provided, followed by a more focused discussion of empirical research on employmen\f diversi\fy, diversi\fy managemen\f, and organiza\fional ou\fpu\fs and ou\fcomes. I\f is found \fha\f al\fhough diversi\fy issues remain salien\f \fo public adminis\fra\fion scholarship, usable knowledge is in shor\f supply. A subs\fan\fial share of \fhis research can be ca\fegorized as focusing on represen\fa\five bureaucracy issues. Few empirical s\fudies \fes\f diversi\fy effec\fs or hypo\fheses. Some empirical work explains fac\fors beyond \fhe con\frol of human resource policies or prac\ficing managers, which makes findings less useful \fo prac\fi\fioners. The research suffers from inadequa\fe da\fa, li\f\fle innova\fion in me\fhodology, and insufficien\f a\f\fen\fion \fo empirical connec\fions be\fween diversi\fy and organiza\fional resul\fs.

Keywords diversi\fy, represen\fa\fion, represen\fa\five bureaucracy 1American Universi\fy, Washing\fon, DC, USA2Indiana Universi\fy, Blooming\fon, IN, USA Corresponding Author:

David W. Pi\f\fs, Depar\fmen\f of Public Adminis\fra\fion and P\molicy, American Universi\fy, 4400 Massachuse\f\fs \m Avenue NW, Washing\fon, DC 20009, USA Email: pi\f\[email protected] Pitts and Wise 45 The purpose of this article is to examine the landscape of research on \forkforce diversity in pu\blic-sector organizations. Workforce diversity has \become one of the most salient management issues in organizations from all sectors. When the Review of Public Personnel Adminis\fra\fion (ROPPA) \fas first pu\blished in 1980, White males accounted for 86% of all Senior Executive Service (SES) employees in the U.S. fed- eral government. By 2008, that num\ber had decreased to 65% (Office of Personnel Management, 2008). In addition to more racial/ethnic and gender diversity, glo\baliza- tion has led to increases in cultural and linguistic diversity as \fell. A\bout 18% of all households in the United States use a language other than English, and a\bout 13% of U.S. residents \fere \born in a different country (Ru\baii Barrett & Wise, 2007\b). The legal environment for diversity has \been altered dramatically \by landmark court deci- sions such as Ba\b\be, Adarand, and the Gra\fz and Gru\f\fer cases, as \fell as legislation such as the Americans \fith Disa\bilities Act of 1990 and the Civil Rights Act of 1991.

Since 1980, 12 states have passed legislation \banning employment discrimination on the \basis of sexual orientation (Human Rights Campaign, 2007). Given these su\bstan- tial shifts in the environment of pu\blic organizations, the 30th anniversary of ROPPA is an excellent juncture at \fhich to assess the state of research on diversity and its util- ity to pu\blic secto\dr practitioners. These changes in \forkforce diversity have required organizations to change ho\f they think a\bout human resources management. Their approach to diversity has his- torically focused on compliance \fith the la\fs and regulations governing recruitment, selection, and separation (for a revie\f, see Riccucci, 2002). Ho\fever, \forld\fide gov- ernment reforms during the 1990s and 2000s elevated the role of performance and strategy in managing pu\blic organizations (Kettl, 2000). A focus on strategy and performance requires organizations to tie management processes to larger goals and o\bjectives, \fhich has led to many organizations to consider ho\f \forkforce diversity can \be used to accomplish their missions. Indeed, organizations often high- light that employee diversity is a “strength,” \fhich implies that diversity can \be used as a lever to improve performance. This idea has developed over time as the “\business case” for diversity \d(for a revie\f, see Kochan et al\d., 2003). Although arguments such as the \business case for diversity are intuitively appeal- ing and politically popular, there is little evidence that organizational diversity can \be used to \boost performance. Whether employee diversity improves organizational per- formance is an empirical question that has not \been adequately tested in the pu\blic- sector context. In 2000, Wise and Tschirhart revie\fed the evidence on connections \bet\feen employee diversity and organizational performance, finding very little research in this area that focused on pu\blic organizations. Their article concluded \fith a call for pu\blic administration scholars to engage in more empirical research on \forkforce diversity. Our primary purpose in this article is to examine the extent to \fhich research in pu\blic administration has responded to Wise and Tschirhart’s (2000) push for greater emphasis on diversity scholarship. Our analysis proceeds in t\fo parts. We \begin \fith an overvie\f of the research on \forkforce diversity that \fas pu\blished in a core set of 46 Re view of Public Pe\fsonnel Administ\fation 3\b(1) pu\blic administration journals since 2000. This part of our analysis is purposefully \broad and meant to include a general cross-section of approaches to diversity research.

For example, \fe include research focusing on particular dimensions of diversity, such as race, ethnicity, or gender; articles focusing on a particular group \fithin a dimension, such as African Americans or \fomen; research on diversity management programs and policies; and analyses and commentary on legal issues related to diversity. Some of the articles are descriptive, \fhereas others are more analytical. We categorize each article \by the dimension of diversity that is examined, the journal in \fhich the article appeared, the methodology used \by the authors, and the year of pu\blication. The second part of our examination is a more targeted analysis of the articles that explore the empirical relationship \bet\feen \forkforce diversity and organizational out- comes. For each article, \fe categorize the dimensions of diversity that are examined, the organizational context, and the authors’ empirical strategy for measuring diversity.

We argue that the second part of our analysis is particularly important, given the evolu- tion of the \business case for diversity and government reforms related to performance. In the next section, \fe provide a \brief overvie\f of ho\f research on \forkforce diversity has evolved over time. We outline the data and method that \fe used to iden- tify \forkforce diversity research, moving to a discussion \based on our initial overvie\f of research pu\blished since 2000. We then more closely examine the empirical research on diversity, diversity manageme\dnt, and outcomes.

The Evo\fution of Research on Workforce Diversity There is general consensus that the roots of pu\blic administration research on \fork- force diversity are in representative \bureaucracy. Representative \bureaucracy research explores the demographic profile of government employees and ho\f it compares to characteristics of citizens and service recipients. The earliest research in this area focused on social class (Kingsley, 1944), \but most of the research since that time has focused on \fomen and people of color (for a revie\f, see Dolan & Rosen\bloom, 2003). Research has typically found that \fomen and people of color are underrepresented in government organizations, \fith steady improvements over time despite persistent shortfalls at the highest ranks (Kellough, 1990; Le\fis, 1992; Riccucci & Saidel, 1997, 2001; Wise, 1990). Some have expanded on this \fork to find that shared social and cultural experiences lead to other results, such as improved services or advancement of policies that \benefit \fomen and minorities (Hindera, 1993; Keiser, Wilkins, Meier, & Holland, 2002; Meier, 1993; Selden, 1997; Wilkins & Keiser, 2006). Other studies focus on employee outcomes in lieu of citizen out- comes and find that advancement and pay gaps tend to exist \bet\feen men and \fomen (Guy, 1994; Le\fis, 1992; Wise, 1994) and \bet\feen Whites and people of color (Naff, 2001; Naff & Kellough, 2003). The frame\fork of la\fs that protect employment equity has \been crucial in promot- ing representation, and this is the focus of another vital stream of diversity research in pu\blic administration. The past 30 years have seen a series of landmark cases that have Pitts and Wise 47 \forked to \both expand and contract Affirmative Action (AA) programs, nota\bly Regen\fs v. Ba\b\be, Adarand v. Pena , Gru\fz v. Bolliner , and Gru\f\fer v. Bollinger . Analysis and commentary on these cases have provided practitioners \fith the appropriate tools for understanding the rights and responsi\bilities associated \fith pu\blic employment (see, e.g., Brad\bury, in press; Carcieri, 2004; Naff, 2004; Naylor & Rosen\bloom, 2004; Sisneros, 2004). Although many legal cases and the research that results from them focus on \fomen and people of color, legal changes that affect other dimensions of diversity have \been examined in this stream of research as \fell, most nota\bly as they relate to age (Wilkins, 2006), disa\bility status (Brad\bury, 2007), and sexual ori- entation (Colvin, 2\d000, 2007; Riccucc\di & Gossett, 1996).\d Although research on representation and AA/Equal Employment Opportunity (EEO) has continued, some scholars have shifted their focus to the impact of employee diversity on \fork-related outcomes. Theory suggests that employee diversity \benefits organizations \by increasing the num\ber of perspectives and solutions to pro\blems, \but that it can also make organizations more likely to experience employee conflict, mis- communication, and mistrust (Adler, 2003; Williams & O’Reilly, 1998). Limited empirical research has examined the extent to \fhich these process-oriented pro\blems out\feigh the \benefits accrued from greater diversity in employee pers pectives, \fith mixed results (for a revie\f, see Williams & O’Reilly, 1998). Kno\fledge is limited on diversity effects in for-profit firms (Kochran et al., 2003), and \fe kno\f even less a\bout the relationship \bet\feen diversity and results in pu\blic agencies (Pitts, 2005; Pitts & Jarry, 2007; Wise & Tschirhart, 2000). This is a challenging area of research, \because the relationship \bet\feen diversity and performance is often nonlinear and contingent on other factors, requiring strong data and often complex modeling strategies (Choi, in press; P\ditts & Jarry, in press). Much of the research on \forkforce diversity has instead chosen to focus on the diversity management programs that \became popular in the early 1990s. Manage- ment researchers had \begun to o\bserve that AA made it possi\ble for a \fider variety of employees to gain entry into firms, \but the organizational cultures often remained the same, \fhich prevented organizations from realizing any \benefits from diversity (Konrad, 2003). Firms responded \by adopting values-\based diversity programs in an attempt to modify organizational culture, \but research found that these initiatives \fere often unsuccessful (Bezrukova & Jehn, 2001). Thomas (1990) \fas one of the first to make the argument that neither the AA/EEO nor the “valuing diversity” paradigm \fas effective, advocating for a third approach: managing for diversity. This approach \fas more pragmatic and permitted organizations to create strategic policies and pro- grams for managing diversity. By emphasizing the potential \bottom line \benefits, Thomas made diversity initiatives more palata\ble to majority employees \fhose inter- est \fas on the \bottom line. Organizations that managed for diversity \fould \be more likely to recruit and retain diverse and effective employees, \fhich \fould in turn lead to performance \benefits. Pu\blic-sector organizations \fere quick to adopt this philoso- phy \because it \fas politically defensi\ble in \fays that AA/EEO never \fas and fit in \fith the performance-\based reforms of the 1990s (Kellough & Naff, 2004; Rangarajan & Black, 2007; Ric\dcucci, 2002). 48 Re view of Public Pe\fsonnel Administ\fation 3\b(1) By 1999, 90% of U.S. federal government agencies had esta\blished a diversity management program, though some did not deviate much from earlier AA/EEO initia- tives, and the actu\dal impact on advanc\dement and pay equit\dy for \fomen and peop\dle of color \fas limited (Kellough & Naff, 2004; Naff & Kellough, 2003). Research on \fhether these programs are effective is limited. Evidence suggests that diversity man- agement programs can \boost jo\b satisfaction and perceptions of performance among people of color (Pitts, 2009). Ho\fever, the causal path \bet\feen diversity management and organizational performance is complex, \fith research demands that are frequently unmet \by existing data (Choi, in press). Much of the existing research on diversity management tends to \be prescriptive and \based on anecdotal data, limited case studies, or theories from other fields of study (see, e.g., Arai, Wanca-Thi\bault, & Shockley- Zala\bak, 2001; Pitts\d, 2007; Von Bergen, Soper, & Foster, 2002).

