LDRSHP Week 5

Managing Diversity in U.S. Federal Agencies 109

Sungjoo Choi is an assistant professor

in the Department of Political Science at

Kennesaw State University. She received her

doctorate from the University of Georgia.

Her research interests include diversity

management, organizational justice, perfor-

mance management in public organizations,

and comparative public administration.

E-mail: [email protected]

Hal G. Rainey is Alumni Foundation

Distinguished Professor in the School of

Public and International Affairs at the

University of Georgia. His book Understand-

ing and Managing Public Organizations was

published in 2009. This year, he received the

Dwight Waldo Award from the American

Society for Public Administration.

E-mail: [email protected]

Diversity in the workplace is a central issue for contemporary organizational management. Concomitantly, managing increased diversity deserves greater concern in public, private, and nonprofi t organizations. Th e authors address the eff ects of diversity and diversity management on employee perceptions of organizational performance in U.S. federal agencies by developing measures of three variables: diversity, diversity management, and perceived organizational performance. Drawing from the Central Personnel Data File and the 2004 Federal Human Capital Survey, their fi ndings suggest that racial diversity relates negatively to organizational performance. When moderated by diversity management policies and practices and team processes, however, racial diversity correlates positively with organizational performance. Gender and age diversity and their interactions with contextual variables produce mixed results, suggesting that gender and age diversity refl ect more complicated relationships. Th is article provides evidence for several benefi ts derived from eff ectively managing diversity.

The American workforce has been increasingly diversifi ed by greater access to jobs for women and minorities. Diversity and representation “politically integrate a diverse nation with a measure of legitimacy” (Brewer 2002, 1), but also enhance social justice (Kellough 1990; Krislov and Rosen- bloom 1981). Understanding the impacts of diversity on organizational outcomes, such as organizational performance, employee satisfaction, and turnover, has become essential (Milliken and Martins 1996). Cox (1993) contends that managing diversity should be signifi cant as a moral imperative, as a legal requirement, and as a factor in organizational performance. In recent years, diversity scholars have reached consensus that the perspective on diversity manage- ment has broadened from

pursuing affi rmative action programs to taking advan- tage of diff erences to improve organizational eff ective- ness (Wise and Tschirhart 2000). Recent research has investigated diversity in relation to organizational eff ectiveness.

Public organizations, through equal employment opportunity (EEO) and affi rmative action programs, have been more committed to workforce diversity than have private organizations, resulting in a higher level of diversity in public organizations (Cornwell and Kellough 1994; Foldy 2004; Riccucci 2002). Public organizations thus face the challenge of managing a diversifi ed workforce, a challenge that calls for more research on the impacts of diversity on organizational eff ectiveness (Wise and Tschirhart 2000). In spite of the greater diversity in government organizations, more research in the fi eld of business management has examined the eff ects of diversity on organizational outcomes. Inconsistent empirical evidence regarding the eff ects of diversity on organiza- tional outcomes suggests a more complex relationship between diversity and organizational eff ectiveness than is often realized (Jackson, Joshi, and Erhardt 2003; Milliken and Martins 1996; Williams and O’Reilly 1998; Wise and Tschirhart 2000). Recent research indicates that mediating or moderating eff ects of contextual factors, such as organizational culture, and demographic characteristics of group members and supervisors, explain the inconsistent research results.

Th e present study shows how managerial eff orts and other contextual variables moder- ate the relationship between diversity and organizational outcomes. First, this article reviews the literature on diversity in work groups and its impacts on organizational outcomes. Th en, it discusses theoretical arguments about the

Sungjoo Choi Kennesaw State University

Hal G. Rainey University of Georgia

Managing Diversity in U.S. Federal Agencies: Eff ects of Diversity and Diversity Management on Employee

Perceptions of Organizational Performance

… [Th is] study shows how managerial eff orts and other

contextual variables moderate the relationship between

diversity and organizational outcomes.

Current Trends in Public Personnel Administration

110 Public Administration Review • January | February 2010

impacts of diversity and diversity management. Next, it examines the eff ects of diversity in race, gender, and age, and of moderators such as diversity management, organizational culture, and team processes on individual- and organizational-level outcomes using a hierarchical regression analysis. Th e last section discusses fi ndings and implications.

Literature on Diversity Early studies of diversity focused on its eff ects on individual and organizational outcomes, indicating that it provides a great oppor- tunity for organizations as well as a challenge (Milliken and Martins 1996; Tsui, Egan, and Xin 1995). Some studies concluded that more heterogeneous work groups consider more perspectives and produce higher-quality solutions than homogeneous groups (Cox, Lobel, and McLeod 1991; McLeod and Lobel 1992; Milliken and Martins 1996; Watson, Kumar, and Michaelsen 1993). Others con- tended that heterogeneous groups show lower levels of integration and have higher levels of dissatisfaction and turnover than homoge- neous groups (Jackson, Joshi, and Erhardt 1993; O’Reilly, Caldwell, and Barnett 1989).

Two perspectives have shaped research on the impacts of diversity (Williams and O’Reilly 1998). One perspective, based on infor- mation and decision-making theories, argues that diversity can benefi t organizations by providing a broad range of ideas, skills, and insights that can improve organizational capabilities to solve prob- lems and make better decisions (Cox 1993; Cox and Blake 1991; Ely 2004). Th e other perspective, based on social categorization and social identity theories (Turner 1987) and the similarity-attraction paradigm (Byrne 1971), contends that diversity may burden organi- zations with high costs of coordination and confl ict resolution, thus compromising organizational eff ectiveness, given that people tend to distinguish between in-group and out-group members, which may cause confl icts and mis- communication (Ely 2004). Although both of these perspectives are supported by research in laboratory settings (Ely 2004; Williams and O’Reilly 1998), the external validity of these fi ndings is questionable, suggesting that researchers should take the eff ects of contex- tual factors into consideration, and seek more sources of evidence outside of the laboratory.

Inconsistent empirical evidence on the impacts of diversity on organizations suggests complex relation- ships between diversity and its consequences. Recently, diversity researchers have developed models to examine the potential impacts of contextual factors that may moderate the relationship between diversity and its outcomes, such as human resources practices, team processes, and organizational culture (Jackson, Joshi, and Erhardt 2003; Kochan et al. 2003; Milliken and Martins 1996; Pitts 2005; Williams and O’Reilly 1998). Williams and O’Reilly (1998), in a review of 40 years of diversity research, proposed that contextual variables should be more carefully examined. Milliken and Martins (1996), in their evaluation of six years (1989–94) of research, iden- tifi ed four types of mediating variables, which are the short-term consequences of diversity that infl uence the long-term outcomes of diversity. More recently, Jackson, Joshi, and Erhardt (2003) reviewed 63 studies published between 1997 and 2002 and found

that contextual factors, such as task characteristics, organizational culture, team processes, and strategic context, signifi cantly aff ected the relationship between diversity and organizational performance or behavior.

Nevertheless, the eff ects of other variables, such as diversity man- agement practices and leadership, need much more attention from researchers. Previous research has used a limited range of samples, such as private fi rms in some fi elds of business or laboratory studies, thereby raising questions about external validity. Although some recent studies examined diversity’s impacts on the organizational eff ectiveness of public organizations, they also used samples drawn from limited areas of government. Th e present study addresses these limitations of previous research, focusing on eff ective management of increased diversity in organizations rather than on the direct eff ects of diversity.

Th e study also draws on a very large sample of actual federal employees. Th e results show that the management of diversity signifi cantly moderates diversity’s impacts on important outcomes at individual and organizational levels. Some diversity scholars, including Cox (1993), Ely (2004), and Foldy (2004), have argued for the importance of eff ective diversity management. Pitts (2009) examined the eff ects of diversity management on federal employ- ees’ perceptions of organizational performance. However, his study relied entirely on survey evidence, and reported no evidence about the levels of diversity in the federal agencies in which the federal employees were located. Th erefore, he could provide no evidence about the relationship between levels of diversity and diversity management. He did not show, as does our present analysis, that diversity management can actually reduce the potential negative impacts of increased levels of diversity and enhance positive impacts

of diversity, which is the goal of diversity management policies and practices. Pitts’s study found that federal employees, especially minority employees, perceived their agen- cies to be more eff ective when they perceived eff ective diversity management. Th e study we report here provides a more advanced analysis of the eff ects of diversity management. We measure the levels of diversity of 67 federal agencies and analyze how diversity manage- ment aff ects the relationship between levels of diversity and perceptions of organizational

performance. We need more evidence about how diversity manage- ment aff ects diversity’s impacts on organizational eff ectiveness. Th is study provides such evidence.

