The Effect of the External Environment on Bureaucratic Representation: Assessing the Passive to Active Representation Link

Abstract

The literature on the link between passive and active representation is well established, but there is still some debate on why and how passive representation translates into active representation in some settings and not others. Some scholars suggest the salience of the policy issue and discretion to act yields a linkage between passive and active representation, while others contend researchers are simply identifying correlates of the two concepts not linkages. However, many studies miss the dynamic relationship between representation and the external environment, so this research explains the manner in which external, environmental pressures influence bureaucratic discretion and active representation. Interdependences between service organizations and the external environment can restrict bureaucratic discretion through resource dependence and cues of expectations and professional norms, or bureaucrats may use the relationship to enhance their discretion. Using an original dataset of the 1,800 largest U.S. school districts over two school terms, I compare the level of disparity in academic grouping across racial groups in two distinct external environments, racially balanced and imbalanced school districts, and find that minority bureaucrats use their discretion to improve the odds of minority students’ assignment to gifted education, but external, environmental control moderates the significance of active representation. The passive to active representation link is weakened based on the operating policy environment. This work offers some insight on the restrictions and limits to active representation, and the empirical findings hold implications for the governance of public organizations and the outcomes of public policies.

Keywords

representative bureaucracy organizational theory education policy desegregation

Although there is an abundance of research on representation in the bureaucracy, questions still remain about the process of representation, the extent to which passive representation actually translates into active representation, and our ability to accurately link the two concepts empirically. Three main arguments hold concerning the passive to active representation linkage. First, many scholars contend that when appropriate conditions are met for bureaucrats—that is, when bureaucrats hold discretion to act on behalf of their identity group, when the professional norms of the agency are less likely to infringe on this discretion, and when the policy issue is salient to the identity group being represented—passive representation translates into active representation (Keiser, Wilkins, Meier, & Holland, 2002; Meier, 1993). Evidence of this process is well established in the research literature (Bradbury & Kellough, 2008; Hindera, 1993; Selden, 1997; Wilkins & Keiser, 2006). Nevertheless, many other scholars argue that passive representation does not translate into active representation and that those who find linkages are simply finding correlates of a passive and active representation link (Lim, 2006; Rehfuss, 1986). The third perspective of the linkage is that passive representation can “translate” into active representation, but the relationship is conditional, solely based on the circumstances of the agency, and less on the discretions and behaviors of the bureaucrat (Hindera & Young, 1998; Rosenthal & Bell, 2003). Favorable conditions in an agency such as having a plurality, majority, or critical mass, for example, provide greater opportunities for passive representation to translate into active representation (Hindera & Young, 1998; Rosenthal & Bell, 2003).

The research supporting these three perspectives on the relationship between passive and active representation are well developed, but they also offer less insight on the broader mechanisms that may affect agencies and the manner in which agencies’ bureaucrats function. The open systems theory of organizations suggests that the external (policy) environment can greatly influence organizational behavior and outcomes, primarily through resources, norms, and support (Shafritz, Ott, & Jang, 2005). As such, external, contextual influences may actually facilitate or mitigate the relationship between passive and active representation and better explain why the linkage occurs in some settings among different levels of bureaucrats (i.e., street-level vs. upper level) but fails to occur in others—even when the theorized conditions of active representation (i.e., discretion, salience) are met. This article seeks to answer the following research questions:

Research Question 1: Why does the linkage between passive and active representation vary across settings?

Research Question 2: What effect does the external environment have on bureaucrats’ active representation?

External environments may affect active representation in two ways. First, the external environment may act as a source of external control due to an organization’s dependence on the external environment for resources, support, and cues of expectations and norms for bureaucrats. Here, active representation may differ based on the external environment, given its level of influence over bureaucratic behavior. On the contrary, the external environment may actually be a source of bureaucratic empowerment in which bureaucrats use their unique relationship with the external environment to enhance bureaucratic discretion and strengthen bureaucratic power and influence in policy implementation, independent of external control. This study explores these two possibilities and the role of the external environment in moderating the linkages between passive and active representation.

The research begins with an examination of previous scholarship on representative bureaucracy and a discussion of representation in the context of the open systems theory and bureaucratic discretion follows. Empirical analyses using two years of national school district data are conducted: First, a test of differential policy outcomes based on race in racially balanced¹ and racially imbalanced school districts to demonstrate a baseline difference in the two types of external environments, and then a test of the effect of bureaucratic representation on these policy outcomes in racially balanced and imbalanced districts. The second test helps to determine the extent to which the passive to active representation linkages are a function of the external environment’s influence, which is assessed as the level of racial balance in a district. Results indicate that while passive representation does empirically lead to active representation as previous scholarship suggests, the external environment can influence bureaucratic representation and yield different levels of active representation across organizations. The article closes with a discussion of the implications of the empirical findings.