Data and Method The purpose of our analysis is to provide an overvie\f of pu\blic administration research on \forkforce diversity pu\blished since Wise and Tschirhart’s (2000) call for greater attention to pu\blic-sector diversity issues. We examined all articles pu\blished in a core set of 12 pu\blic administration journals since 2000 (Ta\ble 1). To dra\f a sample of research that \fas comprised of units that could \be compared \fith one another and across years, \fe did not consider \books, \book chapters, or \book revie\fs. We aimed to include a \broad representation of journals \based in the United States that pu\blish research on pu\blic-sector \forkforce issues, dra\fing heavily from journals used in similar earlier studies (Bre\fer, Douglas, Facer, & O’Toole, 1999). This included some journals focused almost exclusively on \forkforce issues (e.g., Public Personnel Man- agemen\f and Review of Public Personnel Adminis\fra\fion); some focused on pu\blic management more \broadly (e.g., Public Performance and Managemen\f Review or In\ferna\fional Public Managemen\f Journal); and some \fith a primary focus on pu\blic policy \but occasional interest in management (e.g., Journal of Policy Analysis and Managemen\f). Interestingly, the Wise and Tschirhart (2000) 1961-1998 sample that \fas \based on hypothesis-testing \djournal articles focusing on the relationship \bet\feen heterogeneity and p\derformance did not \dcapture any of thes\de journals. We used three criteria in selecting articles for inclusion in our analysis. As a starting point, the research had to focus on \forkforce diversity as a central theme. Diversity is a tough concept to define. Perhaps the most restrictive definition is held \by organi - z ational \behavior scholars, \fho consider diversity to \be a concept of variation or heterogeneity. Greater heterogeneity reflects greater variation among parts of a \fhole, meaning that the highest \forkforce heterogeneity is achieved \fhen employees are split evenly among all categories and groups. This approach measures heterogeneity as it exists on particular dimensions, \fhich in the context of the \forkforce \fould include race, ethnicity, gender, and others. The heterogeneity construct is agnostic to historical representation patterns and discrimination—an organization that is 95% African American and 5% White is less heterogeneous and diverse than one that is 50% White and 50% Asian American. Heterogeneity is typically measured independently for each Pitts and Wise 49 dimension using an index that reflects dispersion across all groups, such as a Blau or Herfindahl Index. Quantitative varia\bles that reflect the aggregate percentage of employees \fho are non-White or from a particular minority group are no\f measures of heterogeneity—they \d are simply indicators of the share of the organization or \fork- force held \by a part\dicular group. Research on \forkforce heterogeneity is an important component of diversity research, \but unlike the Wise and Tschirhart revie\f \fe do not limit our analysis to those studies here. We also included articles that consider the term diversity more Tab\fe 1. Journal and Year of Publica\fion Journal 2000 2001 2002 2003 2004 2005 2006 2007 2008 T o\fal P ercen\fage Review of 3 0 1 2 4 2 5 6 5 28 31.5 Public Pe\fsonnel Administ\fation Public 2 0 1 2 4 3 5 0 1 18 20.2 Administ\f ation Review Public Pe\fsonnel 3 1 1 4 1 0 0 2 1 13 14.6 Manag ement Inte\fnational 0 1 2 2 1 0 0 5 0 11 12.4 J ou\fnal of Public Administ\fation Jou\fnal of Public 2 0 0 1 1 3 1 2 1 11 12.4 Administ\f ation Resea\fch & Theo\fy Administ\fation & 0 1 0 0 1 0 0 2 0 4 4.5 Society Ame\f ican Review 0 0 0 0 0 0 0 1 0 1 2.3 of Public Administ\fation Inte\fnational Publi\cc 0 0 0 1 0 0 0 0 0 1 1.1 Manag ement Jou\fnal Public Administ\fation 0 0 0 0 0 1 0 0 0 1 1.1 Qua\f te\fly Public O\fganization\c 0 0 0 0 0 0 0 1 0 1 — Re view Administ\fative 0 0 0 0 0 0 0 0 0 0 — Theo\f y & P\faxis Jou\fnal of Policy 0 0 0 0 0 0 0 0 0 0 — Analysis & Management Public Pe\ffo\fmance 0 0 0 0 0 0 0 0 0 0 — & Manag ement Review To\fal 10 3 5 12 12 9 11 19 8 89 50 Re view of Public Pe\fsonnel Administ\fation 3\b(1) loosely to apply to a particular minority group (e.g., African Americans, \fomen, persons \fith disa\bilities) \by itself and not in relation to other minority or majority group mem\bers. Studies that hone in on the issues related to one group can \benefit the larger research agenda on diversity, though differently than research on \forkgroup or organizational heterogeneity. For this issue, \fe \fere most concerned \fith \fhether \forkforce diversity—ho\fever measured—\fas a central component of the piece, not just a control varia\ble in analyses that primarily explored other topics. If the article did not discuss \forkforce diversity or one of its dimensions in the title or a\bstract, \fe assumed the contri\bution \fould \be limited enough to \farrant exclusion. Similarly, \fe did not consider articles that focused only on diversity among clients or the target population, although important they are \beyond our scope. A num\ber of articles considered ho\f diversity among service recipients affected policy outcomes, \but these \fere included in our analysis only if wor\bforce diversity \fas also a primary component of the article. Because \forkforce diversity is an inherent component of representative \bureaucracy research and thus important for pu\blic administration, \fe included \fork in that area in our sample. Thus, our sample differs in important \fays from the one dra\fn \by Wise and Tschirhart (2000) and should have greater potential for capturing diversity-related s\dcholarship focusin\dg on the pu\blic sec\dtor. Our search resulted in a sample of 89 research articles that appeared from 2000 to 2008. We evaluated each article using several criteria. We determined \fhether an article focused on one or more specific dimensions of diversity: (a) race or ethnicity, (\b) sex or gender, (c) disa\bility status, (d) social class, (e) age, (f) education or func- tion, (g) sexual orientation, (h) religion, and (i) nationality/language. This assists us in evaluating \fhether \forkforce diversity research is focused primarily on dimen- sions of diversity that have historically \been considered relevant (e.g., race, ethnicity, sex, and gender) or \fhether it is moving into other areas in response to ne\f \forkforce demands (e.g., sexual orientation, age, nationality/language). This is a particularly relevant question for diversity researchers, given that Thomas’ (1990) early defini- tion of the diversity management concept \fas that it \be multidimensional and inclusive of all differences. Second, \fe evaluated the methodology of each article and determined \fhether it used (a) quantitative methods only, (\b) qualitative methods only, (c) mixed methods, (d) used empirical methods to analyze legal issues, or (e) \fas not empirical. Third, \fe identified \fhether each article aimed to understand diversity or diversity management empirically. This could take one of t\fo forms. On one hand, the study could use diver- sity or diversity management as a key independent varia\ble, \fith the corresponding dependent varia\ble \being an organization-level result or outcome. In principle, \fe \fould include individual, group, and organizational outcomes. A study could also use diversity or diversity management as a dependent varia\ble, \fith a series of organiza- tional and/or environmental factors as predictors. The key issue is \fhether research aims to understand diversity or diversity management as it is linked empirically to other factors, an approach that \fe argue is the most likely to result in gains for \both theory and practice—regardless\d of \fhether diversity is on the left or right side of the Pitts and Wise 51 equation. The articles that meet these criteria are then extracted for more comprehen- sive analysis in th\de second part of ou\dr study.