Theoretical Background and Hypotheses

Direct Effects of Diversity Some proponents of diversity have maintained that workforce diver- sity enhances organizational eff ectiveness and productivity (e.g., Ely 2004; Hambrick and Mason 1982; McLeod and Lobel 1992; Richard 2000; Wiersema and Bantel 1992). Ely’s (2004) informa- tion and decision-making theories suggest that diversity improves performance by contributing to higher-quality decisions and by taking advantage of a broader range of alternatives and new ideas (Cox 1994; Foldy 2004; Ospina 2001; Richard 2000). Some studies

We measure the levels of diversity of 67 federal agencies

and analyze how diversity management aff ects the

relationship between levels of diversity and perceptions of organizational performance.

Managing Diversity in U.S. Federal Agencies 111

have found that work teams composed of people with diff erent backgrounds tended to share more information, leading to better performance than homogeneous teams (Bunderson and Sutcliff e 2002). Greater heterogeneity may lead to low consensus in decision making. Some researchers contend, however, that this can improve problem solving (Dutton and Duncan 1987; Katz 1982).

On the other hand, some research has reported a negative or insig- nifi cant relationship between diversity and organizational perfor- mance (e.g., Chatman and Flynn 2001; Chatman et al. 1998; Foldy 2004; Gladstein 1984; Jehn 1995; Jehn, Northcraft, and Neale 1999). Heterogeneous groups may experience problems in integra- tion, coordination, motivation, and confl ict management (Gladstein 1984; Jehn 1995; Jehn, Northcraft, and Neale 1999). Consequently, diverse work groups typically take more energy to accomplish tasks than homogeneous work groups (Greenhaus, Parasuraman, and Wormley 1990; Kanter 1979; Schneider and Northcraft 1999). Research has indicated that individuals from diverse groups can experience exclusion from the group’s internal information networks and from important decision-making processes (Cox 1994; Ibarra 1993; Pettigrew and Martin 1989).

Similarity-attraction theory (Byrne 1971) and social categorization and social identity theories (Turner 1987) support negative perspec- tives on diversity in work groups because people prefer similarity in their interactions (Schneider 1987; Tsui, Egan, and O’Reilly 1992). Individuals feel comfortable in a group they perceive as similar to themselves. Individuals in diverse groups feel less safe and trust each other less. Lower trust increases confl ict within groups. Th eories of selection and socialization view similarity in values and demo- graphic characteristics as conducive to eff ective work environments (Jehn, Northcraft, and Neale 1999).

Social categorization and social identity theories propose that people classify themselves and others into social categories based on social status and membership of social groups (Ashforth and Mael 2001; Tajfel and Turner 1985; Turner 1987). People in a group sharing the same social identity distinguish members of their own social category from others (Ely 2004; Kramer 1991). Such distinctions decrease interaction within or between groups from diff erent social categories and leads to discrimination and self-segregation. Th us, some research fi nds that diversity is associated with less social inte- gration, more confl ict, and less cohesion in groups (Milliken and Martins 1996; Williams and O’Reilly 1998). Diversity in organiza- tions is therefore likely to decrease organizational eff ectiveness.

H1: Higher levels of workforce diversity will decrease organi- zational performance.

Effects of Diversity Management Eff ective diversity management enhances the eff ects of diversity on organizational performance. Research has shown mixed fi ndings about the impacts of diversity in tenure, race, and sex on organi- zational performance. When properly managed, however, diversity can enhance performance (Cox, Lobel, and McLeod 1991; O’Reilly, Williams, and Barsade 1997). Human resource management poli- cies that promote diversity can enhance the diversity–performance relationship (Jehn and Bezrukova 2004; Richard 2000; Richard and Johnson 2001). An organizational culture and managerial attitudes

supporting diversity can also improve performance (e.g., Jehn and Bezrukova 2004; O’Reilly, Williams, and Barsade 1997).

Th e integration and learning perspective proposes that organizations should incorporate employees’ diverse perspectives into organiza- tional tasks and provide opportunities to learn from diff erences (Th omas and Ely 1996). Diversity provides a source of growth, learning, and intuition, but only when it is properly managed. Foldy (2004) contended that the integration and learning perspective (Th omas and Ely 1996, 2001) should link to high learning frames and behaviors (Argyris and Schön 1996). High learning frames and behaviors make three important contributions: (1) they support the perspective that diversity is a source of learning, (2) they call for cultural diversity to be discussed openly, and (3) they ask every member in a group to embrace diversity (Foldy 2004).

Numerous scholars have emphasized the important role of manag- ers in diversity management. Th omas and Ely (1996) suggested that eff ective leadership connects diversity to work eff ectiveness. Eff ective leadership encourages a diverse workforce to appreciate diff erent perspectives and to value diverse opinions and ideas (Th omas and Ely 1996).

H2a: Eff ective diversity management will increase organiza- tional performance.

H2b: Diversity management will moderate the impact of diversity on organizational performance. In agencies that have more eff ective diversity management, higher levels of diversity will increase perceived organizational performance. In agencies that have less eff ective diversity management, higher levels of diversity will decrease perceived organizational performance.

Organizational Tenure Pfeff er (1983, 323) asserted that organizations will be most pro- ductive when employees have organizational tenure “long enough to overcome some initial naiveté and learn the ropes and local practices.” As diverse teams work together longer, teamwork and collaboration increase. Enhanced intragroup contact will reduce the negative eff ects of social categorization (Chatman and Flynn 2000; Ely 2004; Harrison et al. 2002). Th e negative eff ects of surface-level diversity, such as demographic diff erences, become less important, reducing prejudices and stereotyping. Katz (1982) argued that longer organizational tenure stabilizes the organization, reducing goal confl ict. Decreased negative eff ects of diversity imply that organizations can invest more in the improvement of organizational performance, while saving resources that would have been expended for coordination, confl ict management, and control.

H3a: Longer organizational tenure will increase organizational performance.

H3b: Organizational tenure will moderate the impact of di- versity on organizational performance such that, for agencies that have longer average tenure of employees, higher levels of diversity will increase organizational performance. For agen- cies that have shorter average tenure, higher levels of diversity will decrease organizational performance.

112 Public Administration Review • January | February 2010

Effect of Team Processes Individual diff erences may lead to excessive confl ict, so it is neces- sary to develop ways to counteract these eff ects. Relational confl icts, when not properly treated, decrease team members’ satisfaction (De Dreu and Weingart 2003). Team processes can moderate the relationship between diversity and confl ict, so that cooperation and communication among team members mitigate the negative eff ect of diversity (e.g., Ely 2004; Mohammed and Angell 2004). Mohammed and Angell (2004, 1021) argued that team processes enhance interactions between team members (Marks, Mathieu, and Zaccaro 2001) and the synergistic combination of individual eff orts (Kozlowski and Bell 2003). Team processes should moderate the negative eff ects of diversity on organizational performance by reducing confl ict.

H4a: Eff ective team processes will increase organizational performance.

H4b: Eff ective team processes will moderate the impact of di- versity on organizational performance, such that for agencies that have more eff ective team processes, diversity will increase organizational performance. For agencies that have less ef- fective team processes, diversity will decrease organizational performance.

Effect of Organizational Culture Organizational culture is defi ned as “a common set of shared mean- ings or understandings about an organization (e.g., strongly held values, and norms about patterns of group members’ behaviors)” (Chatman and Jehn 1994; Jehn and Bezrukova 2004, 705–6; Reichers and Schneider 1990; Triandis and Suh 2002). Organi- zational culture may aff ect the relationship between diversity and performance by imbuing group members with the organization’s perspectives on diversity (Jehn and Bezrukova 2004).

A results-oriented organizational culture emphasizes individual accomplishment, usually leading to high competition among mem- bers in the organization. High competition may prevent individuals from working effi ciently, leading to ineffi ciency (Jehn and Mannix 2001; Tsui, Egan, and O’Reilly 1992). Furthermore, the organiza- tion may lose opportunities to benefi t from cooperation because of the competitive culture. Diversity in results-oriented organizational culture may decrease organizational performance.