Representative Bureaucracy

The theory of representative bureaucracy argues that the demographic makeup of public agencies should reflect the clientele they serve (Kingsley, 1944). More demographically diverse agencies, scholars contend, are more responsive to the public’s needs and demands because people with shared demographics tend to share values and interests through similar socialization processes and experiences (Krislov, 1974). Bureaucrats are assumed to use these values in their discretionary decision-making process (Dolan & Rosenbloom, 2003; Selden, 1997). Mosher (1968) extends the theory and suggests that agencies can be representative of their diverse clientele in two ways: passive representation or active representation. Passive representation is defined as the extent to which a bureaucratic agency’s employee diversity is proportional to the group’s share in the population. Active representation occurs when a bureaucrat uses his or her position to “press for the interests and desires of those whom he/she is presumed to represent” (Mosher, 1968, p. 11). Scholars have developed a rich body of research that seeks to identify passive and active representation in bureaucratic agencies and the extent to which the concepts matter for service delivery or substantive outcomes such as access to government jobs, housing, or recreational facilities (Esman, 1997). Despite the evidence of the positive effects of passive and active representation (Bradbury & Kellough, 2008; Hindera, 1993; Wilkins & Keiser, 2006), some scholars suggest that there is no relationship between representation and outcomes (Wilkins & Williams, 2008, 2009), or even a negative relationship between the two (Hur, 2013; Nielsen & Wolf, 2001). Such findings may be expected based on Thompson’s (1976) and Meier’s (1993) conditions for the potential of passive representation to lead to active representation. Both contend that a series of conditions must be met for the two concepts to work together to produce substantive outcomes: when groups and institutions recognize and “press” for minority interests; when issues hold obvious ramifications for minority bureaucrats’ groups; when minority bureaucrats hold positions in which they are able to influence policy outcomes through discretion, support, or mobilization; and when the institution holds jurisdiction over an area in which policy outcomes can actually affect the represented group such as social welfare or education (Meier, 1993; Thompson, 1976). Lim (2006) and Selden (1997) suggest two additional conditions for the transition from passive representation to active representation to occur. If bureaucrats believe the costs are high to representation—that is, representing one group may isolate, restrict, or harm another group—active representation may not occur (Lim, 2006). On the contrary, when bureaucrats see themselves as advocates of minority rights or needs, they may act on this feeling, leading to the passive to active representation link (Selden, 1997).

The Importance of Context in Representative Bureaucracy

Despite the abundance of representation literature, scholars know much less about the manner in which the external environment or context influences the process. A growing sector of the representation literature, however, has begun to focus on this relationship. For example, organizations structured to allow bureaucrats greater discretion, those with a more decentralized system of norms and values, and organizations with less hierarchy are more likely to experience the passive to active representation link (Keiser et al., 2002; Meier & Bohte, 2001; Roch & Pitts, 2012). Other scholars point out that factors external to the organization also affect representation. Political and economic environments can influence bureaucratic discretion in a way that may alter the relationship between passive and active representation (Keiser, 1999). In addition, external factors such as the region in which representation occurs and the racial context also appear to directly influence the passive to active representation transition (Grissom, Nicholson-Crotty, & Nicholson-Crotty, 2009; Roch & Edwards, 2017). Black bureaucrats in the south (Grissom et al., 2009) and bureaucrats of more racially homogeneous areas (Roch & Edwards, 2017) are more likely to actively represent clients of their racial group. Such studies suggest that perhaps the benefits of minority representation are contingent on the space in which representation operates.

The current research builds on the previous work as it explores this possibility, focusing on the extent to which external control shapes the level of representation in organizations and consequently organizational outcomes. It makes a few key contributions. First, answering this question may help to extend the study of representative bureaucracy in providing a more clear understanding of why passive representation leads to active representation in some instances, but fails to translate in others, often in the same or similar organizational venues. Applying the rich organizational theory, external control, and bureaucratic control literatures to representative bureaucracy offers an alternate lens to explore the conditions in which active representation is likely to occur. This study may also offer insights on the additional controls on bureaucratic discretion as it explores the manner in which discretion operates with external environmental pressures.

Second, while previous scholars have explored similar questions, many of the analyses are restricted to a single state and focus on a range of other policy issues such as school discipline, student test scores, or welfare generosity (but see Grissom et al., 2009; Nicholson-Crotty, Grissom, & Nicholson-Crotty, 2011). This work uses a nationally representative sample of school districts to assess the effect of external control on bureaucratic representation. The broader sample provides a greater level of generalizability and reduces the chances of state or regional bias in the results. Finally, the research focuses on ability grouping, a less researched policy issue in public administration (but see Grissom et al., 2009; Meier & Stewart, 1991; Meier et al., 1989; Nicholson-Crotty et al., 2011). Unlike discipline or student test scores, ability grouping is more often treated as a “zero-sum game,” in which the benefits are fixed and representation intervention for one group may come at a cost to another group. Positions in gifted education, in particular, are often highly coveted and restricted to an explicit number of students per school or district based on resources (Baker & McIntire, 2003). This makes the policy issue one that holds obvious ramifications for minority bureaucrats’ groups; it is highly contentious and important to parents, district administrators, and the external environment, in general.

