Management Programs? Evidence from Public Organizations David W. Pitts American University Alisa K. Hicklin University of Oklahoma Daniel P. Hawes Kent State University Erin Melton Texas A&M University ABSTRACT As the diversity of the US workforce continues to increase at a rapid pace, public managers are facing pressure to create organizational cultures that permit employees from different backgrounds to succeed. A typical managerial response to this diversity has been the implementation of a formal diversity management program. Although limited empirical research has considered links between diversity management activity and organizational performance, very little research has examined the factors that inuence the implementa- tion of diversity management practices. This article begins with the premise that organizations develop diversity management programs as a means of responding to opportunities and challenges in the internal and external environments. In order to delineate how those environmental phenomena operate, we draw from a specic set of organization theories to formulate three drivers of diversity management implementation: environmental uncertainty, environmental favorability, and institutional isomorphism. We test these drivers empirically using data from public schools, nding that elements of all three forces inuence diversity management implementation but in different ways and in varying degrees. As the diversity of the US workforce continues to increase at a rapid pace, public managers are facing pressure to create organizational cultures that permit employees from different backgrounds to succeed. The one-size-ts-all approach to management that was effective 30 years ago is arguably no longer an appropriate strategy for ensuring maximumemployee performance (Riccucci, 2002). As more women and people of color enter the workforce, An earlier version of this article was presented at the 9th Public Management Research Conference, Tucson, Arizona, October 2527, 2007. The authors thank Ed Kellough for helpful comments and suggestions. Address correspondence to the author at pitts@american.edu. doi:10.1093/jopart/mup044 Advance Access publication on January 11, 2010 The Author 2010. Published by Oxford University Press on behalf of the Journal of Public Administration Research and Theory, Inc. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org managers must adapt to noteworthy changes, including more emphasis on family and work- life balance (Bertelli, 2007; Ezra and Deckman, 1996; Saltzstein et al., 2001), different cultural assumptions about the role of work (Ho, 1987), more emphasis on collective values and teamwork (Azevedo et al., 2002), and different approaches to communication (Arai et al., 2001). The impetus for responding to these changes varies. Managers may feel a nor- mative obligation to accommodate people of different backgrounds (Mosher, 1982; Naff, 2001); they may be under legal obligation to be inclusive (Hicklin, 2007; Kellough and Rosenbloom, 1992; Selden, 1997) or they may try to use diversity as a strategic means of augmenting performance (Andrews et al., 2005; Pitts, 2005; Pitts and Jarry, 2007; Wise and Tschirhart, 2000). Whatever the motivation, the managerial response to workforce diversity has often been the initiation of a formal diversity management program (Kellough and Naff, 2004). These programs involve both human resources management and line management functions as responsibility for managing human capital is increasingly split between the two (for an overview, see Selden, 2008). Research linking diversity management programs to work-related outcomes has been scant, and the fewstudies conducted have yielded mixed results (Cox, 1993; Naff and Kellough, 2003; Pitts, 2009; Sanchez and Brock, 1996). Al- though there remains much work to be done linking diversity management to key outcomes for relevant groups, we focus on a different relationship in this article: the factors behind diversity management implementation. We argue that organizations develop diversity management programs as a means of responding to environmental opportunities and chal- lenges. In order to delineate how those environmental phenomena operate, we draw from a specic set of organization theories to develop three ideas about what drives diversity management implementation. We then use data from a sample of public school districts to test the impact of these three factors empirically to understand why some organizations implement diversity management while others do not. THE DIVERSITY MANAGEMENT CONSTRUCT Diversity management is an organizational response to workforce diversity and its chal- lenges and opportunities. Diversity management initiatives typically address the concerns of historically underrepresented groups, such as people of color and women, but many also focus on other dimensions of diversity (e.g., age, professional background, religion) that inuence work-related outcomes (Kellough and Naff, 2004; Pitts, 2006; Riccucci, 2002; Wise and Tschirhart, 2000). Empirical research shows that the impact of workforce diver- sity on outcomes can be either positive or negative (for a review, see Jayne and Dipboye, 2004). In some cases, diversity can create opportunities for greater innovation and more creative solutions to problems (Richard, 2000; Richard et al., 2003; Watson et al., 1993). In other cases, diversity can create process-oriented problems, such as poor communication, faulty decision making, and mistrust of other employees (Cunningham, 2005; Swann et al., 2004; Timmerman, 2000; Watson et al., 1993). The previous work on the effects of in- creasing diversity suggests that management may be the key to assuring that the organi- zation will be able to fully benet from bringing underrepresented groups into the organization. Some organizations have adopted diversity management initiatives as a way to improve the ability of diverse groups to work together, and limited empirical research has demonstrated that diversity management can improve outcomes in diverse organizations (Kalev et al., 2006; Ng and Burke, 2005; Pitts, 2009). 868 Journal of Public Administration Research and Theory Public organizations have begun to rapidly adopt formal diversity management pro- grams (Kellough and Naff, 2004). However, diversity management is an ambiguous con- cept that does not seem to be implemented consistently across organizations. For example, some organizations may simply repackage their longstanding Afrmative Action/Equal Employment Opportunity ofces, renaming them Diversity Management as a means of staying up-to-date on nomenclature (Kellough and Naff, 2004). Although many argue that Equal Employment issues remain vital to an organization that actively manages diversity, the movement in the eld has been away from these legalistic responses to diversity and more toward strategic ones (Riccucci, 2002; S. C. Selden and F. Selden, 2001; Thomas, 1990; Wise and Tschirhart, 2000). In their survey of federal agency, human resource man- agers Kellough and Naff (2004) found that almost 90% of agencies reported having a di- versity management program, but the components of these programs varied dramatically. For empirical researchers, this creates a serious obstacle for creating useful measures of the diversity management construct. Diversity management is an organizational process with multiple moving parts that have not been examined very thoroughly by the public management literature (Kellough and Naff, 2004; Pitts, 2006). In this article, we rely on Pittss (2006) comprehensive denition of diversity man- agement as a guiding lens. There is the potential for disagreement about the appropriate denition of diversity management, but this is a broad conceptualization that seems to cover most issues relevant to workforce diversity. His model of diversity management includes three interrelated components: recruitment and outreach, valuing differences, and prag- matic policies and programs. These components represent the three primary activities in which an organization engages in order to manage its employee diversity. The rst of these, recruitment and outreach, considers whether an organization ex- tends itself to all potential sources of employees. A strategic plan for recruiting from un- derrepresented groups is a key component of this aspect of diversity management, and recent research has underscored the importance of recruitment for diversity (Rubaii-Barrett and Wise, 2007; Selden, 2006). This involves seeking out employees from the labor market who may not be found through the typical venues. Increasing organizational diversity has the potential to increase performance, making recruitment a potentially vital step in