H5a: Results-oriented organizational culture will increase organizational performance.

H5b: Results-oriented organizational cul- ture will moderate the impact of diversity on organizational performance, such that for agencies that emphasize results-oriented culture, diversity will decrease organiza- tional performance.

Data and Methods

Data Sources and Sample The present study uses data from two major sources: the Central Personnel Data File

(CPDF) and the 2004 Federal Human Capital Survey (FHCS), both published by the U.S. Office of Personnel Management.1 The diversity measures for federal agencies were created using information from the CPDF in 2004. Index variables were developed using the questionnaire of the 2004 FHCS. A total of 67 subagencies under executive branch agencies and inde- pendent agencies participated in the 2004 FHCS, and approxi- mately 150,000 federal government employees responded to the survey.

Diversity Measure Th ree kinds of diversity measures, including the coeffi cient of variation, the Blau index of heterogeneity, and the entropy index of diversity, are the most frequently used measures. Diversity researchers use diff erent measures depending on the characteris- tics of variables. Th e Blau index of heterogeneity and the entropy index are appropriate for categorical variables such as gender, race/ethnicity, and functional and educational background, while the coeffi cient of variation is used for continuous variables such as age and tenure. Th e present study employs the entropy index of diversity to measure race, gender, and age diversity, which are categorical variables in this study. Th e entropy index of diversity, or the Shannon index (uncertainty index), is considered the appro- priate measure of diversity (Miller and Quigley 1990; Pielou 1977; Reardon 1998). Th e entropy index is computed with the following formula:

E = ∑ Qr logn (1/Qr) (r = 1, . . . , n),

where Qr denotes the proportion of the population of group r. Th e entropy index of diversity ranges from a minimum value of 0, which means that the population contains a single group, to a maximum value of 1, which means that all n groups are evenly represented in the population.

Th e indices of each agency’s attributes of diversity—race, gender, and age—are calculated using the entropy index of diversity.2 Our analysis classifi es federal employees into fi ve racial/ethnic groups— American Indian or Alaska Native, Asian or Pacifi c Islander, blacks, Hispanics, and whites—based on the CPDF’s categorization. Gender of federal employees is composed of two groups: male and female. Th e age variable is categorized into fi ve groups: under 29, 30–39, 40–49, 50–59, and over 60.

Dependent Variables Th e appendix presents the questionnaire items for the scales constructed for this study. Th e measure of perceived organiza-

tional performance was developed by using principal components factor analysis with varimax rotation. It includes survey questions about the quality of work, organizational performance, managerial capability, and job-relevant resources and skills. Respondents rated these four questions on a Likert scale from 1 (strongly disagree or very poor) to 5 (strongly agree or very good). Factor analysis of these items shows that factor loadings range between 0.715 and 0.794. Th e scale’s eigen- value is 2.280 and Cronbach’s alpha is 0.738.

Th e present study uses data from two major sources:

the Central Personnel Data File (CPDF) and the 2004

Federal Human Capital Survey (FHCS), both published by the U.S. Offi ce of Personnel

Management.

Managing Diversity in U.S. Federal Agencies 113

Main Independent Variable We measure diversity management in two ways. One method mea- sures the diversity management of agencies according to employees’ perceptions. Th is study develops an index variable through factor analysis, using three survey questions from the 2004 FHCS, that ask about commitment to diversity and policies and practices to promote diversity (see the appendix). Principal components factor analysis and varimax rotation produce a single factor on which these three items loaded. Th e initial eigenvalue of the scale is 2.308 and the Cronbach’s alpha is 0.850. Factor loadings range between 0.868 and 0.886.

Th e other way of measuring the eff ectiveness of diversity manage- ment uses the number of EEO complaints in an agency as a proxy measure of eff ective diversity management. Th e Notifi cation and Federal Employee Antidiscrimination and Retaliation (No FEAR) Act of 2002 requires every federal agency to report complaints about violations of antidiscrimination and whistle-blower protec- tion laws on a quarterly basis.4 Under section 301 of the No FEAR Act, each federal agency must post summary statistical data pertain- ing to complaints of employment discrimination fi led against it by employees, former employees, and applicants for employment under 29 C.F.R. Part 1614 (EEOC 2005). We assume that the eff ectiveness of an agency’s diversity management relates negatively to complaints reported, and that a lower number of EEO com- plaints represents more eff ective diversity management. Th e data come from the 2004 annual report posted on each agency’s public Web site. Th e measure divides the number of EEO complaints by organizational size.

Control Variables Performance-oriented organizational culture. A results-oriented organizational culture emphasizes employees’ performance and productivity by rewarding high performance and penalizing poor performance. Such a culture emphasizes merit-based promotions, rewards, or penalties based on performance evaluations. Th e appen- dix shows the eight items that are integrated into a single factor, with factor loadings between 0.685 and 0.888, through principal factor analysis and varimax rotation. Th e initial eigenvalue of the scale is 5.736 and the Cronbach’s alpha is 0.942.

Team processes. Th is variable measures cooperation and communi- cation in work units. Two questions ask about employees’ coopera- tion to get the job done in work groups and about information and knowledge sharing among employees through communication. Th ey were rated on a fi ve-point Likert scale. Th ese items are also integrated into one index variable through factor analysis. Th e initial eigenvalue is 1.462 and the Cronbach’s alpha is 0.626.

Average tenure. Average tenure represents the average tenure of full-time employees hired in an agency. Th is variable is measured for each agency based on the 2004 CPDF. Th e mean average tenure of federal agencies is 15.1 years, ranging between 7.9 and 22.6 years.

Demographic variables. Demographic variables including gender, minority, and supervisory status are recorded as dummy variables in order to control their eff ects on dependent variables. Previous research has shown that diff erences in gender, minority status, and supervisory status aff ect perceptions of organizational performance. We control for the impacts of these demographic characteristics on dependent variables. Th e gender variable is coded 1 when a respondent is a female and 0 for a male. Th e minority variable is 0 when a respondent is a white male or a white female and is 1 when a respondent is neither a white male nor a white female. Th e CPDF defi nes American Indian or Alaska Native, Asian or Pacifi c Islander, black (not of Hispanic origin), and Hispanic as the minority race/ national origin groups for federal statistics and program adminis- trative reporting (OPM 2006). Supervisory status is classifi ed into two types—supervisor or manager and executive—represented by two dummy variables. For a supervisor or a manager, the supervisor variable is recorded as 1, and when a respondent is an executive, the executive variable is recorded as 1. Otherwise, it is recorded as 0.

Methodology and Model Specifi cation We present hierarchical regressions to test a model of the eff ects of diversity. Moderated multiple regression involves hierarchical regression that fi rst tests the relationship between the predictors (independent variables) of interest and the criterion variable (depen- dent variable), and then tests the relationship of a term that carries information about both predictors (the interaction term). Th e “hier- archical” form of regression indicates that predictors are not entered

Table 1 Descriptive Statistics

Variables Mean Std. Dev. Min Max Unit

Size 11.23 1.83 6.22 13.39 Natural log of the number of employees

Organizational tenure 15.11 2.72 7.9 22.6 Year

Female 0.43 0.50 0 1 Female = 1, male = 0

Supervisor 0.28 0.45 0 1 Supervisor or manager = 1

Executive 0.03 0.18 0 1 Executive = 1

Minority 0.30 0.46 0 1 Minority = 1

Race diversity 0.56 0.08 0.21 0.77 Entropy index

Gender diversity 0.95 0.04 0.81 1 Entropy index

Age diversity 0.89 0.03 0.73 0.96 Entropy index

Diversity management 0 1 –3.07 1.44 Factor score

EEO complaints 0.60 0.28 0.15 1.81 Proportion of the number of EEO complaints

Team processes 0 1 –3.96 1.28 Factor score

Organizational culture 0 1 –2.20 1.87 Factor score

114 Public Administration Review • January | February 2010

into the regression equation simultaneously, but in a logical order. Typically, the continuous predictor and the polychotomous predic- tor are entered in the fi rst step, and the interaction term is entered in the second step (Aiken and West 1991). In addition, researchers decide how many predictors to enter and the order in which they are entered. Determining the order of entry is based on logical or theoretical considerations. F-tests are used to compute the signifi - cance of each added variable (or set of variables) to the explanation refl ected in R2 (Garson 1998).