The Environment as an External Control

Understanding the internal and external factors that contribute to the organizational outcome process is important for predicting when bureaucratic representation might be less effective in providing substantive benefits for clients. Internal processes—the factors that are within the bureaucrat’s or organization’s control—are often easy to manipulate or navigate around to ensure productivity and organizational success. For example, organizations have much more control over employee morale than external political support, and can therefore adjust internal happenings to make this a positive factor for organizational success. However, external processes, factors outside of the organization or bureaucrat’s realm of control, are more difficult to manipulate and address. These factors may consist of clientele preferences or opinions, political elites’ control, or institutional and structural constraints to the organization. Consequently, the actions of organizational bureaucrats, the implementers of policy and handlers of clientele concerns and interests, may also be restricted to this external control. Such restrictions could lead bureaucrats of varying environments under different forms or types of external control to represent differently.

The open systems theory research of organizational theory and behavior is used to guide the argument and expectations for why and how minority bureaucrats in one organizational setting may represent differently from their colleagues in another. Arguments of organizations under the open systems theory note that organizations have interdependences and interactions with their environment that guide their actions, outcomes, and outputs (Katz & Kahn, 1966). Changes and cues from the broader environment or “system” also influence changes in organizations (Shafritz, Ott, & Jang, 2005, p. 476). In fact, Pfeffer and Salancik (1978) contend that the internal operation and happenings of the organization and its bureaucrats reflect the external environment on which it is dependent. In other words, organizational functioning and outputs are largely a function or consequence of the environment and its contingencies and constraints because the organization is dependent on it for resources and support. Unless organizations are able to dominate their resource environment, which is typically not the case for public bureaucracies, they have limited room to establish themselves separate from the environment. This suggests that bureaucrats may lack independent discretion to act on their identities and, instead, may be socialized to actively represent clients based on the support, norms, and cues of the external environment. Simply put, minority bureaucrats will only represent minority clients to the extent and in the manner in which the external environment permits and supports the representative behavior.

Hypothesis 1: External environmental control moderates the passive to active representation link.

If the external environments of two organizations differ, bureaucrats in one environment may represent clients in one way, while those in an alternate environment may represent in another manner or even fail to represent co-ethnic clients. Because this research deals with inequalities that are likely sensitive to minorities, one may not expect the minority representatives to always represent in opposite ways, but Pfeffer and Salancik’s (1978) approach gives reason to expect some divergent behavior in active representation. We could see diverging forms of active representation across different external environments, or possibly very little evidence of active representation in one setting, and much more active representation in another. These findings would suggest that representation is conditional as Hindera and Young (1998) contend, but based on the conditions of the external environment, as Pfeffer and Salancik (1978) note. In sum, passive representation may translate into active representation when the external environment supports active representation; the context of representation moderates the relationship between the two concepts.

Bureaucratic Response to External Control: A Closer Look at Discretion

Pfeffer and Salancik’s (1978) resource dependency theory offers one perspective on why passive representation may translate into active representation in some instances, but not in others. The scholarship on street-level bureaucrats’ use of discretion and their ability to use such discretion to influence, manage, and even control their internal and external environment offers a different perspective and expectation of the passive to active representation link. As previously noted, some scholars of representative bureaucracy are beginning to explore the manner in which various contexts affect representation; however, researchers of bureaucratic control and policy implementation offer much more insight on why street-level bureaucrats’ representative behavior may be tied less to the organization’s dependence on the external environment and how bureaucrats sometimes use this interdependence to influence public policy. Although they are not independent actors and are also subject to a variety of influences—from superiors and political elites, to clients, other agencies, and interest groups of the external environment—street-level bureaucrats wield a considerable amount of independence in their actions and behaviors (Prottas, 1978). They are public sector implementers of public policy goals and statutes that operate on the frontlines of service delivery with considerable amounts of discretion (Lipsky, 1980). They interact directly with clients in the external environment, as well as superiors of the managerial level. This position and their role in the organization permits them to have perhaps greater liberties of power and influence in the organization than is typically assumed, given their lower rank in the organization’s hierarchy.

First, having access to both the managerial level of the organization and the client level of the external environment allows street-level bureaucrats to control the flow of information in and out of an organization. Generally, administrative superiors and other members of an organization interact with clients through street-level bureaucrats, and clients’ interaction with the government is mainly through the street-level bureaucrats—case workers, teachers, police officers (Prottas, 1978). This information asymmetry makes it difficult for superiors to limit bureaucratic discretion, and it makes it difficult for clients to alter its effect on their service delivery.

Second, agencies are dependent on the environment for clients, but the external environment is perhaps more dependent on street-level bureaucrats to access services. Street-level bureaucrats determine access to benefits and services within the organization and the distribution of the benefits and services to the clients in the external environment. Street-level bureaucrats are given relatively large amounts of autonomy and self-direction because they hold specific expertise in the service delivery of their policy area that others in the organization and policy makers may lack (Prottas, 1978). They also work in a complex task environment of uncertainty and unpredictability, which not only expands their need to exercise discretion but also restricts the ability of other administrators and supervisors to monitor or standardize their activities (Meyers & Nielsen, 2012).