Trends in Workforce Diversity Research In contrast to Wise and Tschirhart’s findings, \fe did not find a clear trajectory in the num\ber of articles pu\blished over time. The num\ber ranged from only 3 in 2001 to 19 in 2007, \fith dips and increases in \bet\feen and only 8 pieces in 2008. We found a stronger pattern in the venues that pu\blished diversity research (Ta\ble 1). Almost one- third of the articles appeared in ROPPA, a total of 28 articles across the 9 years of research. Four other journals \fere frequent sources of diversity research: Public Adminis\fra\fion Review (18 articles), Public Personnel Managemen\f (13 articles), and In\ferna\fional Journal of Public Adminis\fra\fion and Journal of Public Adminis\fra\fion Research and Theory, \fith 11 articles each. Although \broad dissemination of diversity scholarship across the major journals of the field \fould \be prefera\ble in some respects, it is also valua\ble for specific journals to develop ongoing streams of research on diversity issues. That appears to \be the case \fith \both ROPPA and Public Personnel Managemen\f—t\fo key sources of\d scholarship on pu\d\blic-sector \forkfor\dce topics. Attention to specific dimensions of diversity varied \fidely, \fith a predicta\ble emphasis on \fomen and people of color (Ta\ble 2). From 2000 to 2008, \fe identified 44 articles that addressed sex and gender issues in the \forkforce, \fith \fide ranging foci that included employment trends (e.g., Kim, 2003; Llorens, Wenger, & Kellough, 2008; Pynes, 2000), active representation \by gender (e.g., Wilkins & Keiser, 2006; Wilkins, 2007), and differences \bet\feen men and \fomen in the \forkplace (e.g., DeHart-Davis, Marlo\fe, & Pandey, 2006; Guy & Ne\fman, 2004; Stackman, Connor, & Becker, 2005). We found 33 articles that addressed issues of race and ethnicity. The \bulk of this research focused on \forkforce trends (e.g., Charles, 2003; McCa\be & Stream, 2000), active representation (e.g., Brudney, Her\bert, & Wright, 2000; Riccucci & Meyers, 2004; So\fa & Selden, 2003), and affirmative action issues (e.g., Gest & Maranto, 2000; Naylor & Rosen\bloom, 2004). Consistent \fith Wise and Tschirhart (2000), \fe find that studies in our sample are primarily focused on sex/ gender and race/eth\dnicity. Our analysis identified only nine articles that addressed age in the \forkplace, despite gro\fing concerns in the 1990s a\bout retirements and the graying of the \fork- force (West & Berman, 1996). Some of this research addressed links \bet\feen age and employee outcomes and \behaviors, such as jo\b satisfaction (Jung, Moon, & Hahm, 2007), turnover intention (Moynihan & Landuyt, 2008), and response to online recruitment (Ru\baii-Barrett & Wise, 2007a). Other articles focused more on develop- ments in antidiscrimination la\fs protecting older \forkers (e.g., Riccucci, 2003; Wilkins, 2006). Six articles from 2000 to 2008 addressed disa\bility issues, including analysis of court decisions limiting the Americans \fith Disa\bilities Act (Riccucci, 2003) and \forkplace issues for managing employees \fith disa\bilities (e.g., Balser, 2007; Brad\bury, 2007). An additional five articles considered issues of sexual 52 Re view of Public Pe\fsonnel Administ\fation 3\b(1) orientation, primarily as it is included in antidiscrimination policies (e.g., Colvin, 2000, 2007; Colvin & Riccucci, 2002). Only four articles addressed religion in the \forkplace and only t\fo of these con- sidered the managerial challenges of religious diversity (Ball & Haque, 2003; Garcia-Zamor, 2003). We identified only one article each that addressed education/ functional diversit\dy, social class dive\drsity, and diversity of n\dationality or langu\dage. More than half of the articles in our study (55.1%, N = 49) used quantitative, empirical research methods to test their questions (Ta\ble 3). Roughly 11% ( N = 10) of the articles used mixed methods, \fhereas qualitative methods \fere used in only 3 articles (3.4%). We identified 15 articles (16.9%) that used empirical methods to examine legal issues, typically an analysis of recent court decisions (e.g., Naylor & Rosen\bloom, 2004; Riccucci, 2003), legislation (e.g., Zeigler, 2006), or state-level adoption of diversity-focused discrimination policies (e.g., Colvin, 2007). A total of 12 of the articles (13.5%) explored diversity \fithout using empirical methods. Some of these focused on developing conceptual or theoretical frame\forks for diversity (e.g., Foldy, 2004; Pitts, 2006; Selden & Selden, 2001), \fhereas some used research from other fields to develop \best practices for diversity management implementation (e.g., Arai et al., 2001; Pitts, 2007). Finally, \fe found that articles \fere most likely to focus on the federal or state government context (Ta\ble 4). Roughly, 42% ( N = 33) considered diversity in the context of federal agencies, and an additional 30.8% ( N = 49) focused on state gov- ernment agencies. Only 14.1% ( N = 11) addressed diversity issues in local government organizations, and 11.5% ( N = 9) considered diversity issues in special districts or pu\blic schools. The context for the analysis is dependent on a num\ber of factors, chief among them \being data availa\bility. Federal-level data are \fidely availa\ble, making those agencies a prime context for research on diversity. In our sample, only one article explicitly addressed nonprofit organizations, \but this is not surprising, Tab\fe 2. Dimension of Diversi\fy by Year of Publica\fion Dimension 2000 2001 2002 2003 2004 2005 2006 2007 2008 T o\fal P ercen\fage Sex/gender 7 1 2 5 8 5 6 5 5 44 49.4 Race/e\fhnici\fy 6 0 0 5 7 4 1 7 3 33 37.1 Age 1 0 0 2 1 1 1 2 1 9 10.1 Disabili\fy 0 0 0 1 1 0 0 4 0 6 6.7 Sexual 1 0 2 0 1 0 0 1 0 5 5.6 orien\fa\fion Religion 0 0 0 2 1 1 0 0 0 4 4.5 Social class 0 0 0 0 0 0 0 1 0 1 1.1 Educa\fion/ 1 0 0 0 0 0 0 0 0 1 1.1 func\fion Na\fionali\fy/ 0 0 0 0 0 0 0 1 0 1 1.1 language Pitts and Wise 53 given that our sample of journals did not include some of the primary venues for nonprofit research.

Trends in Empirica\f R\Mesearch on Diversity and Outcomes\M Of the 89 articles that \fe identified, only 24 examined diversity empirically at the organizational level. For ease of discussion, \fe split them into t\fo groups: those that predict diversity as a dependent varia\ble and those that use diversity as an independent varia\ble to explain one or more organizational outcomes. We proceed \fith our discus- sion in four sections: (a) the dimensions of diversity covered in these studies, (\b) analysis of research using diversity as a dependent varia\ble, (c) analysis of research using diversity as an independent varia\ble explaining organizational outcomes, and (d) data and context issues \dacross studies.

Dimensions of Dive\f\csity The dimensions of diversity explored in this cluster of studies \fere limited. None of the studies in our sample examined interactive effects of different diversity dimen- sions, \but rather appear to make an implicit assumption that all diversity dimensions are of equal importance, regardless of context. The over\fhelming majority focused on race/ethnicity (16 articles) and sex/gender (14 articles). One article each addressed age, disa\bility status, and jo\b function. We find no evidence that diversity research is moving a\fay from race/ethnicity and gender/sex \but there are some negative conse- quences to this attachment. Findings from one dimension of diversity do not necessarily apply to another dimension of diversity (Wise & Tschirhart, 2000), \fhich means that evidence a\bout racial/ethnic and sex/gender diversity does not improve our under- standing of other dimensions. For example, there is nothing a\bout ho\f government agencies and \fork groups are affected \by the presence of non-native speakers of Tab\fe 3. Me\fhodological Approach by Year of Publica\fion Me\fhodology 2000 2001 2002 2003 2004 2005 2006 2007 2008 T o\fal P ercen\fage Quan\fi\fa\five 6 1 2 6 4 7 5 12 6 49 55.1 only Quali\fa\five 0 0 0 1 0 1 0 1 0 3 3.4 only Bo\fh/mixed 2 0 1 2 3 1 0 1 0 10 11.2 me\fhods Empirical 1 0 0 1 4 0 3 4 2 15 16.9 approaches \fo legal issues Nonempirical 1 2 2 2 1 0 3 1 0 12 13.5 T o\fal 10 3 5 12 12 9 11 19 8 89 54 Re view of Public Pe\fsonnel Administ\fation 3\b(1) English. The \business case for diversity might suggest that language diversity leads to \broader cultural perspectives that \benefit performance, \but does it cause communica- tion \breakdo\fns that make performance suffer? Some research suggests that it \fould, particularly in complex tasks that require explanation and discussion (Ham\brick et al., 1998). Are diversity management programs that address language issues effective? If so, on \fhat do they \dfocus? If not, on \d\fhat should they focus? These same questions could \be posed for other dimensions of diversity, such as language, sexual orientation, and religion. The issues here are not simply sym\bolic.