We test the main and moderating eff ects of diversity and diversity management on organizational performance using this method. Some researchers (e.g., Allison 1977; Friedrich 1982; Smith and Sasaki 1979; Wright 1976) have expressed the concern that because a multiplicative term is often highly correlated with its constitu- ent variables, the inclusion of the multiplicative term in a model with its constituent variables leads to multicollinearity, which yields unreliable and unstable coeffi cients. In fact, when we tested our model using the ordinary least squares method, we detected such a problem in the interaction terms of our model. For this reason, many studies in organizational behavior and social psychology (e.g., Aiken and West 1991; Cohen 1978, 861; Cohen and Cohen 1983; Simons and Peterson 2000) have used a hierarchical regression procedure to resolve the problem, in which the eff ect of a multipli- cative term on a dependent variable should be assessed only after the eff ects of its constituent variables on the dependent variable have been partialed out (Friedrich 1982, 802).

Further, larger correlations between predictors can also increase multicollinearity. When the predictors are highly correlated, a statistically signifi cant interaction term may result from a nonlinear multiplicative eff ect, and not because of a linear multiplicative eff ect (Friedrich 1982). Nevertheless, correlations between predictors that compose interaction terms are not substantially high, suggesting that multicollinearity related to interaction terms may not be a criti- cal problem in this model.

Th e model of diversity and organizational performance categorizes the independent variables into eight sets and enters them in the following order: demographic control variables (step 1), the main eff ects of diversity (step 2), the main eff ects of diversity manage- ment (step 3), the main eff ects of team processes and organizational culture (step 4), the moderating eff ects of diversity management (step 5), the moderating eff ects of tenure (step 6), the moderat- ing eff ects of team processes (step 7), and the moderating eff ects of organizational culture (step 8). Th e model of organizational performance is specifi ed as follows:

Step 1: Pi = β0 + β1Ci + ei Step 2: Pi = β0 + β1Ci + β2Di + ei Step 3: Pi = β0 + β1Ci + β2Di + β3Mi + β4Ei + ei Step 4: Pi = β0 + β1Ci + β2Di + β3Mi + β4Ei + β5Ci + β6Ti + β7Ri + ei Step 5: Pi = β0 + β1Ci + β2Di + β3Mi + β4Ei + β5Ci + β6Ti + β7Ri + β8Di * Mi + β9Di * Ei + ei Step 6: Pi = β0 + β1Ci + β2Di + β3Mi + β4Ei + β5Ci + β6Ti + β7Ri + β8Di * Mi + β9Di * Ei + β10 * Yi + ei

Step 7: Pi = β0 + β1Ci + β2Di + β3Mi + β4Ei + β5Ci + β6Ti + β7Ri + β8Di * Mi + β9Di * Ei + β10 * Yi + β11Di * Ti + ei Step 8: Pi = β0 + β1Ci + β2Di + β3Mi + β4Ei + β5Ci + β6Ti + β7Ri + β8Di * Mi + β9Di * Ei + β10 * Yi + β11Di * Ti + β8Di * Ri + ei

where Pi is perceived organizational performance; Ci is the vector of organizational contextual variables (organizational size, average tenure of employees) and demographic characteristics of employees (female, minority status, and supervisory status); Di is the vector of attributes of diversity (race diversity, gender diversity, and age diversity); Yi is organizational tenure; Ti represents team processes; Ri denotes results-oriented organizational culture; Mi is perceived diversity management; Ei is the percentage of reported EEO complaints; and ei is an error term.

Results Th e main eff ects of diversity, diversity management, and other contextual variables and moderating eff ects on organizational per- formance were examined through the eight steps of the hierarchical regression analysis. Th e model fi t of each of the eight blocks is good and the complete model explains 57 percent of the variation in the dependent variable.

Diagnostic tests for normality, linearity, multicollinearity, homosce- dasticity, and the existence of outliers did not indicate serious violations of the assumptions for the ordinary least squares regres- sion, except for one minor problem. Residual analyses detected one outlier, but it does not appear to infl uence the relationship signifi - cantly. Table 3 describes the correlations of variables. Th e largest correlation is 0.66, and the average and median correlations are 0.03 and 0.01, respectively. Th e variance infl ation factor (VIF) test for multicollinearity revealed that the largest VIF value and the average were 1.95 and 1.40 respectively, much lower than the typical cutoff point of 3.5.

Direct Effects of Diversity (H 1 )

Only the eff ect of racial diversity on perceived organizational performance (hereafter, organizational performance) was supported. Racial diversity (–.42, p < .01) is negatively related to organizational performance, in support of hypothesis 1, suggesting that in agencies that have higher levels of racial diversity, organizational performance

Table 2 Summary of Model Fitç

Block F R2 Change in R2

1 224.09*** .0171

2 39.01*** .0186 .0015

3 18090.81*** .3319 .3132

4 21247.24*** .5691 .2373

5 8.67*** .5694 .0003

6 7.00*** .5695 .0001

7 3.85** .5696 .0001

8 5.15*** .5697 .0001

** Signifi cant at .05 level; *** signifi cant at .01 level.

Managing Diversity in U.S. Federal Agencies 115

was perceived lower than that in other agencies. On the other hand, gender diversity (.31, p < .01), which indicates a better balance of the two sexes, is positively related to organizational performance. In agencies where the proportions of male and female employees are similar, organizational performance was perceived higher than it was in others. Th e eff ect of age diversity on organizational performance was not signifi cant.

Effects of Diversity Management (H 2 )

Findings supported our hypotheses of the eff ects of diversity man- agement on organizational performance. We developed the two measures of diversity management—perceived diversity manage- ment and EEO complaints, which are negatively related to each other. Higher levels of perceived diversity management indicate that the agency manages workforce diversity more eff ectively, while higher levels of EEO complaints reported to the agency’s EEO offi ce mean that the agency manages diversity less eff ec- tively. Supporting hypothesis 3a, perceived diversity management (.58, p < .01) is positively related to organizational performance, while EEO complaints (–.031, p < .01) are negatively related to organizational performance. In other words, when agencies have more eff ective diversity management policies and practices, they also perform better than others that are less eff ective at managing diversity.

Th e results of the moderating eff ects of diversity management are complicated to interpret. Perceived diversity management (.13, p < .01) positively moderated the relationship between racial diver- sity and organizational performance, suggesting that in agencies that eff ectively manage diversity, racial diversity is associated with increased perceived organizational performance. Th is indicates that although higher levels of racial diversity will decrease organizational performance, eff ective management will turn negative eff ects on organizational performance into positive eff ects. However, the mod- erating eff ects of perceived diversity management on the other two types of diversity are found to be insignifi cant, in partial support of hypothesis 2b.

Hypothesis 2b also proposed that the eff ects of EEO complaints moderate the relationships between the attributes of diversity and organizational performance such that less eff ective diversity management will enhance the negative eff ects of diversity on orga- nizational performance. Interestingly, in the agencies with higher levels of EEO complaints, which indicates poor management for diversity, higher levels of gender diversity (.85, p < .01) had a posi- tive association with organizational performance, while racial (–.52, p < .01) and age diversity (–1.34, p < .01) related negatively to performance.

Effects of Other Contextual Variables (H 3 –H

5 )

Hypothesis 3a, which examined the eff ect of employees’ organiza- tional tenure on organizational performance, was not supported. Th e fi ndings of the moderating eff ect of tenure were also not consistent with our hypothesis 3b. We predicted that organizational tenure would positively moderate the association between diver- sity and organizational performance. However, the results indicate that higher levels of racial diversity are negatively associated with perceived organizational performance when employees of the agency have longer organizational tenure. In contrast, higher levels of gen- der diversity tend to increase organizational performance when the employees worked together longer. Th e moderating eff ect of tenure on age diversity was not signifi cant.

Th e eff ects of team processes and results-oriented organizational culture on organizational performance supported hypotheses 4a 5a. Both of these two variables were positively associated with organiza- tional performance (.37, p < .01 and .38, p < .01, respectively).