Finally, they often act as enforcers or regulators of public polices and laws; street-level bureaucrats are charged with upholding statutes and ensuring that recipients of goods and services do as well (Meyers & Nielsen, 2012). As such, their level of control over the implementation of organizational policies and goals and their power over their service domain frequently limit both organizational and external control over street-level bureaucrats. While an organization may certainly operate in an open system and depend on the external environment for resources and the external environment may impose some norms, values, and expectations on the organization that become a part of its ethos or culture, street-level bureaucrats are engaged in boundary spanning activities that allow them to maintain significant levels of discretion to operate semi-independent of this control or even influence the manner in which the control “affects” them. Street-level bureaucrats exert influence and power beyond that which is formally authorized (Meyers & Nielsen, 2012). Lipsky (1980) contends that this discretion allows them to be “agents of social control,” significantly affecting their external environment and the behaviors of clients with whom they interact.

Empirical research provides evidence of this behavior among street-level bureaucrats (Keiser, Mueser, & Choi, 2004; Keiser & Soss, 1998; Marvel & Resh, 2015). Similarly, minority bureaucrats may use this coveted discretion to represent minority interests or obtain resources and benefits for their demographic group, especially when the norms and policies of the organization are vague or decentralized (Meier & Bohte, 2001; Watkins-Haynes, 2009). Here, external influences may not moderate the manner in which passive representation translates into active representation. Instead, we should expect passive representation to translate into active representation for minority clients similarly across various organizational contexts. In other words, the external environment will be less significant to minority street-level bureaucrats’ representative behavior as they operate semi-independent from the pressures and influence of the environment, or may even use the external environment to expand discretion and active representation. Empirical findings should also reflect the previous research on representation’s positive effect on minority outcomes.

Hypothesis 2: External environmental control does not affect the passive to active representation link.

Integrating these arguments into the representative bureaucracy literature produces an alternative and perhaps more accurate understanding of how bureaucrats come to represent clients, when and where they are more or less likely to represent their racial group’s interests, and why variation in the evidence of active representation occurs.

The Case of School Desegregation

The case of school desegregation policy and the racial composition of schools provide an appropriate lens for explaining and understanding the role of external environmental control in predicting organizational and bureaucratic behavior for several reasons. First, school districts have been a fertile ground for seeking evidence of the relationship between passive and active representation for scholars; consequently, there is a plethora of research on the relationship between teacher race or gender representation and policy outcomes as an indicator of active representation (Keiser et al., 2002; Meier & Stewart, 1991; Meier, Wrinkle, & Polinard, 1999; Nicholson-Crotty et al., 2011; Pitts, 2007; Roch & Pitts, 2012). Second, education scholars have developed a rich body of research on the differences between racially balanced and imbalanced school districts in terms of academic climate, rigor, resources, and, consequently, the effect such differences have on student achievement and life outcomes. These works also offer some insight on the moderating effect that the differing levels of racial balance may have on minority teachers’ representative behavior.

Racially balanced school districts tend to have a more rigorous academic climate and greater financial and human capital (Southworth & Mickelson, 2007). Teachers of racially balanced schools are generally more experienced, considered of higher quality, and less likely to lack certification (Clotfelter, Ladd, & Vigdor, 2005; Southworth & Mickelson, 2007). As such, students of the more desegregated, racially balanced environment are generally more successful academically in the short term and economically in the long term. They tend to have greater academic attainment levels—they are less likely to drop out of school and more likely to attend and complete college, hold occupational positions of higher prestige, and have white-collar and professional jobs (Crain & Strauss, 1985; Goldsmith, 2009; Guryan, 2004; Wells & Crain, 1994). More segregated, racially imbalanced schools, however, are more likely to have a larger percentage of novice or unqualified teachers and they experience more teacher turnover. The schools have higher levels of poor, homeless, or non-English speaking students and fewer academically advanced students; they offer fewer college preparatory courses and tend to have poorer facilities (Clotfelter et al., 2005; Frankenberg, 2009; Goldsmith, 2011; Loeb, Darling-Hammond, & Luczak, 2005; Southworth & Mickelson, 2007). These schools also tend to be located in more urban areas and many of their students perform below state and national standards on standard achievement assessments (Freeman, Brookhart, & Loadman, 1999). According to Sanders (1984), Black students of racially isolated teachers “gained 27 months less achievement” than their peers of non-isolated teachers. Students in racially imbalanced schools tend to achieve less; they attain less education and hold lower prestige occupations compared with their counterparts in White concentrated or racially balanced schools (Dawkins & Braddock, 1994; Goldsmith, 2011; Wells & Crain, 1994).