The \forkforce is seeing increases in the percentage of employees \fho speak a lan- guage other than English at home (Ru\baii-Barrett & Wise, 2007\b); greater \fillingness of employees to pu\blicly identify themselves as les\bian or gay (Griffith & He\bl, 2002); and a larger variety of religious and spiritual practices among employees (Hicks, 2002). If research is to inform the practice of human resources management, it must identify the consequences of those shifts and understand the management initiatives that are in place \dto affect them.

Dive\fsity as a Dependen\ct Va\fiable A total of 16 articles \fere concerned \fith diversity as a dependent varia\ble—an organi- zational outcome or result—using varying levels of analytical sophistication (Ta\ble 5).

In almost all of these cases, diversity \fas not measured as heterogeneity, \but rather as the percentage of the \forkforce/organization that \fas comprised of particular groups.

T\fo nota\ble exceptions \fere Kim (2005) and Llorens et al. (2008), \both of \fhich used representation ratios that partially reflect heterogeneity. Most of the research used data at the organizational level, an approach that may \be useful in understanding diversity for sym\bolic purposes, \but less helpful in understanding ho\f it relates to outcomes. The \business case for diversity is \based on \benefits that accrue from the interaction of diverse employees, so it is important to discern \fhether diversity exists in different par\fs of the organization. For example, if an organization employs equal Tab\fe 4. Con\fex\f of Research by Year of Publica\fion Con\fex\f 2000 2001 2002 2003 2004 2005 2006 2007 2008 T o\fal P ercen\fage Federal 3 0 2 5 6 3 4 6 4 33 42.3 governmen\f S\fa\fe 4 1 1 3 4 2 2 5 2 24 30.8 governmen\f Local 1 0 0 1 1 2 1 3 2 11 14.1 governmen\f Public schools/ 0 0 1 1 0 2 1 4 0 9 11.5 special dis\fric\fs Nonpr ofi\f 1 0 0 0 0 0 0 0 0 1 1.3 organiza\fions To\fal 9 1 4 10 11 9 8 18 8 78 Tab\fe 5. Empirical Research using Diversi\fy as Ou\fcome of\m In\feres\f S\fudy Naff and Crum \b2000) McCabe and S\fream \b2000) Pynes \b2000) Kim \b2003) Charles \b2003) Naff and Kellough \b2003) Kellough and Naff \b2004) Kim \b2004) Research Ques\fion Do Presiden\fial views of Affirma\five Ac\fion affec\f \fhe represen\fa\fion of women and people of \mcolor in bureaucracy?

How has \fhe represen\fa\fion of women and people of \mcolor changed in governmen\f from 1980 \fo 1995?

Are women underrepresen\fed as leaders in nonp\mrofi\f organiza\fions?

Wha\f affec\fs \fhe represen\fa\fion of women in Korean civil service posi\fions?

Is \fhe represen\fa\fion of older\m workers and people of \mcolor improving in s\fa\fe governmen\f?

Do diversi\fy managemen\f programs lead \fo be\f\fer emplo\mymen\f ou\fcomes for women and people of color?

Wha\f are \fhe componen\fs of \m diversi\fy managemen\f pr\mograms, and wha\f causes age\mncies \fo adop\f \fhem?

Wha\f affec\fs \fhe percen\fages of women and people of \mcolor in s\fa\fe governmen\f agencies? Con\fex\f Sample of employees in U.S. federal governmen\f agencies Aggrega\fe da\fa on federal and s\fa\fe governmen\f employees, 1980-1995 Sample of 200 nonprofi\f organiza\fions in S\f\m. Louis Employmen\f da\fa on all Korean agencies Aggrega\fe da\fa on New Jersey s\fa\fe governmen\f employees, 1993-1999 Aggrega\fe da\fa on U.S. federal governmen\f employees; survey da\fa from 137 agencies Survey da\fa from a sample of 137 U.S. federal agencies Sample of 55 s\fa\fe governmen\f agencies in\m Michigan and Pennsylvania Measure of Diversi\fy Percen\fage of employees a\f differen\f levels who were non-Whi\fe or female Percen\fage of employees by race/e\fhnici\fy an\md sex/ gender Number of organiza\f\mions wi\fh female leaders Percen\fage employees who are women Percen\fage of employees who were older \fhan 65 or \m non-Whi\fe Index measure including \fhe Percen\fage of GS 9-12 posi\fions held by women and people of colo\mr Componen\fs of diversi\fy managemen\f program Percen\fage of employees who were female or non-Whi\fe Dimensions of Diversi\fy Examined Sex/gender, race/ e\fhnici\fy Sex/gender, race/ e\fhnici\fy Sex/gender Sex/gender Age, race/e\fhnici\fy Sex/gender, race/ e\fhnici\fy Mul\fiple Sex/gender, race/ e\fhnici\fy (continued) 55 Tab\fe 5. (continued) S\fudy Kim \b2005) Goode and Baldwin \b2005) Bowling, Kelleher, Jones, and Wrigh\f \b2006) Hsieh and Winslow \b2006) Alozie and Moore \b2007) Llorens, Wenger, and Kellough \b2008) Kogu\f and Shor\f \b2007) Kim \b2007) Research Ques\fion How do Asian American employmen\f pa\f\ferns differ from \fhose of o\fher agencies?

Wha\f influences African American represen\fa\fion in municipal governmen\f?

Wha\f are \fhe \frends in \fhe represen\fa\fion of women in execu\five-level pos\fs in s\fa\fe agencies?

Why does gender diversi\fy vary among racial/e\fhnic\m groups in \fhe U.S. federal workforce?

Wha\f affec\fs \fhe hiring of \mAfrican American and La\fino\m ci\fy managers?

Wha\f predic\fs represen\fa\fion of women and people of \mcolor in s\fa\fe governmen\f?

Has Affirma\five Ac\fion led \fo overrepresen\fa\fion of peopl\me of color in U.S. federal governmen\f?

Are people wi\fh disab\mili\fies represen\fed in U.S. federal governmen\f posi\fions? Con\fex\f Aggrega\fe da\fa on U.S. federal governmen\f employees Sample of 114 municipal workforces Aggrega\fe da\fa on women in execu\five managemen\f posi\fions in all 50\m s\fa\fes, 1970-2000 Aggrega\fe da\fa on U.S. federal employees and ci\fizens Sample of municipali\fies in \fhe Uni\fed S\fa\fes Workforce da\fa on all 50 s\fa\fes from 1987 \fo 2002 Aggrega\fe da\fa on U.S. federal employees and ci\fizens Aggrega\fe da\fa on U.S. federal governmen\f employees Measure of Diversi\fy Represen\fa\fion ra\fio by race/ e\fhnici\fy Percen\fage of employees who were African American Percen\fage of execu\five employees who were female Percen\fage of employees who were women or people of color Race/e\fhnici\fy of ci\fy manager Represen\fa\fion ra\fio by sex/ gender and race/e\fh\mnici\fy Represen\fa\fion ra\fio by race/ e\fhnici\fy Percen\fage of governmen\f employees wi\fh disabili\fi\mes, disaggrega\fed by race/ e\fhnici\fy and sex/g\mender Dimensions of Diversi\fy Examined Race/e\fhnici\fy Race Sex/gender Sex/gender, race/ e\fhnici\fy Race/e\fhnici\fy Sex/gender, race/ e\fhnici\fy Race/e\fhnici\fy Disabili\fy, sex/ gender, race/ e\fhnici\fy 56 Pitts and Wise 57 num\bers of men and \fomen—perfect diversity \by gender—\but \fomen make up 10% of managers and 90% of administrative support staff, there \fould \be little reason to expect performance gains. To\fard this end, the most useful studies that \fe identified in our analysis disaggregated \forkforce diversity statistics \by policy area, jo\b func- tion, or level of responsi\bility (e.g., Naff & Crum, 2000; Naff & Kellough, 2003; Pynes, 2000).Some of the research that examined diversity as an outcome focused on explaining the longitudinal trends in the employment of different groups, primarily \fomen and people of color (e.g., Charles, 2003; Kim, 2005; McCa\be & Stream, 2000; Pynes, 2000). Such studies are certainly useful in understanding \forkforce diversity trends and social progress, \but their utility is limited to conjecture a\bout \fhy they found representational differences \bet\feen groups. Other studies took the analysis a large step further to test propositions a\bout why different groups achieved representation in some organizations \but not others. Some of the determinants explored included issues in the la\bor market and political environment (e.g., Goode & Bald\fin, 2005; Kim, 2003, 2004; Llorens et al., 2008; Naff & Crum, 2000); diversity management pro- grams and Affirmative Action policies (e.g., Kellough & Naff, 2004; Kogut & Short, 2007; Naff & Kellough, 2003); and ho\f racial/ethnic identity can affect the represen- tation of \fomen (e.g., Hsieh & Winslo\f, 2006). Data on different organizational and environmental factors \fere used to test \fhich issues tended to promote (and detract) from \forkforce diversity, an approach that is argua\bly much more useful to \both scholars and practicing managers. These articles help scholars \build on common means of predicting representation and diversity, \fhich \fill promote a more compre- hensive understanding of the issues at hand. On the practical side, such studies assist managers in understanding the levers they may use to recruit more effectively from groups that are \believed to \be underrepresented. If pu\blic administration is an applied field of study that aims to produce prescriptions for pu\blic managers, it should go \beyond descriptive analyses of \forkforce trends to give pu\blic managers something concrete to act on\d.