Hypothesis 4b predicted that team processes would positively moderate the associations of the three attributes of diversity and organizational performance. Team processes showed a pattern of moderating eff ects similar to that of perceived diversity manage- ment. Th e moderating eff ects of team processes (.081, p < .05) were signifi cant only in the relationship between racial diversity and organizational performance, suggesting that in agencies eff ectively

Table 3 Correlations of Variables

1 2 3 4 5 6 7 8 9 10 11 12 13

1.Organizational performance

2. Race diversity –.05

3. Gender diversity .01 .14

4. Age diversity –.01 .09 .27

5. Size .02 –.16 –.35 –.09

6. Tenure .02 –.13 .01 –.45 –.19

7. Gender –.00a .09 .12 .01 –.14 .04

8. Minority –.02 .15 –.00a –.02 –.06 .03 .19

9. Supervisor .08 –.01 –.07 .05 .18 –.14 –.14 –.07

10. Executive .08 –.01 .02 .01 –.01 .02 –.06 –.06 –.12

11. Diversity management .57 –.04 .00a –.01 .03 .03 –.09 –.19 .15 .11

12. EEO complaints –.05 .32 –.03 –.09 –.31 .11 .10 .11 –.04 –.05 –.06

13. Team processes .63 –.04 –.00a –.01 .02 .01 –.06 –.07 .11 .07 .49 –.04

14. Organizational culture .66 –.04 .02 –.00a .04 .04 –.06 –.04 .21 .15 .64 –.07 .52

Note: all of the coeffi cients, except those marked by “a,” are statistically signifi cant at the .01 or .001 levels.

116 Public Administration Review • January | February 2010

cohesion in groups, consequently decreasing organizational perfor- mance (Milliken and Martins 1996; Williams and O’Reilly 1998). In contrast, higher levels of racial diversity decreased perceived organizational performance, while higher levels of gender diver- sity increased perceived organizational performance. One possible interpretation is that gender may lead to fewer confl icts than do other types of diversity, thus boosting the positive eff ects of diversity on organizational performance (Pelled, Eisenhardt, and Xin 1999). Gender diversity may be a less sensitive issue in work groups, as compared to racial diversity, and may require lower costs for confl ict resolution or coordination. On the other hand, work groups with

higher levels of racial diversity may experience more confl icts because of complicated dif- ferences arising from residual eff ects of racial injustice and related factors.

Th e moderating eff ects of diversity manage- ment on the relationship between diversity and organizational performance contribute the most interesting fi ndings of this study. Although policy makers and public managers may understand the importance of workforce diversity very well, they may not realize the crucial need for eff ective management prac- tices to enhance the positive impacts of diver- sity on organizational performance. Th is study developed two measures of diversity manage- ment—a subjective measure and an objective measure. Even though the results of these two measures did not exactly match, the pattern of the impacts is consistent. Th e fi ndings suggest that the eff ects of diversity management can enhance or even reverse the main impacts of diversity on organizational outcomes. When an agency’s employees perceived that leaders managed racial diversity eff ectively, higher levels of racial diversity improved perceived organizational performance. In contrast, higher levels of racial diversity in agencies that

were not successful in managing diversity signifi cantly decreased perceptions of the agency’s performance. Th ese results support the integration and learning perspective on diversity, which proposes that if organizations properly manage employees’ varied perspec- tives, diversity can be a source of growth, learning, and intuition, thus enhancing organizational performance (Foldy 2004; Th omas and Ely 1996, 2001). Th e results are also consistent with Cox and Blake’s (1991) conclusion that managing diversity can bring a number of benefi ts to organizations by improving creativity and problem-solving capacity and reducing turnover.

Along similar lines, when an agency manages diversity poorly, higher levels of age diversity decrease organizational performance. On the other hand, the moderating eff ects of diversity management on the relationship between gender diversity and organizational performance appear to be more complicated to interpret. In agencies that manage gender diversity poorly, higher levels of gender diversity increased organizational performance. As mentioned previously, gender diver- sity may not cause high costs for confl ict resolution and coordination, thereby not requiring managerial capacity or eff orts for diversity.

practicing teamwork and cooperation among members in work- groups, higher levels of racial diversity increased organizational performance.

Hypothesis 5b proposed that organizational culture should negatively moderate the relationship between diversity and orga- nizational performance. However, the moderating eff ects of organizational culture on the relationship between diversity and organizational performance were mixed. Th e results showed that in agencies with high results-oriented culture, higher age diversity increased organizational performance, while the moderating eff ects of organizational culture on racial and gender diversity were insignifi cant.

Results of Control Variables Female employees, supervisors, and Senior Executive Service members perceive higher organizational performance, while racial minority employees perceive lower organiza- tional performance. However, organizational tenure of employees and organizational size were not signifi cantly associated with organi- zational performance.

Discussion and Conclusion Previous empirical research on the eff ects of diversity on organizational performance has found mixed results. Recently, researchers have reported that the actual eff ects of diversity under complex organizational conditions are much more complicated than they had pre- dicted, suggesting the need for a more nuanced interpretation of diversity and for research that analyzes the impacts of diversity taking into account the complex reality of organizations (Bell and Berry 2007; Klein and Harrison 2007). Th is study responds to these needs, ana- lyzing not only the main impacts of diversity but also moderating eff ects of contextual fac- tors on organizational performance. Th e results of this study add to recent research that indicates the importance of contextual variables for understanding impacts of diversity on organizational outcomes (Jackson, Joshi, and Erhardt 2003; Kochan et al. 2003; Milliken and Martins 1996; Williams and O’Reilly 1998). Th e results suggest that contextual variables signifi cantly aff ect the relationship between diversity and organizational performance. Th e most important evidence of such moderating eff ects comes from the fi nding that this very large sample of federal employees perceives lower organizational eff ectiveness where racial diversity is higher, but strikingly, where they perceive that diversity is eff ectively managed, perceived organizational eff ectiveness is higher when racial diversity is greater.

Th e fi ndings suggest that the main eff ects of diversity may vary according to the types of diversity. Th is is consistent with the fi ndings of previous research (Sackett and DuBois 1991; Watson, Kumar, and Michaelsen 1993; Wise and Tschirhart 2000). Th e main eff ect of racial diversity was consistent with social categori- zation and social identity theories that propose greater diversity is associated with less social integration, more confl ict, and less

Recently, researchers have reported that the actual eff ects

of diversity under complex organizational conditions are much more complicated than they had predicted, suggesting the need for a more nuanced interpretation of diversity and for research that analyzes the

impacts of diversity taking into account the complex reality of

organizations.

Th e moderating eff ects of diversity management on the relationship between diversity

and organizational performance contribute the most interesting

fi ndings of this study.

Managing Diversity in U.S. Federal Agencies 117

Th e fi ndings concerning the eff ects of other contextual factors are also interesting. Th e moderating eff ects of organizational tenure were signifi cant, but had only a small eff ect on the relationship between diversity and performance. When employees worked together longer, higher levels of gender diversity increased perfor- mance, but racial diversity decreased performance. In this study, organizational tenure does not strongly aff ect the relationship between diversity and performance, contrary to other studies that have found signifi cant moderating eff ects of organizational tenure on diversity and organizational eff ectiveness (e.g., Chatman and Flynn 2000; Ely 2004; Harrison et al. 2002). Th ese mixed fi nd- ings may be attributed to the oversimplifi cation of the dynamics of diversity. Page (2007) argued that when group members have the same organizational tenure, they are likely to think in the same way. Others, however, contend that depending on the context, diff erences in some attributes such as race/ethnicity and sex can have greater impacts on organizational outcomes than others. Th is suggests that stronger impacts of racial or sex diversity can cause diff erences in tenure and age to fade in signifi cance (Bell and Berry 2007).

Team processes appear to have marginal moderating eff ects on the relationship between diversity and performance. Even so, racial diversity signifi cantly improved performance in the agencies that encourage teamwork and cooperation among employees. Team processes that encourage frequent interactions and communica- tion among members increase positive eff ects of racial diversity on organizational performance.

Th e fi ndings regarding results-oriented organizational culture partially supported our expectation. Only age diversity had a posi- tive association with performance in agencies that emphasized the performance and productivity of employees. As discussed earlier, results-oriented organizational culture tends to promote competi- tion among members in workgroups, potentially increasing confl icts (Lawrence 1988; Pelled, Eisenhardt, and Xin 1999). Higher levels of age diversity in an agency that emphasizes results-oriented culture may increase organizational performance, reducing competition and confl icts among employees.