Teachers and administrators of racially imbalanced schools often report different experiences from their counterparts in more racially balanced schools; such differences also signal the potentially different moderating effects that the racial context of a district may have on minority teachers’ active representation. For example, Milligan and Howley (2015) find that administrators of segregated schools are more likely to manage teachers whose experiences and values are not as aligned with their students, whereas teachers of more diverse, racially balanced schools tend to share values and experiences with their generally middle-class and racially parallel students. They contend that teachers are much more likely to be in open cultural class and value conflict with their student population in more racially imbalanced schools (Milligan & Howley, 2015). Teachers often perceive the socioeconomic and culture differences that they have with their students as barriers to meaningful relationships with their students and contributors to the task complexity of their position (Freeman et al., 1999). Even minority teachers may find themselves at odds and less inclined to “represent” co-ethnic students based on the value incongruence and perceptions of cultural incongruence with students in less racially balanced schools. Teachers of less racially balanced schools also report having less discretion in decision making and greater uncertainty in their task environment when compared with their peers (Freeman et al., 1999). Overall, the research provides evidence of the distinct external environments the level of racial balance in a district creates, and based on the divergent moderating roles each environment has on students and teachers, one might also expect it to moderate the representative behavior of street-level bureaucrats.

Measures and Methods

The study’s data are collected over two school terms, 2000-2001 and 2003-2004, from a series of sources. I use the National Education Survey of U.S. Public Schools, an original survey of the 1,800 largest school districts in the United States, to gather data on the racial composition of the school faculty; the National Center for Education Statistics’ (NCES) School District Demographic System and Common Core Database for data on schools’ student racial compositions and community resource variables; the Office for Civil Rights (OCR) for school policy outcome variables; and the 2000 Census to create a measure of residential segregation. The units of analysis are public school districts with a 1999-2000 school enrollment above 5,000. The survey had a response rate of 95%, yielding 3,567 cases. About 15% of the cases are missing key variables such as a student’s odds of being grouped into gifted education and the percentage of Latino and African American students in gifted education. Difference of means tests that compared non-missing cases with missing cases revealed that the cases included in the models on Latino education are more likely to be larger school districts in the south that have larger Latino student populations, hold a Latino population above the U.S. average (17.6%), have higher percentages of gifted Latino students, are less racially balanced, but have a greater proportion of educated and high-income Latinos to Whites when compared with the missing cases. Similarly, the school districts missing from the model that tests the odds of Blacks’ assignment to gifted education are more likely to have higher percentages of Black teachers and administrators; they are more racially balanced, have lower levels of residential segregation, have lower levels of Whites in poverty, and have larger Black populations based on Census data. These districts also have smaller student populations and larger percentages of Black students in gifted education, and Black students have a higher odds of being grouped into gifted education in districts not included in the model. Given the significant differences between the two groups of school districts, I used the mean imputation method² to reduce the number of missing cases on the dependent variables and key explanatory variables—the proportional odds of gifted assignment for Latino and African American students and the percentage of Latino and African American teachers and administrators—as well as the total student enrollment figures.

Dependent Variables

A student’s odds of being placed in gifted education, an academic grouping measure, is used to identify within-school segregation procedures and to examine any differences in how minority teacher representation affects the processes across the two external environments (U.S. Department of Education, OCR, 2010³). A proportional index measure (odds index) is constructed to capture the extent to which African American or Latino students are disproportionately assigned to gifted education classes. It assumes that students of each racial group represented in a school or district are assigned to such courses at a rate equal to their population size (Meier, Stewart, & England, 1989). It can be best illustrated as

P = \frac{\frac{h_{g}}{H_{t}}}{\frac{G_{t}}{N}}

where $h_{g}$ equals the number of Latino students grouped in gifted education courses, $H_{t}$ is the total Latino student population, $G_{t}$ represents the total number of students grouped in gifted education, and $N$ is the total student population. Traditional of any ratio calculation, an index value of 1.0 indicates that students are proportionally represented in gifted courses relative to their share of the total student population; values above 1.0 indicate overrepresentation in the courses and values below 1.0 indicate underrepresentation (Meier & Stewart, 1991; Meier et al., 1989). An examination of the variables’ distribution suggested skewness due to extreme values, so a log transformation of the measure is used in each model. Log transformation makes a skewed distribution more normal and reduces the weight placed on extreme-value cases in a regression. The interpretation of the results change; a change in the regression slope is now a percentage change rather than a one unit change in the dependent variable.

This measure allows me to observe the overrepresentation or underrepresentation of students in academic groups to determine the extent to which minority representatives and/or the external environment alter the groupings. However, the measure fails to capture some important aspects of educational inequality that are fundamental to the research. First, the discriminatory process of academic grouping, the inequitable class assignment of students based on measured or perceived school ability, and performance for particular subjects often follows discriminatory tracking policies—the disproportionate assignment of minority students in college-bound or vocational tracks (Oakes, 1985; Gamoran, 1992). The OCR data do not allow me to capture the tracking aspect of desegregation and the effect that representation and the external environment may have on this process, though previous scholarship suggests that there is reason to believe that students in a racially balanced environment are more likely to be disproportionately funneled into the lower, vocational track and less likely to be fostered into the college-bound track (Oakes, 1985).