Dive\fsity as an Indepen\cdent Va\fiable Perhaps the largest unresolved issue in \forkforce diversity research is \fhether the \business case for diversity stands up to empirical scrutiny. An ans\fer to that question can \be formulated only through testing diversity as an independent varia\ble that affects organizational outcomes, \but only eight articles in our sample that did so (Ta\ble 6).

These studies \fere more likely than others to focus on \forkforce heterogeneity, either through an index of dissimilarity (e.g., Pitts, 2005; Pitts & Jarry, 2007) or representa- tion ratio (e.g., Andre\fs, Boyne, Meier, O’Toole, & Walker, 2005; Pitts, 2007). The usefulness of the outcome measures tested in these articles varied. Three of the studies tested ho\f diversity affected perceptions of organizational outcomes or results that some may consider “performance” \but are pro\ba\bly more accurately an\feceden\fs of performance. For example, Antonova (2002) examined ho\f gender affected employee Tab\fe 6. Empirical Research linking Diversi\fy and Organiza\m\fional Ou\fcomes S\fudy Gilber\f \b2000) An\fonova \b2002) Pi\f\fs \b2005) Andrews, Boyne, Meier, O’Toole, and Walker \b2005) Meier, O’Toole, and Goerdel \b2006) Pi\f\fs \b2007) Research Ques\fion Do employee percep\fions of organiza\fional resources vary by sex, race, and managemen\f level?

Wha\f issues do women face in public ser\mvice careers in Russia?

How does \feacher racial/e\fhnic diversi\fy affec\f organiza\fional \m performance?

How does racial/e\fhnic diversi\fy in\ferac\f wi\fh managemen\f s\fra\fegy \fo influence organiza\fional performance?

Do differences be\fween male and female managemen\f s\fra\fegies\m affec\f organiza\fional \m performance?

Does racial/e\fhnic \m represen\fa\fion benefi\f organiza\fio\mnal performance?

Con\fex\f Sample of 83 federal governmen\f employees in a Sou\fhwes\fern U.S. field office Sample of 365 Russian public servan\fs Sample of 4,014 Texas public school dis\fric\fs Sample of 365 English local governmen\f au\fhori\fies Sample of 1,485 Texas public school dis\fric\fs Sample of 6,994 Texas public school dis\fric\fs Measure of Diversi\fy Sex/gender, race/e\fhnici\fy, func\fion of individual employee responden\f Sex/gender of indi\mvidual employee responden\f Blau Index of raci\mal/ e\fhnic diversi\fy Represen\fa\fion ra\fio by race/e\fhnici\fy Gender of school d\mis\fric\f superin\fenden\f Represen\fa\fion ra\fio by race/e\fhnici\fy Ou\fcome\bs) of In\feres\f Employee percep\fions of organiza\fional resources Percep\fions of equi\fy \m in \fhe organiza\fion\mal cul\fure S\fuden\f dropou\f ra\fe, percen\fage of s\fuden\fs passing a s\fandardized gradua\fion exam, average SAT score Ci\fizen sa\fisfac\fion\m Nine measures of s\fuden\f ou\fcomes Six measures of s\fuden\f ou\fcomes, broken down by racial/e\fhnic group Dimensions of Diversi\fy Examined Sex/gender, race/ e\fhnici\fy, func\fion Sex/gender Race/e\fhnici\fy Race/e\fhnici\fy Sex/gender Race/e\fhnici\fy (continued) 58 S\fudy Pi\f\fs and Jarry \b2007) Johansen \b2007) Research Ques\fion Does \fhe impac\f of racial/e\fhnic diversi\fy on performance vary by level of au\fhori\fy?

How do differences be\fween male and female managemen\f s\fra\fegies affec\f performance?

Con\fex\f Sample of 6,691 Texas public school dis\fric\fs Sample of Texas public school dis\fric\fs Measure of Diversi\fy Blau Index of racial/e\fhnic diversi\fy Gender of school d\mis\fric\f managers Ou\fcome\bs) of In\feres\f S\fuden\f dropou\f ra\fe, percen\fage of s\fuden\fs passing a s\fandardized gradua\fion exam, average SAT score S\fuden\f achievemen\f ou\fcomes Dimensions of Diversi\fy Examined Race/e\fhnici\fy Sex/gender Tab\fe 6. (continued) 59 60 Re view of Public Pe\fsonnel Administ\fation 3\b(1) perceptions of different characteristics of organizational culture, and Ivancevich and Gil\bert (2000) considered links \bet\feen diversity and perceptions of resource availa\bility. Andre\fs et al. (2005) tested the relationship \bet\feen racial/ethnic repre- sentation and citizen satisfaction \fith local government services. Given recent emphasis on treating citizens as clients or customers, su\bjective measures such as the one used in Andre\fs et al. (2005) have \become increasingly relevant as outcome metrics. The other studies in this area \fere dra\fn from the pu\blic education policy setting, and these five cases all used data from Texas pu\blic schools. This is a policy context \fith unam\biguous outcome measures, the most common \being student test scores.

Test scores argua\bly do not reflect true student competency or the quality of instruc- tion, \but they are nonetheless crucial to the assessment of performance \by elected officials, particularly in Texas. Other results used in this area included dropout rates (Pitts, 2005; Pitts & Jarry, 2007) and college readiness (Meier, O’Toole, & Goerdel, 2006). Overall, the\dse studies can \be t\daken as indirectly \dsupporting expecta\dtions \based on the anticipated linkage \bet\feen diversity and performance. These metrics are polit- ically salient and easy to measure, \but it is difficult to formulate a strong causal story a\bout \fhy they are likely to \be influenced \by employee diversity. The measures are true outcomes that are largely influenced \by external factors that may not reflect orga- nizational diversity at all. Proper specification and the inclusion of control varia\bles \become vital to ensure that \bias does not lead one to find diversity effects that actually reflect something \delse.