im- proving organizational outcomes (Pitts, 2005; Pitts and Jarry, 2007; Wise and Tschirhart, 2000). Organizations with greater staff diversity are more likely to generate innovative solutions and creative approaches to services (Adler, 2002; Richard, 2000; Richard et al., 2003). The second of these processes, valuing differences, is more normative in nature and considers whether employees and managers appreciate the different cultural assumptions and biases that employees bring to their work. Previous work argues that multicultural un- derstanding is imperative for managers who oversee the work of diverse employees, and such understanding is arguably improved through programs aimed at bridging cultural gaps (Adler, 2002; Ely and Thomas, 2001; Thomas and Ely, 1996). These programs might take a number of forms, including diversity training and cultural awareness events. Foldy (2004) argues that organizational processes aimed at encouraging multicultural learning will lead to greater benets from existing diversity. Unfortunately, much of the evidence on this point is anecdotal: Empirical research has not produced much support for the effectiveness of values-based programs (Cox, 1993; Sanchez and Brock, 1996), and these types of di- versity training sessions can sometimes promote majority backlash (Riccucci, 1997). Pitts et al. What Drives the Implementation of Diversity Management Programs? 869 Although more research is needed before the eld can be condent about the relationship between values-based diversity training and performance, it remains a key aspect of the diversity management construct and should be considered in any related empirical research. Finally, pragmatic programs and policies consist of a strategic set of management tools that an organization can use to promote employee job satisfaction and performance (Pitts, 2006). These programs and policies provide an explicit means for underrepresented or minority employees to work in a culture that is supportive of their needs. Pragmatic programs and policies typically aim to effectively integrate employees from diverse back- grounds into the organization, such that diversity is a consideration both pre- and posthire. Such programs might involve mentoring opportunities for those outside the majority or- ganizational culture; family friendly programs for those who have children or parents at home that require care; exible working hours for those who have competing demands and are unable to work a consistent, 9-to-5 day; and collaborative assignments for those who work better in groups than individually (Kellough and Naff, 2004; Salzstein et al., 2001). The list varies by organization, but it reects a strategy on the part of managers to assess employee needs and create workplace policies that respond to them. DRIVERS OF DIVERSITY MANAGEMENT IMPLEMENTATION The motivations for engaging in diversity management vary by organization, but we expect that forces in an organizations environment will be key. Research on organizations has focused on the central importance of environmental factors since moving toward open sys- tems theories in the 1950s and 1960s (for a review, see Hatch and Cunliffe, 2006). Organ- izations adapt to their surroundings, and as social, political, and economic environments shift, managers make decisions about how to alter strategies and processes in response. In doing so, they adapt to the values of external society (Selznick, 1957). Empirical research on public organizations has demonstrated that extensively both the external and internal environments are crucial variables in public management models. In their logic of gover- nance, Lynn et al. (2001: 17) note that any governance regime, from the local to the in- ternational, is embedded in a wider social, scal, political, and cultural context. Research on network management has shown a very consistent relationship between external man- agement activity and organizational performance (Goerdel, 2006; Meier and OToole, 2001, 2003, 2005). Reforms during the 1990s and early 2000s have arguably increased the central role of the environment for public organizations as they are marketized and tasked to compete for the right to offer services (Donahue and Nye, 2002). Although it is evident that environmental pressures affect management choices, the specic causal mechanisms are not always clear. In the case of diversity management, there are a number of environmental forces that could inuence implementation. The purpose of this article is to isolate three different environmental forces and examine them as drivers of diversity management implementation. We argue that organizations implement diversity management programs (a) as a response to environmental uncertainty, (b) as a result of environmental favorability and resource municence, and/or (c) in order to adapt to en- vironmental norms and mimic the actions of peer organizations (table 1). Put more simply, they manage diversity (a) because they have to, (b) because they can, or (c) because ev- erybody else is. Some organizations may be motivated by more than one of these factors, and the importance of any given factor is likely to change over time. Although we begin with a conceptual review of these forces as distinct motivations, our empirical analysis 870 Journal of Public Administration Research and Theory permits us to test whether organizations seem to be inuenced by multiple factors at the same time. In the following sections, we outline these three arguments in the context of public schools. Environmental Uncertainty Environmental uncertainty often encourages organizations to develop strategies for min- imizing ambiguity. Organizations can either buffer the technical core from unknown threats, or else they can exploit the changing and uncertain environment as a strategy for success (OToole and Meier, 1999). In the context of our research, organizations may implement diversity management as a means of fending off environmental uncer- tainty. Similar to the actions of a risk-averse individual, organizations try to minimize the uncertainty in their environment by making decisions that promote stability. Classic research in contingency theory showed that organizations should structure themselves in ways that reected the level of uncertainty in the environment (Aldrich and Mindlin, 1978; Galbraith, 1973; Lawrence and Lorsch, 1967). For example, horizontal organizations were best for uncertain environments since they permitted fast response to environmental shifts, whereas vertical organizations were best for certain environments since they pro- moted efciency (for a review, see Daft, 2008). Contingency theory has expanded beyond structure to include a number of other management decisions, key among them human resources policies like diversity management. Perhaps more importantly, research has re- ned contingency theory to suggest that it is perceptions of the environment that guide behavior, not necessarily a set of objective environmental conditions that affect all organ- izations equally (Aldrich and Mindlin, 1978; Galbraith, 1973). Managers make decisions based on the enacted environment, which means that managers perceptions are more Table 1 Drivers of Diversity Management Implementation Driver Example Theoretical Basis Hypothesis Environmental uncertainty School districts implement diversity management as a means of buffering against environmental uncertainty Contingency theory (Aldrich and Mindlin, 1978; Lawrence and Lorsch, 1967) Organizations in environments characterized by uncertainty are more likely than others to adopt diversity management programs. Environmental favorability School districts implement diversity management because they are already performing well in other areas, have the funds with which to do so, and/or have signicant political support in place Resource municence (Dess and Beard, 1983), capacity (Aldrich, 1979), contact hypothesis (Allport, 1954) Organizations in resource-rich environments are more likely than others to adopt diversity management programs. Institutional isomorphism School districts implement diversity management because of professional norms or encouragement from other actors in the external environment Institutional isomorphism (Ashworth et al., 2009; DiMaggio and Powell, 1983; Scott, 2003) Organizations with peers that have implemented diversity management are more likely than others to adopt it. Pitts et al. What Drives the Implementation of Diversity Management Programs? 