Further research should address potential limitations of this study. Because this research analyzed cross-sectional data, it does not conclusively establish the causal directions of the relations observed. Some variables were measured by self-reported responses from FHCS, so monomethod biases could aff ect the results. Th e analy- sis, however, does include variables that use multiple sources (e.g., CPDF, archival data, and No FEAR Act), reducing the possibil- ity of such biases. In addition, subjective evaluations of diversity management may not capture the actual eff ectiveness of diversity management in an organization. Kalev, Dobbin, and Kelly (2006) found that diversity management programs targeting individual bias or network isolation were less eff ective than organizational struc- tures that strengthen responsibility for diversity (e.g., affi rmative action programs and diversity committees). Future research should develop more rigorous measures of diversity management and more independent, “objective” measures of performance. Even so, exam- ining how people in organizations feel about their organization’s performance, and about other matters analyzed here, has great value (e.g., Brewer 2006).

Table 4 Hierarchical Regression Analysis Predicting Organizational Performance

Coeffi cient Std. Err.

Step 1: Controls

Size .0013 .0028

Tenure –.0011 .0015

Gender .0335*** .0073

Supervisor .2107*** .0080

Executive .1149*** .0232

Minority –.0173* .0081

R2 = .0171 Adjusted R2 = .0170

F = 224.09***

Step 2: Diversity Effects

Race diversity (RD) –.4163*** .0389

Gender diversity (GD) .3102*** .0945

Age diversity (AD) –.0719 .1147

R2 = .0186 Adjusted R2 = .0185

F = 162.62***

Step 3: Management Effects

Diversity management .5781*** .0030

EEO complaints –.0310*** .0102

R2 = .3319 Adjusted R2 = .3318

F = 3484.66***

Step 4: Teamwork and Organizational Culture Effects

Team processes .3669*** .0029

Results-oriented culture .3783*** .0033

R2 = .5691 Adjusted R2 = .5690

F = 7840.82***

Step 5: The Moderating Effects of Diversity Management

Diversity management * RD .1264*** .0256

Diversity management * GD –.0873 .0612

Diversity management * AD –.0479 .0714

EEO * RD –.5225*** .1180

EEO * GD .8450*** .2924

EEO * AD –1.3439*** .3246

R2 = .5694 Adjusted R2 = .5693

F = 5370.71***

Step 6: The Moderating Effects of Organizational Tenure

Tenure * RD –.0609*** .0148

Tenure * GD .0654* .0290

Tenure * AD .0373 .0285

R2 = .5695 Adjusted R2 = .5694

F = 4640.38***

Step 7: The Moderating Effects of Team Processes

Team processes * RD .0811** .0296

Team processes * GD .0501 .0705

Team processes * AD –.1387 .0825

R2 = .5696 Adjusted R2 = .5694

F = 4084.45***

Step 8: The Moderating Effects of Organizational Culture

Organizational culture * RD –.0022 .0347

Organizational culture * GD –.0223 .0825

Organizational culture * AD .3608*** .0974

R2 = .5697 Adjusted R2 = .5695 F = 3647.97***

* Signifi cant at .10 level; ** signifi cant at .05 level; *** signifi cant at .01 level.

118 Public Administration Review • January | February 2010

Our results have important implications for public managers and policy makers. Especially for racial diversity, the evidence shows the importance of managing diversity eff ectively. Th e benefi ts of racial diversity are enhanced when leaders work well with employees of diverse backgrounds, show a commitment to a workforce that is rep- resentative of all society, and establish policies and procedures that pro- mote diversity. Th e evidence here supports decisions to invest resources in developing such leadership behaviors, policies, and procedures.

Appendix: Construction of Index Variables (Five-point Likert-type scales)

Organizational Performance

(Cronbach’s alpha = .74, eigenvalue = 2.280, factor loadings = .72–.80)

How would you rate the overall quality of work done by your work group?

Overall, how good a job do you feel is being done by your im- mediate supervisor/team leader?

Th e workforce has the job-relevant knowledge and skills neces- sary to accomplish organizational goals.

Th e skill level in my work unit has improved in the past year.

Diversity Management

(Cronbach’s alpha = .85, eigenvalue = 2.308, factor loadings = .87–.89)

Supervisors/team-leaders in my work unit are committed to a workforce representative of all segments of society.

Policies and programs promote diversity in the workplace (for example, recruiting minorities and women, training in awareness of diversity issues, mentoring).

Managers/supervisors/team leaders work well with employees of diff erent backgrounds.

Results-Oriented Organizational Culture

(Selected items—contact authors for the others)

(Cronbach’s alpha = .94, eigenvalue = 5.74, factor loadings = .69–.89)

In my work unit, steps are taken to deal with a poor performer who cannot or will not improve.

Employees are rewarded for providing high quality products and services to customers.

Awards in my work unit depend on how well employees per- form their jobs.

In my work unit, diff erences in performance are recognized in a meaningful way.

Team Processes

(Cronbach’s alpha = .63, eigenvalue = 1.46, factor loading = .86)

Th e people I work with cooperate to get the job done. Employees in my work unit share job knowledge with each

other.

•

•

•

•

•

•

•

•

•

•

•

• •

Notes 1. Th e Central Personnel Data File, an information system intended to support

statistical analysis of federal personnel management programs, is composed of two primary data fi les—the status fi le and the dynamics fi le. Th is study extracted 2004 employees’ demographic data of the target agencies of the federal govern- ment from the dynamic fi le, which has information about personnel actions such as accessions, separations, and promotions over a period of time. Th e CPDF is limited to federal civilian employees and covers all agencies under the executive branch except for the Board of Governors of the Federal Reserve, Central Intel- ligence Agency, Defense Intelligence Agency, National Imagery and Mapping Agency, National Security Agency, Offi ce of the Vice President, Postal Rate Commission, Tennessee Valley Authority, U.S. Postal Service, and White House Offi ce. Th e coverage of Federal Bureau of Investigation, the legislative branch, and the judicial branch is limited. All data of the CPDF are submitted by the agencies from their own separate personnel systems on a quarterly basis (OPM 2006).

2. Previous research on diversity employed various attributes of diversity, such as age, education, ethnicity, gender, institutional location, job category, national origin, supervisory status, and race. Because of limited data availability, this research focuses on three dimensions of diversity: race, gender, and age diversity. Archival data on the race, gender, and age of federal employees by agency come from the CPDF. Th e data represent one time period and were collected in 2004.

3. Response rates involving the questions about institutional location (27.2 per- cent), ethnic background (27.1 percent), pay category/grade (27.2 percent), ten- ure in federal government (24.2 percent) and in current agency (23.9 percent), and age (27.2 percent) are very low and limited to only the defense agencies.

4. On May 15, 2002, President George W. Bush signed the Notifi cation and Fed- eral Employee Antidiscrimination and Retaliation of 2002 Act into law. Th e No FEAR Act, with an eff ective date of October 1, 2003, requires federal agencies to be more accountable for violations of antidiscrimination and whistle-blower protection laws. Th e act requires each federal agency to post quarterly on its public Web site certain statistical data relating to federal sector equal employ- ment opportunity complaints fi led with such agency and for other purposes. Under the act, every federal agency is required to notify employees and appli- cants for employment about their rights under the discrimination and whistle- blower laws; to post statistical data relating to federal sector equal employment opportunity complaints on its public Web site; to ensure that their managers have adequate training in the management of a diverse workforce, early and alternative confl ict resolution, and essential communications skills; to conduct studies on the trends and causes of complaints of discrimination; implement new measures to improve the complaint process and the work environment; to initiate timely and appropriate discipline against employees who engage in misconduct related to discrimination or reprisal; to reimburse the Judgment Fund for any discrimination and whistleblower related settlements or judgments reach in federal court; and to produce annual reports of status and progress to Congress, the attorney general, and the U.S. Equal Employment Opportunity Commission.

5. Specifi cally, the signifi cance of the interaction term regression weights may be attributable only to the overlap between the interaction terms and untested but signifi cant nonlinear trends, and not to an actual interaction between the variables.

References Aiken, Leona S., and Stephen G. West. 1991. Multiple Regression: Testing and Inter-

preting Interactions. Newbury Park, CA: Sage Publications. Allison, Paul D. 1977. Testing for Integration in Multiple Regression. American

Journal of Sociology 83: 144–53. Argyris, Chris, and Donald A. Schön. 1996. Organizational Learning II: Th eory,

Method, and Practice. Reading, MA: Addison-Wesley.