Second, I am not able to disentangle the extent to which students are discouraged to pursue certain academic tracks or more positive opportunities. It is unclear how much of the grouping is related to self-selection and the amount attributed to teacher or administrator behavior. Although the theory contends that teachers of certain backgrounds affect the process through their recommendations, a full picture of physical recommendation (submitting formal paperwork to group a student) and verbal recommendation to individual students would be ideal to capture academic grouping assignments. Nevertheless, the measure is the best measure for capturing within-school segregation and active representation.

Explanatory Variables

Minority teacher representation, the main independent variable, is measured at the district level as the percentage of African American and Latino teachers per district. A second key explanatory variable in this study is the level of school district racial balance. Racial balance is assessed as a Taeuber dissimilarity index for each district, indicating the overall evenness of schools within a district (Taeuber, 1964). The measure itself captures the percentage of students that would need to transfer to different schools to make all of the schools within the district equally mixed among the races and ethnicities. I convert the measure into a similarity score to simplify interpretation as follows:

1 - (\frac{1}{2} \sum_{i = 1}^{n} | \frac{b}{B} - \frac{w}{W} |)

where b equals the Black students in an individual school, B equals the number of Black students in the entire district, w equals the White students in an individual school, and W equals the White students in the entire district (Rodgers & Bullock, 1976).

A similarity index score of zero (total racial isolation) shows that a district is completely segregated and nearly all of the students would need to change schools to equally distribute the races across the district, while a score of 1 (complete racial balance) indicates complete desegregation. Reverse coded from the literature to simplify interpretation, a similarity index score of .30 or below signals high levels of segregation, scores between .40 and .50 are considered moderate levels of segregation, and values of .60 and above typically signal maximum desegregation or a close approximation (Rodgers & Bullock, 1976). Because a score of .60 or above is generally accepted as an appropriate level of desegregation, the .60 cutoff is used in the analyses below to separate more racially balanced districts from less racially balanced districts. The average racial balance score is .73 for both African American and Latino students, suggesting that, in general, minority students in the sample attend schools that are significantly racially balance. Only about 17% of the districts in the sample are characterized as racially imbalanced.

Control Variables

A set of resource variables expected to improve the equity of grouping policies for minority students are also included in the analyses. Greater resources—financial or political—increase minorities’ ability to pressure school districts to create a more equitable academic setting for students (Meier & Stewart, 1991; Meier et al., 1989). Resource controls include the ratio of Blacks and Latinos with college degrees to White degree holders, the percentage of Black and Latino homeowners, the median Black and Latino family income to the White median income, and the percentage of Whites in poverty for each school district.

Outside of the resource variables, the study accounts for district characteristics that may affect grouping procedures. A measure of district enrollment is used to control for district size. Larger districts are generally more professionalized, are more aware of civil rights laws and regulations, are typically under more public scrutiny when it comes to questions of equity, and are, therefore, also more likely to be sensitive to even covert discriminatory practices like inequitable grouping and tracking policies (Meier & Stewart, 1991; Meier et al., 1989). I expect more equitable polices in larger districts. Upper level administrator representation is also included in the models; the percentage of Latino and African American administrators in a district is used as a control of alternate bureaucratic representation. The measure also serves as a measure of internal control over street-level bureaucratic discretion. Higher levels of administrative representation may reduce the amount of street-level active representation provided or increase the odds of more “professionalized” versus discretionary behaviors.

Dummy variables are included to control for geographic region—southern states and Latino-concentrated states. Southern states (1 = southern state, 0 otherwise⁴) are expected to have greater levels of within-school segregation in the form of grouping measures given the nature in which desegregation was adapted in the region. The Latino-concentrated state variable (1 = Latino population at or above the national average, 0 otherwise⁵) is based on the overall Latino population in the state. These states also have a history of Latino segregation and are more likely to use within-school segregation given the size of the Latino population (Bowman, 2001; Clotfelter, 2004; Meier & Stewart, 1991).

A control for the level of residential segregation in the school districts is also included. Orfield and Eaton (1996) credit the changes in the racial composition of schools to regional residential trends. Lower levels of residential segregation increase the probability that students attend more racially balanced schools that are also more likely to segregate the students internally. I include an interaction index, an exposure measure, scaled from zero to one that measures the probability of Blacks or Latinos interacting with Whites in their school district, given the census block restrictions. Index values are calculated as follows:

P_{k k} = \sum_{i = 1}^{I} (\frac{n_{i j}}{N_{j}}) (\frac{n_{i k}}{n_{i}})

where $n_{i j}$ is the population of group $j$ in census block $i,$ $N_{j}$ is the total population of group j in the district, $n_{i k}$ is the total population of group $k$ in census block $i,$ and $n_{i}$ is the total population in census block $i .$ Greater index values indicate higher levels of exposure (less segregation) and values closer to 0 represents a smaller probability of interacting with Whites (White & Kim, 2005).

Ordinary least squares regressions are used to test the hypotheses. Diagnostic tests revealed threats of heteroscedasticity; robust standard errors were used to address the concern. The diagnostic tests did not reveal any major concerns of omitted variable biases or specification issues.