871 valuable to understanding this process than objective data about environmental stability and complexity (Duncan, 1972; Weick, 1979). Resource dependence theory would further predict that managers will make decisions based on the reactions that they anticipate from important nodes in their network of actors, creating power/dependence relationships that work to their advantage (Pfeffer and Salancik, 1978). In the case of public schools, environmental uncertainty might encourage the adoption of diversity management as educators struggle to stabilize themselves from environmental instability and complexity. That uncertainty could arise from a number of sources: chang- ing student demographics, labor market instability, and changing funding mechanisms are three key examples. Such uncertainty could reect threats, but it could also reect oppor- tunities for growth or more benign changes. Ultimately, it does not matter how the uncer- tainty presents itself but rather the extent to which the manager believes it to exist. The resulting hypothesized relationship would be that managers who perceived the environment to be uncertain would be more likely to implement diversity management than those who did not. Environmental Favorability and Resource Municence A second explanation behind diversity management implementation is quite different: The environment does not require it in the formof uncertainty but rather promotes it through the municence of resources, time, and intergroup harmony. Organization theorists have long established that capacity (Aldrich, 1979) and resource municence (Dess and Beard, 1984) play a large role in determining whether an organization will choose to implement a new program or policy. Research on organizational suggests that slack resources provide a cli- mate where programs like diversity management are more likely to prosper (Boyne, 2003; Fernandez and Rainey, 2006). If resources are plentiful, the rules for adoption become more relaxed, meaning that potentially contentious programs like diversity management are more likely to be accepted. Limited research specic to diversity management suggests that nancial resources are key to implementation success (Pitts, 2007). These resources may be drawn from both the internal and external environments, the distinction between which is often vague in public organizations (Rainey, 2003). This is particularly the case in public school districts, where funding is drawn from a variety of sources at the federal, state, and local levels. In addition to nancial resources, time can play a role in creating new programs. For public school districts, this means whether or not other goals must take priority over di- versity. School districts that are already performing well on standardized tests, graduating students, and sending graduates to college have the luxury of creating a diversity manage- ment program. On the other hand, school districts with high dropout rates and more fun- damental education problems may have to handle those crises rst. If one assumes that the amount of time a school district can spend on goals is xed, then it carries that districts that are already performing well in other areas would be able to devote more attention to diversity management. Finally, it is likely that the environment will encourage the development of a diversity management program if relations between disparate groups are already harmonious. Re- search shows that this type of harmony in diverse communities is not necessarily the norm. For example, members of minority communities may hunker down and avoid intermingling with other minority communities in order to improve their likelihood of survival and 872 Journal of Public Administration Research and Theory success (Putnam, 2007). That should change over time as the contact hypothesis suggests that intergroup harmony improves as the number of out groups increases and contact be- tween different ethnic groups becomes more frequent (Allport, 1954; Welch and Sigelman, 2000). It is possible that a harmonious environment, marked by signicant out groups and general inter-group tolerance, will be more likely to implement diversity management because of a general political will and social desirability. Institutional Isomorphism Our third explanation for diversity management implementation is based in institutional isomorphism, where organizations in the same eld gradually adapt to the same norms (Aldrich, 1979; DiMaggio and Powell, 1983; March and Olsen, 1989). This isomorphism can take place for a number of reasons, but chief among them for diversity management would be mimetic and normative forces (Scott, 2003). Mimetic forces might encourage an organization to change because the change is culturally supported and helps to reduce un- certainty in the environment. On the other hand, normative forces would encourage an organization to change out of duty or obligation, with the primary motivation being moral. These mimetic and normative forces are often very strongDurant (2000) argues that iso- morphism occurs with public sector reform even when organizations do not completely understand the reform itself. Some organizations adopt programs simply because they are perceived as new or innovative, but it is unclear whether they are fully implemented (Feller, 1981; Perry et al., 1993). In an examination of isomorphism in English local government, Ashworth et al. (2009) discuss this process of conformity as having two parts: compliance and conver- gence. Public organizations may grow to look like one another out of compliance with a lawor mandate. For example, organizations may be sanctioned on the basis of Afrmative Action or Equal Employment Opportunity violations, which could spur diversity programs as a mandated response. On the other hand, organizations can also converge for other reasons that may not be legal, or even rational, in nature (see also DiMaggio and Powell, 1983). For example, organizations may converge because of shifts in the normative envi- ronment. It may not be that organizations rationally or strategically create diversity man- agement programs to address a threat or uncertainty but rather that they implement them because peer organizations or other external actors seem to think they are socially necessary. For diversity management, we believe that this isomorphism is likely to take one of two paths. It could be that the superintendent learns about diversity management programs and initiatives through formal education. For example, managers who are more educated may be more likely to learn about the movement in the eld toward diversity management programs, which would then inuence their choice to create those programs in their dis- tricts. It could be that managers who are educated more recently are more likely to be ex- posed to diversity management since it is a relatively new management tool that may not have been a key component of older educational programs. However, it may be that di- versity management is not so much an innovation but rather a program that can result from normative pressure from ones peers. In this case, one would expect well-networked man- agers to implement diversity management as a result of pressure from one or more peers in the network. This convergence is more likely to take place as networking increases since opportunities for converge increase as the dyads grow in number. Pitts et al. What Drives the Implementation of Diversity Management Programs? 873 DATA AND METHODS We use data from the public education policy setting in order to explore these three drivers of diversity management implementation. Public education organizations provide an in- teresting laboratory for investigating diversity management. They are highly profession- alized with strong norms and values, which makes diversity management more likely to be implemented there than in policy settings with less professional training and socialization. Public education organizations tend to be relatively at and invest signicant discretion in street-level bureaucrats. Given research that demonstrates the signicant inuence that bu- reaucrat ethnicity can wield on student outcomes (Meier et al., 2000, 2006; Pitts, 2005; Pitts and Jarry, 2007), diversity management is particularly vital to organizational performance in public schools. The data for this analysis are drawn froma 2007 survey of public school district super- intendents in Texas (see, e.g., Meier and OToole, 2001). The survey was sent to super- intendents in 1,110 school districts, of whom 757 responded, producing a 68% response rate. After omitting cases with missing data, we generated 586 useable responses for our rst model and 596 for our second (see below). Our sample is representative of Texas school districts as a whole: differences between sample districts and population districts in demographics, nances, and performance are not statistically signicant. 