Managing Diversity in U.S. Federal Agencies 119

Ashforth, Blake E., and Fred Mael. 1989. Social Identity Th eory and the Organization. Academy of Management Review 14(1): 20–39. Bell, Myrtle P., and Daphne P. Berry. 2007. Viewing Diversity through Diff erent Lenses: Avoiding a Few Blind Spots. Academy of Management Perspectives 21(4): 21–25. Brewer, Gene A. 2002. Public Workforce Integration: A Field Essay. Paper presented at the Annual Meeting of the Southern Political Science Association, November 6–9,

Savannah, GA. ———. 2006. All Measures of Performance Are Subjective: More Evidence on U.S. Federal Agencies. In Public Service Performance, edited by George A. Boyne, Kenneth J.

Meier, Laurence J. O’Toole, Jr., and Richard A. Walker, 35–54. Cambridge: Cambridge University Press. Bunderson, J. Stuart, and Kathleen M. Sutcliff e. 2002. Comparing Alternative Conceptualizations of Functional Diversity in Management Teams: Process and Performance

Eff ects. Academy of Management Journal 45(5): 875–93. Byrne, Donn. 1971. Th e Attraction Paradigm. New York: Academic Press. Chatman, Jennifer A., and Francis J. Flynn. 2001. Th e Infl uence of Demographic Composition on the Emergence and Consequences of Cooperative Norms in Groups.

Academy of Management Journal 44(5): 956–74. Chatman, Jennifer A., and Karen A. Jehn. 1994. Assessing the Relationship between Industry Characteristics and Organizational Culture: How Diff erent Can You Be? Acad-

emy of Management Journal 37(3): 522–53. Chatman, Jennifer A., Jeff rey T. Polzer, Sigal G. Barsade, and Margaret A. Neale. 1998. Being Diff erent yet Feeling Similar: Th e Infl uence of Demographic Composition and

Organizational Culture on Work Processes and Outcomes. Administrative Science Quarterly 43(4): 749–80. Cohen, Jacob. 1978. Partialed Products Are Interactions; Partialed Powers Are Curve Components. Psychological Bulletin 85(4): 858–66. Cohen, Jacob, and Patricia Cohen. 1983. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum. Cornwell, Christopher, and J. Edward Kellough. 1994. Women and Minorities in Federal Agencies: Examining New Evidence from Panel Data. Public Administration Review

54(3): 265–70. Cox, Taylor, Jr. 1993. Cultural Diversity in Organizations: Th eory, Research, and Practice. San Francisco: Berrett-Koehler. Cox, Taylor, Jr., and Stacy Blake. 1991. Managing Cultural Diversity: Implications for Organizational Competitiveness. Academy of Management Executive 5(3): 45–56. Cox, T. J., Sharon A. Lobel, and Poppy L. McLeod. 1991. Eff ects of Ethnic and Group Cultural Diff erences on Cooperative and Competitive Behavior on a Group Task.

Academy of Management Journal 34(4): 827–47. De Dreu, Carsten K., and Laurie R. Weingart. 2003. Task versus Relationship Confl ict, Team Performance, Team Member Satisfaction: A Meta-Analysis. Journal of Applied

Psychology 88(4): 741–49. Dutton, Jane E., and Robert B. Duncan. 1987. Th e Criteria of Momentum for Change through the Process of Strategic Issue Diagnosis. Strategic Management Journal 8(3):

279–96. Ely, Robin J. 2004. A Field Study of Group Diversity, Participation in Diversity Education Programs, and Performance. Journal of Organizational Behavior 25(6): 755–80. Foldy, Erica Gabrielle. 2004. Learning from Diversity: A Th eoretical Exploration. Public Administration Review 64(5): 529–38. Friedrich, Robert J. 1982. In Defense of Multiple Terms in Multiple Regression Equations. American Journal of Political Science 26(4): 797–833. Garson, David G. 1998. http://faculty.chass.ncsu.edu/garson/PA765/regress.htm. Gladstein, Deborah L. 1984. A Model of Task Group Eff ectiveness. Administrative Science Quarterly 29(4): 499–517. Greenhaus, Jeff rey H., Saroj Parasuraman, and Wayne M. Wormley. 1990. Eff ects of Race on Organizational Experiences, Job Performance Evaluations, and Career Outcomes.

Academy of Management Journal 33(1): 64–86. Hambrick, Donald E., and Phyllis A. Mason. 1982. Upper Echelons: Th e Organization as a Refl ection of Its Top Managers. Academy of Management Review 9(2): 193–206. Harrison, David A., Kenneth H. Price, Joanne H. Gavin, and Anna T. Florey. 2002. Time, Teams, and Task Performance: Changing Eff ects of Surface- and Deep-Level Diver-

sity on Group Functioning. Academy of Management Journal 45(5): 1029–45. Ibarra, Herminia. 1993. Personal Networks of Women and Minorities in Management. Academy of Management Review 18(1): 56–87. Jackson, Susan E., Aparna Joshi, and Niclas L. Erhardt. 2003. Recent Research on Team and Organizational Diversity: SWOT Analysis and Implications. Journal of Manage-

ment 29(6): 801–30. Jehn, Karen A. 1995. A Multimethod Examination of the Benefi ts and Detriments of Intragroup Confl ict. Administrative Science Quarterly 40(2): 256–82. Jehn, Karen A., and Katerina Bezrukova. 2004. A Field Study of Group Diversity, Workgroup Context, and Performance. Journal of Organizational Behavior 25(6): 703–29. Jehn, Karen A., and Elizabeth A. Mannix. 2001. Th e Dynamic Nature of Confl ict: A Longitudinal Study of Intragroup Confl ict and Group Performance. Academy of Manage-

ment Journal 44(2): 238–51. Jehn, Karen A., Gregory B. Northcraft, and Margaret A. Neale. 1999. Why Diff erence Make a Diff erence: A Field Study of Diversity, Confl ict, and Performance in Work-

groups. Administrative Science Quarterly 44(4): 741–63. Kalev, Alexandra, Frank Dobbin, and Erin Kelly. 2006. Best Practices or Best Guesses? Assessing the Effi cacy of Corporate Affi rmative Action and Diversity Policies. Academy

of Management Perspectives 71(4): 589–617. Kanter, Rosabeth M. 1979. Power Failure in Management Circuit. Harvard Business Review 57(4): 65–75. Katz, Ralph. 1982. Th e Eff ects of Group Longevity on Project Communication and Performance. Administrative Science Quarterly 27(1): 81–104. Kellough, J. Edward. 1990. Integration in the Public Workplace: Determinants of Minority and Female Employment in Federal Agencies. Public Administration Review 50(5):

557–66. Klein, Katherine J., and David A. Harrison. 2007. On the Diversity of Diversity: Tidy Logic, Messier Realities. Academy of Management Perspectives 21(4): 26–33. Kochan, Th omas, Katerina Bezrukova, Robin Ely, Susan Jackson, Aparna Joshi, Karen Jehn, Jonathan Leonard, et al. 2003. Th e Eff ects of Diversity on Business Performance:

Report of the Diversity Research Network. Human Resource Management 42(1): 3–21. Kozlowski, Steve W. J., and Bradford S. Bell. 2003. Work Groups and Teams in Organizations. In Handbook of Psychology, vol. 12, Industrial and Organizational Psychology,

edited by Walter C. Borman, Daniel R. Illgen, and Richard J. Klimoski, 333–75. New York: Wiley. Kramer, Roderick M. 1991. Intergroup Relations and Organization Dilemmas: Th e Role of Categorization Processes. Research in Organizational Behavior 13: 191–228. Krislov, Samuel, and David H. Rosenbloom. 1981. Representative Bureaucracy and the American Political System. New York: Praeger.