Findings

Descriptive Findings

Tables 1 and 2 show the odds of Black, Latino, and White students being grouped in gifted education, first in an overall comparison, and then a comparison based on the external environment. In Table 1, Black and Latino students are underrepresented in gifted courses. An odds ratio of .48 indicates that Black students are underrepresented by 52 percentage points in gifted classes, and Latino students are underrepresented by 50 percentage points. White students, however, are overrepresented in gifted courses; they are nearly 3 times more likely than Black students to be enrolled in gifted classes.

Table 1.

Grouping Policy Ratios by Race.

Grouping policy	Blacks		Latinos		Whites
Grouping policy	M	SD	M	SD	M	SD
Gifted education	0.4837	0.4241	0.5008	0.4154	1.417	0.6796

Table 2.

Grouping Policy Ratios by Race and Level of Racial Balance.

Grouping Policy	Blacks		Latinos		Whites
	Racially balanced	Racially imbalanced	Racially balanced	Racially imbalanced	Racially balanced	Racially imbalanced
	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)
Gifted education	0.4657*** (0.3995)	0.5645 (0.5131)	0.4960 (0.3959)	0.5215 (0.4910)	1.352*** (0.4862)	1.705 (1.171)

Note. Ratios between racial group in racially balanced and imbalanced districts statistically different at specified level.

p < .1. **p < .05. ***p < .01.

A deeper look at this data by external environment in Table 2 shows that Black and Latino students are still underrepresented in gifted classes in both racially imbalanced and balanced districts, but to a greater extent in racially balanced districts. Black students are underrepresented by 43 percentage points in racially imbalanced districts, but they are underrepresented at a rate of 10 additional percentage points in racially balanced districts. The difference between the two external environments is statistically significant, and the finding is consistent with previous literature (McBee, 2010; Meier et al., 1989). White students remain overrepresented in gifted classes regardless of the academic setting, though their odds of being placed in gifted education are significantly higher in more racially imbalanced school districts. In racially imbalanced districts, White students are overrepresented in gifted classes by 70 percentage points, but they experience a significant decline in overrepresentation in racially balanced districts at 35 percentage points. There is no statistical difference observed for Latino students.

In general, the preliminary descriptive findings seem to highlight the potential environmental distinction between racially balanced and imbalanced school districts, which could lead to a differential effect on minority teacher representation, as Pfeffer and Salancik suggest. The level of disparity may also cue the salience of discriminatory grouping policies for minority bureaucrats, prompting them to engage in greater levels of active representation to reduce the discrimination (Meier, 1993).

External Environment Influence on Representation in Public Schools

Tables 3 and 4 include a comparison of the results for the effect of Black and Latino teacher representation on grouping policies in racially balanced and imbalanced school districts for Black and Latino students. The models of Table 3 offer evidence of active representation and then they compare the effect of Black teacher representation in racially imbalanced school districts with the effect of Black teacher representation in racially balanced school districts on academic grouping for Black students. First, Model 1 shows that when the external environment is not considered, teachers, street-level bureaucrats, can use their discretion to provide active representation for Black students, though this effect is very limited substantively. A 1% increase in African American teachers is associated with a 0.25% increase in an African American students’ odds of being grouped into gifted education. Substantively, this means that in the average school district, an African American student is underrepresented by 51.75 percentage points when passive representation is present, versus 52 percentage points when it is not present. A 50% increase in the number of African American teachers on staff only ensures a 12.5% increase in the Black gifted ratio, so African American students remain underrepresented by about 40 percentage points.

Table 3.

Effect of Bureaucratic Representation on Gifted Education Grouping Outcomes for Black Students Across External Policy Environments.

Variables	Coefficient (SE)	Coefficient (SE)	Coefficient (SE)
Black representation (%)	0.002** (0.001)	0.001 (0.001)	0.006** (0.003)
Black–White Racial Balance Index		−0.127* (0.073)	0.049 (0.067)
Representation × Racial Balance			−0.007* (0.004)
Black administrator representation (%)		0.001 (0.001)	0.000 (0.001)
Black/White college graduate ratio	0.028 (0.025)	0.028 (0.025)	0.017 (0.024)
Black/White income ratio	0.003 (0.022)	0.004 (0.022)	0.006 (0.018)
Homeownership (%)	0.003*** (0.000)	0.003*** (0.000)	0.001** (0.000)
Whites below poverty (%)	0.019*** (0.003)	0.018*** (0.003)	0.009*** (0.002)
Level of residential segregation	−0.050 (0.053)	−0.027 (0.055)	0.013 (0.042)
South	−0.180*** (0.021)	−0.183*** (0.021)	−0.057*** (0.016)
Latino gifted (%)	0.013*** (0.001)	0.013*** (0.001)	0.007*** (0.001)
District size (100,000s)	0.025* (0.015)	0.013 (0.015)	−0.007 (0.010)
Constant	−1.168*** (0.042)	−1.081*** (0.073)	0.256*** (0.057)
Observations	3,129	3,129	3,297
R ²	.207	.207	.123
Adjusted R²	.204	.205	.120

Note. Robust standard errors in parentheses.

p < .10. **p < .05. ***p < .01.