1 The survey data are paired with other data on school district characteristics collected in 2007 by the Texas Education Agency (TEA). We indicate the source for each variable in table 2. The diversity management survey data are available only for 2007, which limits us to a cross- sectional research design. Weuse bothordinaryleast squares (OLS) andorderedprobit regressionwithrobust stan- dard errors (SEs) to estimate our model. Our dependent variables are both ordinal, which would typically require the use of an ordered regression method. However, in our case, the coefcients statistical signicance and magnitude are practically identical in OLS and ordered probit analyses. Given the relative ease of interpretation gained from OLS, we discuss those results in the body of the article, but both sets of results appear in table 4. In the section below, we identify all the variables that we use in the analysis. The theory on which we base our work does not suggest interactive or curvilinear relationships, so all these variables are included as additive terms. Many of our variables are drawn from the same survey, so it is possible that common-source bias limits the strength of our nd- ings. However, given that we have a reasonable distribution of responses in our data and recent research has demonstrated that the impacts of common-source bias are typi- cally minimal, we do not believe that this should create much of an issue for our results (Crampton and Wagner, 1994; Doty and Glick, 1998; Spector and Brannick, 1995). Finally, we recognize that endogeneity could affect our ndings. For example, it is possible that diversity management inuences some of our explanatory variables, such as nonwhite teachers and harmonious intergroup relations. We do not believe that the impacts are sub- stantial, given that most of these variables are largely determined by broader environmental factors. To the extent that diversity management affects teacher race/ethnicity and/or intergroup relations, those impacts should take place gradually over time. In the immediate 1 t Statistics for difference in means ranged from 0.21 to 1.12 on the eight variables we used to test for the representativeness of the sample. 874 Journal of Public Administration Research and Theory case, we do not have data available to create instrumental variables and use a more ad- vanced analysis. VARIABLES Diversity Management We use two dependent variables to examine howthese factors inuence the implementation of particular aspects of diversity management (table 2). To construct those variables, we used two questions from the superintendent survey. Superintendents were asked to rate the extent to which they agreed with the two statements: Promoting under-represented groups to positions of authority is a priority in my district [recruitment and retention], 2 and Table 2 Variables Concept Label/Variable Description Diversity management: recruitment and retention Recruitment and retention: Promoting under-represented groups to positions of authority is a priority in my district (superintendent survey) Diversity management: programs and initiatives Programs and initiatives: There are special programs in place in my district to manage diversity among principals, teachers, and staff (superintendent survey) Environmental uncertainty Environmental uncertainty: There is a great deal of uncertainty in the environment in which my district operates (superintendent survey) Environmental favorability: resource municence Revenue per pupil (in thousands of dollars, from TEA) Environmental favorability: capacity 1. Student enrollment (in hundreds, from TEA) 2. Superintendent tenure in district (in years, from superintendent survey) 3. Professional development: How would you rate the quality of professional development in your district? (superintendent survey) 4. District adopts new ideas: Our district is always among the rst to adopt new ideas and practices (superintendent survey) Environmental favorability: harmonious culture Harmonious culture: I would characterize relations between diverse groups in my district as harmonious (superintendent survey) Institutional isomorphism: socialization by education 1. Superintendent education (superintendent survey) 2. Superintendent time since degree (number of years since completion of highest degree, from superintendent survey) Institutional isomorphism: socialization by networking Superintendent networking, which is the composite score of responses on 10 networking dimensions:Indicate howfrequently you interact with individuals in the following groups: school board members, teachers associations, parent groups, local business leaders, other superintendents, federal education ofcials, state legislators, TEA, City/County Government, Local police/re departments (superintendent survey) Demographic control variables 1. Teacher diversity (by race/ethnicity, from TEA) 2. Student diversity (by race/ethnicity, from TEA) 2 We are unable to disentangle the extent to which responses to this question specically reect recruitment or retention since the question is worded generally to incorporate both into the same statement. Pitts et al. What Drives the Implementation of Diversity Management Programs? 875 There are special programs in place in my district to manage diversity among principals, teachers, and staff [diversity programs and initiatives]. Responses to all questions ranged from strongly disagree to strongly agree, using a four-point scale. In the context of our denition of diversity management, the rst question measures recruitment and outreach initiatives, whereas the second variable measures the existence of diversity-centered pro- grams, a broad categorization that encompasses both efforts to promote valuing differences and efforts that would be categorized as pragmatic programs. Although we are limited in the availability of data regarding the specic content of these programs, 3 these variables allow us to tap into whether the district has made efforts to engage in diversity management more generally. 4 The distribution of responses on the dependent variable was fairly broad, with mean- ingful percentages at both ends of the scale and the bulk of respondents in the middle. For our recruitment and retention variable, 9% strongly agreed, 42% agreed, 43% disagreed, and 6% strongly disagreed. For our diversity programs and initiatives variable, 13% strongly agreed, 52% agreed, 30% disagreed, and 5% strongly disagreed. There is the po- tential for bias on these questions since there are typically socially desirable responses when it comes to questions about diversity. One expects that responses about diversity management will be articially positive since managers likely believe that they should be encouraging of underrepresented groups. We do not believe that this should affect our ndings since we do not think that propensity to provide a socially desirable answer covaries with any other relevant characteristics of the respondent or organization. If social desirability were, indeed, driving the responses, we would expect to see considerably more agree responses than disagree, which does not appear to be the case. For these data, we note that the diversity programs and initiatives variable is only slightly skewed, whereas the recruit and retention variable is fairly close to a normal distribution. 5 Environmental Uncertainty Our rst driver of diversity management speculates that the decision to engage in these programs may be a function of uncertainty in the environment. We measure this as the extent to which the superintendent agrees with the statement, There is a great deal of uncertainty in the environment in which my district operates. Since contingency theorists argue that managers make decisions on the basis of the enacted environment, perceptions are much more useful in empirical research than more objective measures of environmental uncertainty. Responses were distributed across all four categories (strongly disagree, dis- agree, agree, strongly agree), with a mean value just over 2. Environmental Favorability Our second driver links diversity management efforts to factors that would contribute to a more favorable environment, hypothesizing that organizations engage in diversity 3 Our interest is primarily in exploring the factors that lead to the presence of diversity management practices. Future research should investigate the factors that lead to different types of diversity management practicesrecruitment/ retention, valuing differences, and the more pragmatic policies and programs. 4 Additionally, these questions do not allowus to tap into the extent to which these programs have been implemented throughout the district. However, the wording of the questions (programs in place, promoting . . . is a priority) suggests that these initiatives have already been implemented, at least to some extent. 