Data and Policy Context The 24 empirical studies that \fe have identified here reflect the lack of data avail- a\ble for use in pu\blic administration research on \forkforce diversity. Many of these articles use U.S. federal government data, either from the U.S. Office of Personnel Management’s (OPM, 2008) Federal Human Capital Survey or Central Personnel Data File (e.g., Hsieh & Winslo\f, 2006; McCa\be & Stream, 2000; Naff & Crum, 2000;). These archival data are good options for diversity research. They typically include multiple years of data, allo\fing for longitudinal research, and survey sam- pling strategies are rigorous. They also span multiple policy areas, making it easier to generalize these results than those in a single policy context. On the other hand, fed- eral agencies are enormous organizations, and it is difficult to \be sure that diversity measured at the organizational level actually filters do\fn to \fork groups. Studies that disaggregate diversity figures \by policy area or supervisory status ameliorate this to some extent, \but there is no \fay to kno\f the extent to \fhich dissimilar individuals are interacting on the\d jo\b. Five of the articles use data from Texas pu\blic schools, a source used in a num\ber of other pu\blic administration research areas. It is argua\bly difficult to generalize these findings to many other policy contexts, given that education organizations are very professionalized, highly decentralized, and dra\f on funding sources different from Pitts and Wise 61 many other government agencies (see, e.g., Pitts, 2005). Perhaps more consequential is the fact that these represent five of the eight articles testing the link \bet\feen diver- sity and performance. It is pro\blematic that \fe have so little evidence accumulating in pu\blic administration on the \business case for diversity, \but it is even more of an issue \fhen five of the eight pieces come from the same data source. A theory \base cannot develop from a single data set in one policy setting. Although there are scholarly advantages in multiple studies dra\fing from the same data\base, a disadvantage in this case is the age of the survey data on \fhich the studies are \based, particularly given the rapid changes in acceptance of diversity in recent decades (Wise & Tschirhart, 2000).The lack of data sources stems from several practical pro\blems that thus far have eluded solution. A num\ber of factors limit collection of individual-level data to ans\fer specific research questions and unaccepta\bly lo\f survey response rates often compro- mise collected data. Pu\blic organizations in the United States typically collect extensive data on employee demographics facilitating straightfor\fard tests of some diversity effects. But diversity effects cannot \be generalized from one dimension of diversity to another (Wise & Tschirhart, 2000) and given the array of salient diversity dimensions, the possi\bilities f\dor replication are\d \feakened. More confusion comes in figuring out ho\f to measure outcomes. Performance mea- surement is a persistent pro\blem in pu\blic administration that is complicated \by the political environment and goal am\biguity that many pu\blic agencies face (Boyne et al., 2007). As diversity research in the for-profit arena identifies contingencies, interac- tions, and nonlinearities in relationships \bet\feen diversity and outcomes, a host of other varia\bles \become necessary for empirical tests, and those are typically hard to come \by \fithout an original survey (Choi, in press; Pitts & Jarry, in press). Bringing diversity managemen\f into the equation most certainly requires survey data, archival research, and/or co\dntent analysis (Kel\dlough & Naff, 2004). These concerns are compounded \by the pro\blem of organizational access. To collect the necessary data, researchers must convince organizations to “let them in,” a propo- sition that is particularly shaky in an area such as diversity that is fraught \fith such important normative issues. Fe\f organizations \fant to run the risk of \being exposed as having a su\bpar diversity management program or discriminatory organizational culture. There are reasons to \be optimistic, ho\fever, given efforts \by diversity research- ers to collect ne\f data that can \be used in research. For example, Ed Kellough and Katherine Naff conducted a comprehensive survey of federal government agencies on diversity management programs, producing data that supplemented existing demo- graphic information availa\ble through other means (Kellough & Naff, 2004; Naff & Kellough, 2003). Other studies reflect original data collection in a variety of other contexts, including nonprofit organizations (e.g., Pynes, 2000), municipal govern- ments (e.g., Alozie & Moore, 2007; Goode & Bald\fin, 2005), and agencies in other countries (e.g., Antonova, 2002; Kim, 2003). Some research took advantage of exist- ing archival data produced \by the U.S. Census in creative ne\f \fays (e.g., Hsieh & Winslo\f, 2006; Llorens et\d al., 2008). 62 Re view of Public Pe\fsonnel Administ\fation 3\b(1) As researchers collect more data, they should \be mindful to include qualitative methods. All of the articles included in this analysis used quantitative methods, \fhich limits our a\bility to understand the nuanced relationships that are at \fork. Of course, many issues in the diversity sphere are \best discussed in terms of num\bers. For exam- ple, to measure \forkforce heterogeneity, it is pro\ba\bly \better to start \fith the num\bers of employees in each racial/ethnic group than \fith a ver\bal description of the mix.

The causal mechanisms underlying the relationships \bet\feen diversity, diversity management, and organizational outcomes are contingent and very complex (Foldy, 2004). Quantitative data are not likely to fully capture the dynamics. Comprehensive case studies using mixed methods have the potential \both to solve this pro\blem and to assist quantitative researchers in identifying the varia\bles that should \be included in large- N models.

Conc\fusion Using a relatively \broad definition of “diversity” \fe found evidence that interest in diversity scholarship has continued since 2000. Examination indicates that pu\blica- tion of diversity-related studies tends to \be limited to a su\bset of pu\blic administration journals, a couple of diversity dimensions, and a fe\f organizational frame\forks. A num\ber of key journals in our field pu\blished no research on diversity during the study period that fell \fithin our guidelines. A\bout half of the studies \fe found focused on issues of sex or gender and more than a third focused on race and ethnicity. Research on other diversity dimensions such as disa\bility, age, or sexual orientation \fas rare, consistent \fith Wise and Tschirhart’s earlier meta-ana\dlysis.

Workforce diversity research among scholars of pu\blic administration is popu- lated primarily \fith studies of representative \bureaucracy, \fith only a small portion of current research providing practical, action-\based findings for pu\blic managers and human resource practitioners. There is little research that \fould allo\f pu\blic sec- tors managers to step \beyond \best guesses for \fhat does and does not \fork for managing diversity. The large \body of \fork focusing on issues related to equita\ble employment outcomes is delivered \fith little information a\bout \fhich approaches have the \best results for promoting equita\ble employment practices and integrating diverse others into the pu\blic \forkforce. With a fe\f exceptions, pu\blic sector scholars are focusing on factors that managers cannot manipulate. Future research might address this shortcoming \by comparing employment outcomes at the organizational level \based on comparing the relative successes of different approaches \fith manag- ing \forkforce diversity such as training programs to reduce \bias or change attitudes, \forkplace diversity committees, formulation of affirmative action strategies and stra- tegic plans related to diversity, or efforts to promote inclusion through mentoring and net\forking. These findings need to \be replicated \by other scholars; a single study reporting a finding for or against a particular management strategy for enhancing rep- resentativeness or employee diversity is not a sufficient \basis for action and may lead to the implementat\dion of programs th\dat \fill ultimately \dfail in many organizations. Pitts and Wise 63 We did find some empirical research on diversity and organization-level results that has some potential for informing practice, \but \fe argue this research can only \be vie\fed as indirectly supporting the assumed positive relationship \bet\feen diversity and outcomes. Performance is often measured \fith antecedents and proxies that may or may not translate into \better performance or an organization’s level of effi- ciency or effectiveness. In these studies, \fe kno\f little a\bout \fhat \fas happening inside the \black \box of organizational context. This highlights t\fo empirical ques- tions that \farrant further investigation: Can \fe confirm that diversity leads to certain favora\ble organizational outcomes? What are the contextual characteristics and diver- sity management strategies in place \fhen favored outcomes \fere or \fere not attained? Our main conclusion echoes that of Wise and Tschirhart (2000) nearly 10 years ago: Diversity research has limited utility for pu\blic sector managers. Human resource managers should \be cautious a\bout the extent to \fhich they rely on research findings in deciding ho\f to address to diversity issues and \fhich programs or policies to implement. The pool of research is too shallo\f to use the results \fith confidence, and existing research is not designed to ans\fer the question “What \forks?” Similarly, studies investigating the linkage \bet\feen \forkforce diversity and organizational per- formance for several reasons fall short of \being a\ble to offer sound advice to human resource managers a\bout ho\f their organizations might leverage diversity for greater efficiency or effectiveness. One path scholars might pursue in the future to help inform practitioners a\bout \fhat \forks for managing diversity \fould \be the area of cultural competencies (Rice, 2007). Empirical studies on this topic are relatively limited \but squarely focused on the health care sector, \fhich can \be seen as an advantage for interpreting the relia\bility of findings. These studies suggest some promising effects for pu\blic service delivery and citizen satisfaction in that culturally competent pu\blic officials can provide ser- vices \fithin the context of different social systems that \better meet citizens’ needs and preferences (Weech-Maldonado, 2002). Confirming this assumption for health care \forkers and applying this research to other sectors of government \fould advance the utility of diversity\d-related research. We argue that pu\blic administration research can play a vital role \by producing \better information a\bout the impacts of \forkforce diversity on organizational out- comes. Only through empirical research can the relationships among diversity, diversity management, and organizational outcomes \become clear. More data are necessary, particularly from ne\f policy contexts and types of organizations, and performance outcomes to advance our kno\fledge a\bout the consequences of \forkplace heterogene- ity and diversity programs and policies. By understanding these complex relationships, research can provide more effective assistance to pu\blic-sector HR managers \fho are charged \fith the task of \balancing demands for equity and performance.

Dec\faration of Conf\M\ficting Interests The authors declared no conflicts of interest \fith respect to the authorship and/or pu\blication of this article. 64 Re view of Public Pe\fsonnel Administ\fation 3\b(1) Funding The authors receive\dd no financial supp\dort for the resear\dch and/or authorshi\dp of this article.\d References Adler, N. J. (2003). In\ferna\fional dimensions of organiza\fional behavior (4th ed. ). Cincinnati:

South\festern College Pu\blishing.

Alozie, N., & Moore, C. (2007). Black and Latinos in city management: Prospects and challenges in council-manager governments. In\ferna\fional Journal of Public Adminis\fra\fion , 40 , 47-63.

Andre\fs, R., Boyne, G., Meier, K., O’Toole, L. Jr., & Walker, R. (2005). Representative \bureaucracy, organizational strategy, and pu\blic service performance: An empirical analysis of English local government. Journal of Public Adminis\fra\fion Research and Theory, 15, 489-504.

Antonova, V. (2002). Women in pu\blic service in the Russian Federation: Equal capa\bility and unequal opportunit\dy. Review of Public P\mersonnel Adminis\fra\fion, 22, 216-232.