120 Public Administration Review • January | February 2010

Lawrence, Barbara S. 1988. New Wrinkles in a Th eory of Age: Demography, Norms, and Performance Ratings. Academy of Management Journal 31(2): 309–37. Marks, Michelle A., John E. Mathieu, and Stephen J. Zaccaro. 2001. A Temporally Based Framework and Taxonomy of Team Processes. Academy of Management Review

26(3): 356–76. McLeod, Poppy L., and Stephen Lobel. 1992. Th e Eff ects of Ethnic Diversity on Idea Generation in Small Groups. Paper presented at the Annual Meeting of the Academy of

Management, Las Vegas, NV. Miller, Vincent P., and John M. Quigley. 1990. Segregation by Racial and Demographic Group: Evidence from the San Francisco Bay Area. Urban Studies 27(1): 3–21. Milliken, Frances J., and Luis L. Martins. 1996. Searching for Common Th reads: Understanding the Multiple Eff ects of Diversity in Organizational Groups. Academy of

Management Review 21(2): 402–34. Mohammed, Susan, and Linda C. Angell. 2004. Surface- and Deep-Level Diversity in Workgroups: Examining the Moderating Eff ects of Team Orientation and Team Process

on Relationship Confl ict. Journal of Organizational Behavior 25(8): 1015–39. O’Reilly, Charles A., III, David F. Caldwell, and William P. Barnett. 1989. Work Group Demography, Social Integration, and Turnover. Administrative Science Quarterly 34(1):

21–37. O’Reilly, Charles A., III, Katherine Y. Williams, and Sigal G. Barsade. 1997. Group Demography and Innovation: Does Diversity Help? In Research in the Management of

Groups and Teams, vol. 1, edited by Elizabeth A. Mannix and Margaret A. Neale, 183–207. Greenwich, CT: JAI Press. Ospina, Sonia. 2001. Managing Diversity in the Civil Service: A Conceptual Framework for Public Organizations. In Managing Diversity in the Civil Service, 11–29. Amster-

dam: IOS Press. Page, Scott. 2007. Making the Diff erence: Applying a Logic of Diversity. Academy of Management Perspectives 21(4): 6–20. Pelled, Lisa Hope, Kathleen M. Eisenhardt, and Katherine R. Xin. 1999. Exploring the Black Box: An Analysis of Work Group Diversity, Confl ict, and Performance. Adminis-

trative Science Quarterly 44(1): 1–28. Pettigrew, Th omas F., and Joanne Martin. 1989. Organizational Inclusion of Minority Groups: A Social Psychological Analysis. In Ethnic Minorities: Social Psychological Per-

spectives, edited by Jan Pieter Van Oudenhoven and Tineke M. Willemsen, 169–200. Berwyn, PA: Swets North America. Pfeff er, Jeff rey. 1983. Organizational Demography. In Research in Organizational Behavior, vol. 5, edited by L. L. Cummings and Barry M. Staw, 299–357. Greenwich, CT:

JAI Press. Pielou, E. C. 1977. Mathematical Ecology. New York: Wiley. Pitts, David W. 2005. Diversity, Representation, and Performance: Evidence about Race and Ethnicity in Public Organizations. Journal of Public Administration Research and

Th eory 15(4): 615–31. Reardon, S. F. 1998. Methods of Measuring Diversity and Segregation in Multi-Group Populations: With Examples Using Racial School Enrollment Data. Unpublished draft,

supported by the Spencer Foundation Small Grants Program. Reichers, Arnon E., and Benjamin Schneider. 1990. Climate and Culture: an Evolution of Constructs. In Organizational Climate and Culture, edited by Benjamin Schneider,

5–39. San Francisco: Jossey-Bass. Riccucci, Norma M. 2002. Managing Diversity in Public Sector Workforces. Boulder, CO: Westview Press. Richard, Orlando. 2000. Racial Diversity, Business Strategy, and Firm Performance: A Resource-Based View. Academy of Management Journal 43(2): 164–77. Richard, Orlando, and Nancy Brown Johnson. 2001. Understanding the Impact of Human Resource Diversity Practices on Firm Performance. Journal of Management 13(2):

177–95. Sackett, Paul R., and Cathy L. DuBois. 1991. Rater–Ratee Race Eff ects on Performance Evaluation: Challenging Meta-Analytic Conclusion. Journal of Applied Psychology

76(6): 873–77. Schneider, Benjamin. 1987. Th e People Make the Place. Personnel Psychology 40(3): 437–53. Schneider, Sherry K., and Gregory B. Northcraft. 1999. Th ree Dilemmas of Workplace Diversity in Organizations: A Social Identity Perspective. Human Relations 52(11):

1445–67. Simons, Tony L., and Randall S. Peterson. 2000. Task Confl ict and Relationship Confl ict in Top Management Teams: Th e Pivotal Role of Intragroup Trust. Journal of Applied

Psychology 85(1): 102–11. Smith, Kent W., and M. S. Sasaki. 1979. Decreasing Multicollinearity: A Method for Models with Multiplicative Functions. Sociological Methods and Research 8(1): 35–56. Tajfel, Henri, and J. C. Turner. 1985. Th e Social Identity Th eory of Intergroup Behavior. In Th e Psychology of Intergroup Relations, edited by Stephen Worchel and William G.

Austin, 7–24. Chicago: Nelson-Hall. Th omas, David A., and Robin J. Ely. 1996. Making Diff erences Matter: A New Paradigm for Managing Diversity. Harvard Business Review 74(5): 79–90. ———. 2001. Cultural Diversity at Work: Th e Eff ects of Diversity Perspectives on Work Group Processes and Outcomes. Administrative Science Quarterly 46(2): 229–73. Triandis, Harry C., and Eunkook M. Suh. 2002. Cultural Infl uences on Personality. Annual Review of Psychology 53: 133–60. Tsui, Anne S., Terri D. Egan, and Charles A. O’Reilly III. 1992. Being Diff erent: Relational Demography and Organizational Attachment. Administrative Science Quarterly

37(4): 549–79. Tsui, Anne S., Terri D. Egan, and Katherine R. Xin. 1995. Diversity in Organizations: Lessons from Demography Research. In Diversity in Organizations: New Perspectives

from a Changing Workplace, edited by Martin M. Chemers, Stuart Okampo, and Mark A. Constanzo, 191–219. Th ousand Oaks, CA: Sage Publications. Turner, John C. 1987. Rediscovering the Social Group: A Self-Categorization Th eory. Oxford, UK: Blackwell. U.S. Equal Employment Opportunity Commission (EEOC). 2005. 29 C.F.R. Part 1614. http://edocket.access.gpo.gov/2006/E6-12432.htm [accessed September 27, 2009]. U.S. Offi ce of Personnel Management (OPM). 2006. FedScope: Federal Human Resources Data. http://www.fedscope.opm.gov/datadefn/acpdf.asp [accessed September 27,

2009]. Watson, Warren E., Kamales Kumar, and Larry K. Michaelsen. 1993. Cultural Diversity’s Impact on Interaction Process and Performance: Comparing Homogeneous and

Diverse Task Groups. Academy of Management Journal 36(3): 590–602. Wiersema, Margarethe F., and Karen A. Bantel. 1992. Top Management Team Demography and Corporate Strategic Change. Academy of Management Journal 35(1): 91–121.

Managing Diversity in U.S. Federal Agencies 121

Williams, Katherine Y., and Charles A. O’Reilly III. 1998. Demography and Diversity in Organizations: A Review of 40 Years of Research. Research in Organizational Behavior 20: 77–140.

Wise, Lois Recascino, and Mary Tschirhart. 2000. Examining Empirical Evidence on Diversity Eff ects: How Useful Is Diversity Research for Public-Sector Managers? Public Administration Review 60(5): 386–94.

Wright, Gerald C. 1976. Linear Models for Evaluating Conditional Relationships. American Journal of Political Science 20(2): 349–73.

FOUNDATIONS OF PUBLIC ADMINISTRATION

Let PAR be your Guide to discovering the Treasures in the ongoing and historical work of Public Administration….

Th e Foundations of Public Administration Series is a collection of articles written by experts in 20 content areas, providing introductory essays and recommending top articles in those subjects. Introducing:

Human Resource Management as a Core Dimension of Public Administration

By: Patricia W. Ingraham and Nadia Rubaii-Barrett Binghamton University

Th e Foundations of Public Administration Series also includes a sample Syllabus and Select Readings from PAR on Human Resource

Management as a Core Dimension.

Login at http://www.aspanet.org and click on “Public Administration Review”, and ‘Foundations of Public Administration’. Select ‘Human Resource Management’

Copyright of Public Administration Review is the property of Blackwell Publishing Limited and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.