5 The p values for normality are .02 and .09, respectively. 876 Journal of Public Administration Research and Theory management when it is easiest to do so. We include a series of six variables that capture different aspects of the larger concept of environmental favorability: four variables that measure organizational capacity, one variable that measures resources, and one variable that reects harmonious relationships in the environment. Four variables measure organizational capacity, which in our case reects both inter- nal and external factors. One of these variables is drawn from a question that asks super- intendents to rate the extent to which the school district is often among the rst districts to adopt new ideas. We also include variables for organizational size (student enrollment) and managerial tenure in the organization (the number of years that the superintendent has been employed by the district). Given that the superintendent serves at the pleasure of the school board, longer tenure often results in a looser rein for implementing new programs like di- versity management. The last variable is a basic rating of the quality of the professional development in the district. This variable taps into the concept of slack resources since districts with large professional development programs are those that are most likely to have the slack resources necessary to implement a program like diversity management. We also include a more explicit measure of resources by controlling for nancial wealth, which we measure as revenue per pupil. Finally, our variable for environmental harmony captures the extent to which the organization is already diversity friendly. This variable represents the response to the survey questions: I would characterize relations between diverse groups in my district as harmonious. Institutional Isomorphism Our third driver speculates that the likelihood to implement diversity management pro- grams will be a function of the norms to which the executive manager is exposed. The motivations behind isomorphism cannot be measured directed, so we use two proxy meas- ures that reect key aspects of isomorphism in our policy context. We are particularly in- terested in the possibility that the education and socialization of managers may be a key explanatory factor in diversity management as many graduate programs in education have recently incorporated issues related to diversity management into the curriculum. To ex- plore the effect of professional norms, we include two variables: whether the superinten- dent holds a doctorate (a dichotomous variable where 1 represents having a doctorate), and the time since the superintendent has received the most recent degree (measured in years). We would anticipate that those with doctorates and those who have completed grad- uate work more recently would be more likely to hold those norms that would lead the manager to pursue diversity initiatives. Finally, we are also interested in exploring whether exposure to other organizations is positively related to diversity management. Here, we hypothesize that as managers become more engaged in their environment, including interactions with other districts and local organizations, they will be more likely to engage in diversity management. Managers who interact with others in the external environment are more likely to learn about new ideas and innovations, as well as feel social pressure to implement programs that are being developed in other districts. Our measure of interaction with other groups is a composite score that combines the level of activity between the superintendent and the superintendents network, a measure that has been used in a number of previous studies (see, e.g., Hicklin, 2004; Meier and OToole, 2001; Fernandez and Pitts, 2007). This mea- sure includes nine items, representing the level of interaction with other superintendents, Pitts et al. What Drives the Implementation of Diversity Management Programs? 877 teachers associations, parent groups, local business leaders, federal education ofcials, state legislators, TEA, city and county government, and local police and re departments (Cronbachs alpha: .742). Control Variables We include two variables that measure the racial/ethnic diversity of student and teachers. We use a Blau (1977) index of dissimilarity, which is a common means of measuring var- iation in a categorical variable. The Blau index ranges from 0 to 1, where 0 represents perfect homogeneity and 1 represents perfect heterogeneity with an innite number of cat- egories. In our case, with ve categories, the practical limit is 0.8. For example, a school district with all African-American students would score a 0 since it would be perfectly homogeneous. A school district with students evenly distributed among all racial/ethnic groups would score 0.8. This permits us to gauge the extent to which the environment could produce diversity-oriented conict. We conjecture that organizations with high student and teacher diversity will see diversity management as an important and necessary response to these environment characteristics. Not surprisingly, the average district was more diverse along student lines than among teachers (table 3). FINDINGS Results from our data analysis are presented in Table 4. The rst two columns show results from our model predicting recruitment and retention, whereas the second two columns show results from the model for diversity programs and initiatives. Within each, the rst column of results shows coefcients and SEs from the OLS regression, and the second shows results from the ordered probit analysis. Given the similarities between the two sets of results, we discuss the OLS coefcients for ease of interpretation. Environmental uncertainty appears to be related to programs and initiatives more than recruitment and retention. Increasing the perceived uncertainty of the environment by one point on a four-point scale corresponds to a 0.132-point increase in diversity programs and initiatives, holding constant the other independent variables in the model. This is the equiv- alent of changing ones response from agree to strongly agree. The largest possible shift moving from strongly disagree to strongly agree or three pointswould correspond to a 0.396-point shift in the dependent variable or about half of 1 standard deviation (SD). This relationship is not substantial, but it is not trivial, either. These results suggest that organizations in environments marked by uncertainty may be more likely to adopt and/ or implement programs aimed at managing diversity. The coefcient is statistically signicant in the model for programs and initiatives, but it is not statistically signicant in the model for recruitment and retention. Uncertainty is, perhaps, more likely to be important to the decision to create broader, more strategic pro- grams and initiatives than narrowly tailored recruitment and retention policies. Large-scale diversity programs and initiatives are more likely than recruitment efforts to garner the attention of other actors in the environment, which would enhance the school districts power and potential resource base (Pfeffer and Salancik, 1978). It is also possible that school districts are attempting to exploit diversity-related possibilities in the environment. If organizations view diversity as an environmental issuewhether positive or negative then creating broad policies and programs may make it able to respond more quickly to the changes that diversity may introduce. 878 Journal of Public Administration Research and Theory Table 3 Descriptive Statistics and Correlations Variable Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Recruitment and retention 2.535 0.732 2. Programs and initiatives 2.728 0.752 0.287*** 3. Teacher diversity 0.113 0.114 0.196*** 0.084** 4. Student diversity 0.228 0.114 0.289*** 0.076* 0.256*** 5. Environmental uncertainty 2.077 0.726 0.040 0.050 0.136*** 0.077 6. Harmonious culture 3.162 0.543 0.003** 0.271*** 20.064* 20.052* 20.263*** 7. District adopts new ideas 2.667 0.653 0.103** 0.232*** 0.084* 20.095 20.047 0.157*** 8. Professional development 3.800 0.765 0.112 0.302*** 0.080* 20.027 20.204*** 0.267*** 0.297*** 9. Revenue per pupil 10.037 3.504 20.061* 20.028 20.042*** 20.078** 20.029 0.068 20.020 20.083* 10. Superintendent tenure in district 4.853 4.393 0.112 0.128** 20.022 20.028 20.046 0.126* 0.046 0.220*** 0.027 11. Student enrollment (in hundreds) 37.454 116.266 0.241*** 0.128*** 0.232*** 0.149*** 0.046 20.031* 0.083** 0.164*** 20.134*** 20.004* 12. Superintendent education 0.341 0.475 0.162*** 0.044 0.084*** 0.154*** 0.114 20.042 0.037 0.000 20.147** 0.024* 0.240*** 13. Superintendent time since degree 18.665 10.249 20.062 0.064 0.024 20.089 20.012 0.116 20.075* 0.112** 20.022 0.301*** 0.001 20.275*** 14. Superintendent networking 0 0.878 0.147*** 0.162*** 20.004 0.144** 0.023 0.094* 0.049* 0.117** 20.018*** 20.087 0.117** 0.130*** 20.088 Statistical signicance: *p , .05, **p , .01, ***p , .001. 