Arai, M., Wanca-Thi\bault, M., & Shockley-Zala\bak, P. (2001). Communication theory and training approaches for multiculturally diverse organizations: Have academics and practitio- ners missed the con\dnection? Public Personnel M\managemen\f, 30, 445-455.

Ball, C., & Haque, A. (2003). Diversity in religious practice: Implications of Islamic values in the pu\blic \forkplac\de. Public Personnel M\managemen\f, 32, 315.

Balser, D. (2007). Predictors of \forkplace accommodations for employees \fith disa\bilities. Adminis\fra\fion & Soc\mie\fy, 39, 656-683.

Bezrukova, K., & Jehn, K. A. (2001). The effec\fs of diversi\fy \fraining programs. Unpu\blished manuscript, Solomon Asch Center for the Study of Ethnopolitical Conflict, University of Pennsylvania.

Boyne, G. A. (2003). What is pu\blic service improvement? Public Adminis\fra\fion, 81, 211-227.

Brad\bury, M. D. (2007). The legal and managerial challenge of o\besity as a disa\bility: Evidence from the federal co\durts. Review of Public P\mersonnel Adminis\fra\fion, 27, 79-93.

Brad\bury, M. D. (2008). Parents involved in community schools v. Seattle School District No. 1: Du\bious prospects for diversity as a compelling governmental interest. Review of Public Personnel Adminis\fra\fion, 28, \m385-391.

Bre\fer, G. A., Douglas, J. W., Facer, R. L. I., & O’Toole, L. J. (1999). Determinants of graduate research productivity in doctoral of programs of pu\blic administration. Public Adminis\fra- \fion Review, 59, 1-13.

Brudney, J., Her\bert, F. T., & Wright, D. (2000). From organizational values to organizational roles: Examining representative \bureaucracy in state administration. Journal of Public Adminis\fra\fion Resea\mrch and Theory, 10, 491-512.

Carcieri, M. D. (2004). The University of Michigan affirmative action cases and pu\blic person- nel decisions. Review of Public P\mersonnel Adminis\fra\fion, 24, 70-84.

Charles, J. (2003). Diversity management: An exploratory assessment of minority group repre- sentation in state\d government. Public Personnel M\managemen\f, 32, 561-568.

Choi, S. (in press). Diversity in the U.S. Federal Government: Diversity management and employee turnover in federal agencies. Journal of Public Adminis\fra\fion Research and Theory . Pitts and Wise 65 Colvin, R. A. (2000). Improving state policies prohi\biting pu\blic employment discrimination \based on sexual ori\dentation. Review of Public P\mersonnel Adminis\fra\fion, 20, 5-22.

Colvin, R. A. (2007). The rise of transgender-inclusive la\fs: Ho\f \fell are municipalities implementing supportive nondiscrimination pu\blic employment policies? Review of Public Personnel Adminis\fra\fion, 27, 336-350.

Colvin, R. A., & Riccucci, N. M. (2002). Employment nondiscrimination policies: Assessing implementation and measuring effectiveness. In\ferna\fional Journal of Public Adminis\fra- \fion, 25, 95-108.

DeHart-Davis, L., Marlo\fe, J., & Pandey, S. (2006). Gender dimensions of pu\blic service moti- vation. Public Adminis\fra\fion Revie\mw, 66, 873.

Dolan, J. A., & Rosen\bloom, D. H. (2003). Represen\fa\five bureaucracy: Classic readings and con\finuing con\froversies. Armonk, NY: ME Sharpe.

Foldy, E. (2004). Learning from diversity: A theoretical perspective. Public Adminis\fra\fion Review, 64, 529-538.

Garcia-Zamor, J. C. (2003). Workplace spirituality and organizational performance. Public Adminis\fra\fion Revie\mw, 63, 355-363.

Gest, R., & Maranto, R. (2000). Gaining practical insights from experience: Reflections on cases of racial discrimination in federal service. Review of Public Personnel Adminis\fra\fion, 20, 55-67.

Goode, S. J., & Bald\fin, J. N. (2005). Predictors of African-American representation in munici- pal government. Review of Public P\mersonnel Adminis\fra\fion, 25, 29-55.

Griffith, K. H., & He\bl, M. R. (2002). The disclosure dilemma for gay men and les\bians: “Com- ing out” at \fork. Journal of Applied Psychology\m, 87, 1191-1199.

Guy, M. E. (1994). Organizational architecture, gender and \fomen’s careers. Review of Public Personnel Adminis\fra\fion, 14, 77-90.

Guy, M. E., & Ne\fman, M. (2004). Women’s jo\bs, men’s jo\bs: Sex segregation and emotional la\bor. Public Adminis\fra\fion Revie\mw, 64, 289.

Ham\brick, D. C., Davison, S. C., Snell, S. A., & Sno\f, C. C. (1998). When groups consist of multiple nationalities: To\fards a ne\f understanding of the implications. Organiza\fion S\fudies, 19, 181-205.

Hicks, D. A. (2002). Spiritual and religious diversity in the \forkplace: Implications for leader- ship. Leadership Quar\ferl\my, 13, 379-396.

Hindera, J. J. (1993). Representative \bureaucracy: Further evidence of active representation in the EEOC district offices. Journal of Public Adminis\fra\fion Research and Theory , 3 , 415-429.

Hsieh, C., & Winslo\f, E. (2006). Gender representation in the federal \forkforce: A comparison among groups. Review of Public P\mersonnel Adminis\fra\fion, 26, 276-295.

Human Rights Campaign. (2007). The s\fa\fe of \fhe wor\bplace. Retrieved Decem\ber 6, 2008, from http://\f\f\f.hrc.org/documents/State_o\df_the_Workplace.pdf Ivancevich, J. M., & Gil\bert, J. A. (2000). Diversity management. Time for a ne\f approach. Public Personnel M\managemen\f, 29, 75-92.

Johansen, M. (2007). The effect of female strategic managers on organizational performance. Public Organiza\fion Review, 7, 269-279.

Jung, K., Moon, J., & Hahm, S. D. (2007). Do age, gender, and sector affect employee jo\b sat- isfaction. Review of Public P\mersonnel Adminis\fra\fion, 27, 125-146. 66 Re view of Public Pe\fsonnel Administ\fation 3\b(1) Keiser, L. R., Wilkins, V. M., Meier, K. J., & Holland, C. A. (2002). Lipstick and logarithms:

Gender, institutional context, and representative \bureaucracy. American Poli\fical Science Review, 96, 553-564.

Kellough, J. E. (1990). Integration in the pu\blic \forkplace: Determinants of minority and female employment i\dn federal agencies.\d Public Adminis\fra\fion Revie\mw, 50, 557-566.

Kellough, J. E., & Naff, K. C. (2004). Responding to a \fake-up call: An examination of federal agency diversity man\dagement programs. \dAdminis\fra\fion & Soc\mie\fy, 36, 62.

Kettl, D. F. (2000). The transformation of governance: Glo\balization, devolution, and the role of government. Public Adminis\fra\fion Revie\mw, 60, 488-497.

Kim, C. (2003). Women in the Korean civil service. In\ferna\fional Journal of Public Adminis\fra- \fion, 26, 61-78.

Kim, C. (2004). Women and minorities in state government agencies. Public Personnel Man- agemen\f, 33, 165.

Kim, C. (2005). Asian American employment in the federal civil service. Public Adminis\fra\fion Quar\ferly, 28, 430-459.

Kim, C. (2007). Federal employees \fith disa\bilities \fith regards to occupation, race, and gender. Public Personnel M\managemen\f, 36, 115-125.

Kingsley, J. D. (1944). Represen\fa\five bureaucracy: An in\ferpre\fa\fion of \fhe Bri\fish civil service. Yello\f Springs, OH: \dAntioch Press.

Kochan, T., Bezrukova, K., Ely, R., Jackson, S., Joshi, A., & Jehn, K. (2003). The effects of diversity on \business performance: Report of the diversity research net\fork. Human Resource Managemen\f, 42, 3-21.

Kogut, C. A., & Short, L. E. (2007). Affirmative action in federal employment: Good intentions run amuck? Public Personnel M\managemen\f, 36, 197.

Konrad, A. M. (2003). Special issue introduction: Defining the domain of \forkplace diversity scholarship. Group & Organiza\fion Manageme\mn\f, 28, 4.

Le\fis, G. B. (1992). Men and \fomen to\fard the top: Backgrounds, careers, and potential of federal middle manag\ders. Public Personnel M\managemen\f, 21, 473-491.

Llorens, J. J., Wenger, J. B., & Kellough, J. E. (2008). Choosing pu\blic sector employment: The impact of \fages on the representation of \fomen and minorities in state \bureaucracies.

Journal of Public Adminis\fra\fion Resea\mrch and Theory, 18, 397-412.

McCa\be, B., & Stream, C. (2000). Diversity \by the num\bers: Changes in state and local govern- ment \forkforces 1980\d-1995. Public Personnel M\managemen\f, 29, 93-107.

Meier, K. J. (1993). Latinos and representative \bureaucracy: Testing the Thompson and Henderson hypotheses. Journal of Public Adminis\fra\fion Resea\mrch and Theory, 3, 393-414.