8 7 9 Table 4 Regression Results Recruitment and Retention Programs and Initiatives OLS Ordered Probit OLS Ordered Probit Environmental uncertainty Environmental uncertainty 0.011 (0.044) 0.028 (0.074) 0.132** (0.043) 0.223** (0.072) Environmental favorability Harmonious culture 20.015 (0.065) 20.025 (0.109) 0.289** (0.065) 0.499** (0.110) District adopts new ideas 0.073 (0.048) 0.121 (0.081) 0.181** (0.048) 0.305** (0.081) Professional development 0.047 (0.043) 0.072 (0.073) 0.183** (0.042) 0.302** (0.069) Revenue per pupil (thousands) 0.004 (0.010) 0.008 (0.017) 0.003 (0.012) 0.007 (0.020) Superintendent tenure in district 0.009** (0.003) 0.016** (0.005) 0.005 (0.002) 0.008 (0.004) Student enrollment (hundreds) 0.001** (0.0002) 0.002** (0.0006) 0.0006* (0.0003) 0.001* (0.0005) Institutional isomorphism Superintendent education (1 5 doctorate) 0.086 (0.065) 0.131 (0.110) 0.041 (0.062) 0.065 (0.101) Superintendent time since degree (years) 20.005 (0.003) 20.009 (0.005) 0.003 (0.003) 0.005 (0.005) Superintendent networking 0.097** (0.031) 0.163** (0.152) 0.097** (0.036) 0.157** (0.059) Demographic control variables Teacher diversity (Blau index) 0.527 (0.413) 0.843 (0.690) 0.030 (0.383) 0.026 (0.631) Student diversity (Blau index) 1.544** (0.283) 2.571** (0.494) 0.507 (0.270) 0.833 (0.446) Constant 1.667** (0.282) 0.050 (0.281)
N 586 586 596 596
Adjusted R 2 /pseudo R 2 .174 .088 .205 .104 Cut point 1 20.369 2.669 Cut point 2 1.429 4.084 Cut point 3 2.983 5.819 Note: Robust SEs in parentheses. Statistical signicance: *p , .05, **p , .01, two-tailed tests. 880 Journal of Public Administration Research and Theory Findings are mixed for our second potential driver of diversity management, environ- mental favorability. The harmonious culture variable is statistically signicant for diversity programs and initiatives but not for recruitment and retention. These results suggest that school districts where employees value ethnic and cultural differences may be more likely to use diversity programs and initiatives but no more likely than others to engage in re- cruitment and retention. This makes sense, given that environments marked by such har- mony and cultural awareness probably do not suggest a need for strategic recruitment and retention initiatives. The organization may already be as diverse as it needs to be, given the diversity present in the target population, making strategic recruitment less important. In addition, if the environment is harmonious with regard to racial and ethnic differences, retention is likely not an issue that requires a formal program. These organizations may choose instead to devote their efforts toward diversity programs and initiatives that continue that harmony and ensure that differences continue to be valued. This relationship is one of the strongest substantive ndings across both models: Increasing the harmonious culture variable by one point corresponds to a 0.289-point shift in diversity programs and initiatives. Our results also suggest that organizational capacity may affect broad programs and initiatives more than recruitment and retention efforts. A one-point increase in the extent to which the district adopts new ideas corresponds to a 0.181-point increase in diversity pro- grams and initiatives, and a one-point increase in the quality of professional development corresponds to a 0.183-point increase in diversity programs and initiatives. These relation- ships are moderate in scope, but they demonstrate that a solid infrastructure for professional development and innovation may be important in implementing broad diversity programs and policies. Both these variables are statistically signicant at the .01 level in the programs and initiatives model, but neither is signicant in the recruitment and retention model. Our primary resource municence variable, average revenue per pupil, is not statis- tically signicant for either recruitment and retention or diversity programs and initiatives. It does not seem to matter whether organizations have abundant resources for developing diversity management programs. We considered that this could be the result of a method- ological issueperhaps revenue per pupil does not measure resource municence as well as a different measure. However, we are unable to nd any measure of resources that cor- responded to diversity management in either direction, positive or negative. Multicollinear- ity does not seem to be an issue as the highest variance ination factor was 1.28. Perhaps the role of resources is more indirect, and it really affects diversity management through other factors, such as professional development opportunities or the adoption of innovative ideas. Ultimately, we believe that resource municence is a factor that should be taken into account in this line of research, but we do not nd it to be important in this sample of organizations. Our nal driver of implementation is a set of isomorphism variables that suggest that organizations adopt diversity management measures as a means of mimicking the strategies of other school districts or inuential actors in the environment. This is tested through three variables: whether the superintendent holds a doctorate, how recently the superintendent completed his or her highest degree, and the extent to which the superintendent interacts with a series of external actors. Support for this driver is mixed. Neither of the two edu- cation variables is statistically signicant. Superintendents who earned doctorates are no more likely to engage in diversity management than those who earned masters degrees. Pitts et al. What Drives the Implementation of Diversity Management Programs? 881 One might expect that well-educated superintendents would be more likely than others to be exposed to diversity management or that superintendents educated more recently would be more likely to be exposed to diversity management, but our results do not demonstrate this to be the case. On the other hand, our results suggest that interaction with actors in the external environment may lead organizations to engage in diversity management. The co- efcient for management networking is positive and statistically signicant at the .01 level for both recruitment and retention strategies and diversity programs and initiatives. The substantive relationships are roughly the same in both models: a 1 SD increase in network management corresponds to roughly a 0.1-point increase in diversity management. This variable incorporates networking with multiple nodes, and some nodes are undoubtedly more inuential than others. In auxiliary analyses, we tested each node separately. All were positively related to diversity management implementation, but teacher associations and local business leaders appeared to be the most inuential across both models. Future re- search should consider how different actors in the network inuence diversity management practices. These results suggest that isomorphic factors may not differ between the two com- ponents of diversity management. There are no differences between the isomorphic forces that correspond to recruitment and retention strategies and those that correspond to diver- sity programs and initiatives. Neither education variable is signicant in either model, but network management is signicant for both and has roughly the same coefcient for both. Although consistent, these results do not suggest that isomorphismis a primary explanation behind the implementation of diversity management. Only one of three variables is sig- nicant, and its substantive signicance is somewhat weak. It is possible that isomorphism operates through other avenuesperhaps the convergence of norms is not reected by the education variables that we use as proxy measures in our analysis (Ashworth et al., 2009). Future research should work to further clarify how isomorphic pressures operate in the context of diversity. CONCLUSION This analysis demonstrates that diversity management springs froma series of complex and nuanced environmental factors. It is not possible to directly compare the three drivers of diversity management implementation that we formulate, but we do show varying levels of support for each. It seems that isomorphic inuences are the least important, but environ- mental favorability and uncertainty are both relevant predictors. Motivations are different for recruitment and retention than for diversity programs and initiatives. Results indicate that organizations appear to be responding to environmental issues. For example, organizations with high student and teacher diversity are most likely to create recruitment and retention strategies, which is a good means of directly addressing these environmental demographics. We also nd that organizations in uncertain environments are more likely to implement broad diversity programs and initiatives. This should be an effective approach, insofar as broad and overarching diversity policies may permit the or- ganization to better adapt to environmental changes. Future research should examine the impact of the environment more explicitly through a resource dependence lens since it seems possible that organizations may implement diversity management programs as a means of gaining legitimacy with actors in the environment who believe that such pro- grams are important. It is beyond the scope of this article to test the relationship between 882 Journal of Public Administration Research and Theory diversity management and performance, but previous research suggests that these actions may be the best way for organizations to avoid the pitfalls of diversity while taking ad- vantage of its benets (Pitts, 2009). Practicing managers might use these results to think about how the environments in which they operate could affect their implementation of diversity management initiatives. For example, managers may choose to adopt diversity management if their organization is similar to those that already have. These results show that organizations that frequently adopt new ideas and already have strong professional development programs seem more likely to adopt broad policies on diversity. A harmonious culture with regard to diversity seems to encourage diversity programs and policies, whereas teacher and student diversity appear to encourage recruitment and retention programs. To the extent that managers wish to align themselves with other organizations in their environment, these ndings give them a glimpse into the relevant factors to consider. These conclusions must be considered in the context of a few methodological caveats. Future research should consider ways of expanding upon our approach to include a more nuanced and comprehensive measurement of diversity management. Our dependent var- iables reect responses to only two survey questions, and our ndings would be reinforced by a more complete measure of diversity management implementation or adoption. Several of our measures are attitudinal, and there is the potential for social desirability bias that could yield inconclusive results. We are also limited by data from one policy area, which limits the generalizability of our ndings. These results would be strengthened by similar ndings in another policy setting or type of organization. Although we believe our results are robust and meaningful for other education organizations, we cannot claimwith certainty that they would operate the same for organizations that were less professionalized, more hierarchical, or more routinized. Our models are able to explain only about one-fth of the variation in diversity management implementation, indicating that factors that we did not include in our model could be important inuences. The eld should also work to develop panel data sets that make it possible to test these relationships as they shift over time since cross-sectional data are limited in their ability to demonstrate causal relationships. Finally, the eld would benet from additional qualitative inquiry about diversity management. Well-executed case studies would improve our ability to make causal assumptions and complement the quantitative evidence we bring to bear in this article. Future research should consider how diversity management programs conspire with the environment to produce particular performance outcomes. If diversity management is a key moderator between diversity and performance, how might these environmental motivations affect that relationship? For example, if diversity management is a response to environmental threat and uncertainty, does it lead to performance gains that are higher than in scenarios where environmental threats and uncertainty do not exist? The role of the organizational environment is crucial to appropriately modeling the relationship between diversity management and organizational performance. REFERENCES Adler, N. J. 2002. International dimensions of organizational behavior, 4th ed. Cincinnati, OH: Southwestern College. Aldrich, H. E. 1979. Organizations and environments. Englewood Cliffs, NJ: Prentice Hall. Aldrich, H. E., and S. Mindlin. 1978. Uncertainty and dependence: Two perspectives on environment. In Organization and environment, ed. L. Karpik, 14970. Beverly Hills, CA: Sage. Pitts et al. What Drives the Implementation of Diversity Management Programs? 883 Allport, G. W. 1954. The nature of prejudice. Boston, MA: Addison-Wesley. Andrews, R., G. A. Boyne, K. J. Meier, L. J. OToole, and R. Walker. 2005. Representative bureaucracy, organizational strategy and public service performance: An empirical analysis of English local governments. Journal of Public Administration Research and Theory 15:489504. Arai, M., D. M. Wanca-Thibault, and D. P. Shockley-Zalabak. 2001. Communication theory and training approaches for multiculturally diverse organizations: Have academics and practitioners missed the connection? Public Personnel Management 30:445445. Ashworth, R., G. Boyne, and R. Delbridge. 2009. Escape from the iron cage? Organizational change and isomorphic pressures in the public sector. Journal of Public Administration Research and Theory 19:16587. Azevedo, A., E. A. Drost, and M. R. Mullen. 2002. Individualism and collectivism: Toward a strategy for testing measurement equivalence across culturally diverse groups. Cross Cultural Management 9:1929. Bertelli, A. M. 2007. Determinants of bureaucratic turnover intention: Evidence from the department of the treasury. Journal of Public Administration Research and Theory 17:23558. Blau, P. M. 1977. Inequality and heterogeneity. New York: Free Press. Boyne, G. A. 2003. Sources of public service improvement: Acritical reviewand research agenda. Journal of Public Administration Research and Theory 13:36794. Cox, T. 1993. Cultural diversity in organizations. San Francisco, CA: Berret-Koehler. Crampton, Suzanne M., and John A. Wagner. 1994. Percept-percept ination in microorganizational research: An investigation of prevalence and effect. Journal of Applied Psychology 79:6776. Daft, R. L. 2008. Organization theory and design. Cincinnati, OH: Thomson. Dess, G. G., and D. W. Beard. 1983. Dimensions of organizational task environment. Administrative Science Quarterly 29 (1): 5273. . 1984. Dimensions of organizational task environments. Administrative Science Quarterly 29:52. DiMaggio, P. J., and W. W. Powell. 1983. The iron cage revisited: Institutional isomorphism and col- lective rationality in organizational elds. American Sociological Review 48:14760. . 2001. The new institutionalism in organizational analysis. Chicago, IL: Univ. of Chicago Press. Donahue, J. D., and J. S. Nye. 2002. Market-based governance. Washington, DC: Brookings Institution Press. Doty, D. Harold., and William H. Glick. 1998. Common methods bias: Does common methods variance really bias results? Organizational Research Methods 1:374406. Duncan, R. B. 1972. Characteristics of organizational environments and perceived environmental un- certainty. Administrative Science Quarterly 17:31327. Durant, R. F. 2000. Whither the neoadministrative state? Toward a polity-centered theory of adminis- trative reform. Journal of Public Administration Research and Theory 10:79109. Ely, R. J., and D. A. Thomas. 2001. Cultural diversity at work: The effects of diversity perspectives on work group processes and outcomes. Administrative Science Quarterly 46:22973. Ezra, M., and M. Deckman. 1996. Balancing work and family responsibilities: Flextime and child care in the federal government. Public Administration Review 56:1749. Feller, I. 1981. Public sector innovation as conspicuous production. Policy Analysis 7 (1): 120. Fernandez, S., and D. W. Pitts. 2007. Under what conditions do public managers favor and pursue or- ganizational change? American Review of Public Administration 37:32441. Fernandez, S., and H. G. Rainey. 2006. Managing successful organizational change in the public sector. Public Administration Review 66:16885. Foldy, E. G. 2004. Learning from diversity: A theoretical exploration. Public Administration Review 64:52938. Galbraith, J. R. 1973. Designing complex organizations. Boston, MA: Addison-Wesley. Goerdel, H. T. 2006. Taking initiative: Proactive management and organizational performance in net- worked environments. Journal of Public Administration Research and Theory 16:35167. Hatch, M. J., and A. L. Cunliffe. 