Meier, K. J., O’Toole, L. J., & Goerdel, H. (2006). Management activity and program perfor- mance: Gender as ma\dnagement capital. \dPublic Adminis\fra\fion Revie\mw, 66, 24-36.

Moynihan, D. P., & Landuyt, N. (2008). Explaining turnover intention in state government: Examining the roles of gender, life cycle, and loyalty. Review of Public Personnel Adminis- \fra\fion, 28, 120-143.

Naff, K. C. (2001). To loo\b li\be America: Disman\fling barriers for women and minori\fies in governmen\f. Boulder, CO: Westvie\f Press.

Naff, K. C. (2004). From Ba\b\be to Gru\f\fer and Gra\fz: The Supreme Court as a policymaking institution. Review of Policy R\mesearch, 21, 405-427. Pitts and Wise 67 Naff, K. C., & Crum, J. (2000). The president and representative \bureaucracy: Rhetoric and reality. Public Adminis\fra\fion Revie\mw, 60, 98-110.

Naff, K. C., & Kellough, J. E. (2003). Ensuring employment equity: Are federal diversity pro- grams making a difference? In\ferna\fional Journal of Public Adminis\fra\fion, 26, 1307-1336.

Naylor, L. A., & Rosen\bloom, D. H. (2004). Adarand, Grutter, and Gratz: Does affirmative action in federal employment matter? Review of Public Personnel Adminis\fra\fion , 24 , 150-183.

Pitts, D. W. (2005). Diversity, representation, & performance: Evidence a\bout race and ethnicity in pu\blic organizations. Journal of Public Adminis\fra\fion Research and Theory, 15 , 615-631.

Pitts, D. W. (2006). Modeling the impact of diversity management. Review of Public Personnel Adminis\fra\fion, 26, 245.

Pitts, D. W. (2007). Representative \bureaucracy, ethnicity, and pu\blic schools: Examining the link \bet\feen repres\dentation and perfor\dmance. Adminis\fra\fion & Soc\mie\fy, 39, 497-513.

Pitts, D. W. (2009). Diversity management, jo\b satisfaction, and organizational performance: Evidence from federa\dl agencies. Public Adminis\fra\fion Revie\mw, 69, 328-338.

Pitts, D. W., & Jarry, E. M. (2007). Ethnic diversity and organizational performance: Assess- ing diversity effects at the managerial and street levels. In\ferna\fional Public Managemen\f Journal, 10, 233-254.

Pitts, D. W., & Jarry, E. M. (in press). Getting to kno\f you: Time, ethnicity, and performance in pu\blic organizations. Public Adminis\fra\fion.

Pynes, J. E. (2000). Are \fomen underrepresented as leaders of nonprofit organizations? Review of Public Personne\ml Adminis\fra\fion, 20, 35-49.

Rangarajan, N., & Black, T. (2007). Exploring organizational \barriers to diversity: A case study of the Ne\f York State Education Department. Review of Public Personnel Adminis\fra\fion, 27, 249-261.

Riccucci, N. M. (2002). Managing diversi\fy in public sec\for wor\bforces. Boulder, CO: West- vie\f Press.

Riccucci, N. M. (2003). The U.S. Supreme Court’s ne\f federalism and its impact on antidis- crimination legisl\dation. Review of Public P\mersonnel Adminis\fra\fion, 23, 3-22.

Riccucci, N. M., & Gossett, C. W. (1996). Employment discrimination in State and local gov- ernment: The les\bian and gay male experience. American Review of Public Adminis\fra\fion, 26, 175-192.

Riccucci, N. M., & Meyers, M. (2004). Linking passive and active representation: The case of frontline \forkers in \felfare agencies. Journal of Public Adminis\fra\fion Research and Theory, 14, 585-597.

Riccucci, N. M., & Saidel, J. R. (1997). The representativeness \d of state-level \bureaucratic leaders: A missing piece of the representative \bureaucracy puzzle. Public Adminis\fra\fion Review, 57, 423-430.

Riccucci, N. M., & Saidel, J. R. (2001). The demographics of gu\bernatorial appointees: To\fard an explanation of \dvariation. Policy S\fudies Journ\mal, 29, 11-22.

Rice, M. (2007). A post-modern cultural competency frame\fork for pu\blic administration and pu\blic service deliv\dery. In\ferna\fional Journa\ml of Public Sec\for \mManagemen\f, 20, 622-637.

Ru\baii-Barrett, N., & Wise, L. R. (2007a). From \fant ads to We\b sites: What diversity messages are state governmen\dts projecting? Review of Public P\mersonnel Adminis\fra\fion, 27, 21-38. 68 Re view of Public Pe\fsonnel Administ\fation 3\b(1) Ru\baii-Barrett, N., & Wise, L. R. (2007\b). Language minorities and the digital divide. Journal of Public Managemen\m\f and Social Policy\m, 12, 5-27.

Selden, S. C. (1997). The promise of represen\fa\five bureaucracy: Diversi\fy and responsiveness in a governmen\f age\mncy. Armonk, NY: M. E. Sharpe.

Selden, S., & Selden, F. (2001). Rethinking diversity in pu\blic organizations for the 21st cen- tury: Moving to\fard \da multicultural mo\ddel. Adminis\fra\fion & Soc\mie\fy, 33, 303.

Sisneros, A. (2004). Marching in the procession of precaution? A re\buttal to Martin D. Carcieri’s “The University of Michigan affirmative action cases and pu\blic personnel decisions.” Review of Public Personne\ml Adminis\fra\fion, 24, 175-183.

So\fa, J., & Selden, S. (2003). Administrative discretion and active representation: An expan- sion of the theory of representative \bureaucracy. Public Adminis\fra\fion Review, 63, 700-708.

Stackman, R., Connor, P., & Becker, B. (2005). Sectoral ethos: An investigation of the personal values systems of female and male managers in the pu\blic and private sectors. Journal of Public Adminis\fra\fion Resea\mrch and Theory, 16, 577-597.

Thomas, R. R. (1990). From affirmative action to affirming diversity. Harvard Business Review, 107-117. Retrieved Octo\ber 2, 2009, from http://\f\f\f.runet.edu/~kvharri\dng/docs/ HRMDocs/3affirmingdiv.pdf U.S. Office of Personnel Management. (2008). Federal employmen\f s\fa\fis\fics. Retrieved Decem\ber 6, 2008, f\drom http://\f\f\f.opm.gov/feddata/ind\dex.asp Von Bergen, C. W., Soper, B., & Foster, T. (2002). Unintended negative effects on diversity management. Public Personnel M\managemen\f, 31, 239-252.

Weech-Maldonado, R. (2002). Racial/ethnic diversity and culture competency: The case of Pennsylvania hospit\dals. Journal of Heal\fhcar\me Managemen\f, 47, 111-124.

West, J. P., & Berman, E. M. (1996). Managerial responses to an aging municipal \forkforce. Review of Public P\mersonnel Adminis\fra\fion, 16, 38-51.

Wilkins, V. M. (2006). A mixed \bag: The Supreme Court’s ruling on the ADEA and disparate impact. Review of Public P\mersonnel Adminis\fra\fion, 26, 269-274.

Wilkins, V. M. (2007). Exploring the causal story: Gender, active representation, and \bureau- cratic priorities.\d Journal of Public Adminis\fra\fion Resea\mrch and Theory, 17, 77-94.

Wilkins, V. M., & Keiser, L. R. (2006). Linking passive and active representation \by gender: The case of child support agencies. Journal of Public Adminis\fra\fion Research and Theory, 16, 87-102.

Williams, K. Y., & O’Reilly, C. A. (1998). Demography and diversity in organizations: A revie\f of 40 years of rese\darch. Research in Organiza\fional Behavi\mor, 20, 77-140.

Wise, L. R. (1990). Social equity in civil service systems. Public Adminis\fra\fion Review, 50 , 567-575.

Wise, L. R. (1994). Factors affecting the size of performance a\fards among mid-level civil ser- vants in the United States. Public Adminis\fra\fion Quar\ferly , 18 , 260-278.

Wise, L. R., & Tschirhart, M. (2000). Examining empirical evidence on diversity effects: Ho\f useful is diversity research for pu\blic-sector managers? Public Adminis\fra\fion Review, 60, 386-394.

Zeigler, S. (2006). Litigating equality: The limits of the equal pay act. Review of Public Personnel Adminis\fra\fion , 26 , 199-215. Pitts and Wise 69 Bios David W. Pitts ([email protected]\d), PhD, is an assistant professor in the School of Pu\blic Affairs at American University, Washington, DC. His research focuses on \forkforce diversity and pu\blic-sector h\duman resources man\dagement.

Lois Recascino Wise (\[email protected]),\d PhD, is a professor in the School of Pu\blic and Environmental Affairs and Director of West European Studies and the European Union Center at Indiana University, Bloomington, IN. Her research centers on comparative administrative reform and includes\d a special focus o\dn the pu\blic sector\d and managing for d\diversity.