2006. Organization theory. Oxford, NY: Oxford Univ. Press. Hicklin, A. K. 2004. Network stability: Opportunity or obstacle? Public Organization Review 4:157398. 884 Journal of Public Administration Research and Theory . 2007. The effect of race-based admissions in public universities: Debunking the myths about hopwood and proposition 209. Public Administration Review 67:33140. Ho, M. K. 1987. Family therapy with ethnic minorities. Newbury Park, CA: Sage. Jayne, M. E. A., and R. L. Dipboye. 2004. Leveraging diversity to improve business performance: Research ndings and recommendations for organizations. Human Resource Management 43:40924. Kalev, A., F. Dobbin, and E. Kelly. 2006. Best practices or best guesses? Assessing the efcacy of corporate afrmative action and diversity policies. American Sociological Review 71:589617. Kellough, J. E. 1990. Integration in the public workplace: Determinants of minority and female em- ployment in federal agencies. Public Administration Review 50:557. Kellough, J. E., and K. C. Naff. 2004. Responding to a wake-up call: An examination of federal agency diversity management programs. Administration Society 36:6290. Kellough, J. E., and D. H. Rosenbloom. 1992. Representative bureaucracy and the EEOC: Did civil service reform make a difference? In The promise and paradox of civil service reform, eds. P. W. Ingraham and D. H. Rosenbloom, 24566. Pittsburgh, PA: Univ. of Pittsburgh Press. Lawrence, P. R., and J. W. Lorsch. 1967. Differentiation and integration in complex organizations. Administrative Science Quarterly 12 (1): 112. Lynn, L. E., C. J. Heinrich, and C. J. Hill. 2001. Improving governance. Anewlogic for empirical research. Washington, DC: Georgetown Univ. Press. March, J. G., and J. P. Olsen. 1989. Rediscovering institutions. The organizational basis of politics. New York: Free Press. Meier, K. J. 1975. Representative bureaucracy: An empirical analysis. American Political Science Review 69:526. Meier, K. J., C. Doerer, D. Hawes, A. K. Hicklin, and R. R. Rocha. 2006. The role of management and representation in improving performance of disadvantaged students: An application of Bum Phillipss Don Shula Rule. Review of Policy Research 23:1095110. Meier, K. J., and L. J. OToole, Jr. 2001. Managerial strategies and behavior in networks: A model with evidence from U.S. public education. Journal of Public Administration Research and Theory 11:27194. . 2003. Public management and educational performance: The impact of managerial networking. Public Administration Review 63:68999. . 2005. Managerial networking: Issues of measurement and research design. Administration and Society 37:52341. Meier, K. J., J. L. Polinard, and R. D. Wrinkle. 2000. Bureaucracy and organizational performance: Causality arguments about public schools. American Journal of Political Science 44:590602. Mosher, F. C. 1982. Democracy and the public service, 2nd ed. Oxford, NY: Oxford Univ. Press. Naff, K. C. 2001. To look like America: Dismantling barriers for women and minorities in government. New York: Westview. Naff, K. C., and J. E. Kellough. 2003. Ensuring employment equity: Are federal diversity programs making a difference? International Journal of Public Administration 26:1307. Ng, E. S. W., andR. J. Burke. 2005. Personorganizationt andthewar for talent: Does diversitymanagement make a difference? International Journal of Human Resource Management 16:1195210. OToole, L. J., Jr., and K. J. Meier. 1999. Modeling the impact of public management: Implications of structural context. Journal of Public Administration Research and Theory 9:50526. Perry, J. L., K. L. Kraemer, D. Dunkle, and J. King. 1993. Motivations to innovate in public organizations. In Public management: The state of the art, ed. B. Bozeman. San Francisco, CA: Jossey-Bass. Pfeffer, J., and G. R. Salancik. 1978. The external control of organizations: A resource dependence perspective. New York: Harper and Row. Pitts, D. W. 2005. Diversity, representation, and performance: Evidence about race and ethnicity in public organizations. Journal of Public Administration Research and Theory 15:61531. . 2006. Modeling the impact of diversity management. Review of Public Personnel Administration 26:24568. . 2007. Implementation of diversity programs in public organizations: Lessons from policy im- plementation research. International Journal of Public Administration 30:157390. Pitts et al. What Drives the Implementation of Diversity Management Programs? 885 . 2009. Diversity management, job satisfaction, and organizational performance: Evidence from U.S. federal agencies. Public Administration Review 69:32838. Pitts, D. W., and E. M. Jarry. 2007. Ethnic diversity and organizational performance: Assessing diversity effects at the managerial and street levels. International Public Management Journal 10:23354. Putnam, R. D. 2007. E Pluribus Unum: Diversity and community in the twenty-rst century. Scandinavian Political Studies 30:13774. Rainey, H. G. 2003. Understanding and managing public organizations, 3rd ed. San Francisco, CA: Jossey-Bass. Riccucci, N. M. 1997. Cultural diversity programs to prepare for work force 2000: Whats gone wrong? Public Personnel Management 26 (1): 3541. . 2002. Managing diversity in public sector workforces. Boulder, CO: Westview Press. Richard, O., A. McMillan, K. Chadwick, and S. Dwyer. 2003. Employing an innovation strategy in racially diverse workforces: Effects on rm performance. Group Organization Management 28:10726. Richard, O. C. 2000. Racial diversity, business strategy, and rm performance: A resource-based view. Academy of Management Journal 43:16477. Rubaii-Barrett, N., and L. R. Wise. 2007. From want ads to web sites: What diversity messages are state governments projecting? Review of Public Personnel Administration 27:2138. Salzstein, A. L., Y. Ting, and G. H. Salzstein. 2001. Work-family balance and job satisfaction: The impact of family-friendly policies on attitudes of federal government employees. Public Administration Review 61:45267. Sanchez, J. I., and P. Brock. 1996. Outcomes of perceived discrimination among hispanic employees: Is diversity management a luxury or a necessity? Academy of Management Journal 39:70419. Scott, W. R. 2003. Organizations: Rational, natural, and open systems, 5th ed. Upper Saddle River, NJ: Prentice Hall. Selden, S. C. 1997. The promise of representative bureaucracy: Diversity and responsiveness in a gov- ernment agency. Armonk, NY: M.E. Sharpe. . 2006. A solution in search of a problem? Discrimination, afrmative action, and the new public service. Public Administration Review 66:91123. . 2008. Human capital: Tools and strategies for the public sector. Washington, DC: CQ Press. Selden, S. C., and F. Selden. 2001. Rethinking diversity in public organizations for the 21st century: Moving toward a multicultural model. Administration Society 33:30329. Selznick, P. 1957. Leadership and administration. New York: Harper Collins. Spector, Paul E., and Michael T. Brannick. 1955. The nature and effects of method variance in orga- nizational research. In International Review of Industrial and Organizational Psychology, eds. Cary L. Cooper and Ivan T. Robertson, 24974. West Sussex, UK: John Wiley. Swann, W. B., J. T. Polzer, D. C. Seyle, and S. J. Ko. 2004. Finding value in diversity: Verication of personal and social self-views in diverse groups. Academy of Management Review 29:927. Thomas, D. A., and R. J. Ely. 1996. Making differences matter: A new paradigm for managing diversity. Harvard Business Review 74:7990. Thomas, R. R. 1990. From afrmative action to afrming diversity. Harvard Business Review 90:1108. Timmerman, T. A. 2000. Racial diversity, age diversity, interdependence, and team performance. Small Group Research 31:592606. Watson, W. E., K. Kumar, and L. K. Michaelsen. 1993. Cultural diversitys impact on interaction process and performance: Comparing homogeneous and diverse task groups. Academy of Management Journal 36:590602. Weick, K. E. 1979. The social psychology of organizing. New York: McGraw-Hill. Welch, S., and L. Sigelman. 2000. Getting to know you? Latino-Anglo social contact. Social Science Quarterly 81:67. Wise, L. R., and M. Tschirhart. 2000. Examining empirical evidence on diversity effects: How useful is diversity research for public-sector managers? Public Administration Review 60:38694. 886 Journal of Public Administration Research and Theory Copyright of Journal of Public Administration Research & Theory is the property of Oxford University Press / UK 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.