Table 4.

Effect of Bureaucratic Representation on Gifted Education Grouping Outcomes for Latino Students Across External Policy Environments.

Variables	Coefficient (SE)	Coefficient (SE)	Coefficient (SE)
Variables	(1)	(2)	(3)
Latino representation	0.010*** (0.001)	0.008*** (0.002)	0.014*** (0.005)
Latino–White Racial Balance Index		0.336*** (0.087)	0.414*** (0.105)
Representation × Racial Balance			−0.010 (0.006)
Latino administrator representation (%)	0.022 (0.025)	0.002 (0.001)	0.002 (0.001)
Latino/White college graduate ratio	0.154*** (0.059)	0.027 (0.025)	0.028 (0.025)
Latino/White income ratio	0.003*** (0.001)	0.151** (0.060)	0.151** (0.060)
Homeownership (%)	0.025*** (0.003)	0.003*** (0.001)	0.003*** (0.001)
Whites below poverty (%)	0.167** (0.066)	0.027*** (0.003)	0.027*** (0.003)
Level of residential segregation	0.000*** (0.000)	0.105 (0.069)	0.097 (0.069)
District size (100,000s)	0.164*** (0.021)	0.074*** (0.019)	0.069*** (0.019)
Latino region	0.014*** (0.001)	0.149*** (0.021)	0.148*** (0.021)
Black gifted (%)	−1.569*** (0.060)	0.014*** (0.001)	0.014*** (0.001)
Constant	0.010*** (0.001)	−1.778*** (0.077)	−1.829*** (0.086)
Observations	3,253	3,253	3,253
R ²	.162	.166	.166
Adjusted R²	.159	.163	.164

Note. Robust standard errors in parentheses.

p < .10. **p < .05. ***p < .01.

The second model of Table 3 suggests that when the external environment’s control is considered, even these small benefits of passive representation are lost and it does not lead to statistically significant active representation. Here, the external environment influences a Black student’s odds of being equitably grouped into gifted education. A 1% increase in the level of racial balance, that is, moving closer to a more demographically equal district, is associated with a 12.7% decrease in the Black gifted class ratio (p < .10). Black students’ underrepresentation in gifted education increases from 52 percentage points to about 65 percentage points in the average school district.

The final model of Table 3 tests the influence of the external environment on the passive to active representation linkage. Model 3 suggests that passive representation leads to active representation, but when the external environment’s influence—balanced versus imbalanced—is considered alongside passive representation, African American students’ odds of being grouped into gifted education decreases. A 1% increase in Black teacher representation is related to only a 0.60% increase in the Black gifted class ratio, so Black students remain underrepresented in gifted education course by 51.4 percentage points. When Black teacher representation increases by 50%, Black students are only underrepresented by 23.9 percentage points.

However, when the external environment increases its racial balance by 1%, a 1% increase in passive representation leads to a 0.70% decline in active representation. Substantively, this means that Black students’ underrepresentation almost remains unchanged in the average school district (52 percentage points vs. 52.70). Figure 1 allows for a more complete view of where and how the external environment affects Black teachers’ active representation. Here, the external environment reduces active representation to the greatest extent in racially balanced districts, leading Black representation in the imbalanced districts to differ from representation in the most balanced districts. Among the entire range of racially imbalanced districts, that is, districts with a racial balance score below the .60 threshold, Black teachers have small positive, but diminishing effects on Black students’ odds of being assigned to gifted education. In the most racially balanced districts, which are also the districts in which minority students experience greater levels of inequality, the effect is lost; Black teachers fail to have a statistically significant effect on gifted education groupings for Black students.

Figure 1.

Marginal effect of Black teacher representation on Black students’ odds of gifted education assignment as policy environments change.

Table 4 includes the results for Latino street-level bureaucratic representation’s effect on grouping practices. Latino students have a greater odds of being grouped into gifted education when passive representation increases and this finding is consistent across all three models of Table 4, despite external control or influence. In Model 1, a 1% increase in Latino teacher representation increases the Latino gifted class ratio by 1%. A 50% increase in the number of Latino teachers increases a student’s chances by 50%. A 50% increase in the Latino gifted class ratio means that Latino students would be grouped at ratio proportional to their population size if Latino teachers’ share of the staff increases by 50% in the average school district.

In Model 2, Latino students experience the greatest gains from the external environment’s pressure. Latino student’s odds of being grouped into gifted education increases by 33.6% when a school district increases its level of racial balance by 1% (see Model 2). Such an increase means that Latino students are only underrepresented in gifted education by 17 percentage points, compared with the 49 percentage point disparity of Model 1. Latino passive representation is also associated with increases in active representation in Model 2, but the substantive effect of the relationship is much smaller than that of Model 1 and the external environment effect in Model 2. Here, a 1% increase in Latino representation is associated with a 0.80% increase in the Latino gifted class ratio.

Finally, Model 3 shows that Latino students experience their greatest gains from Latino representation in Model 3; a 1% increase in Latino teacher representation is associated with a 1.4% increase in Latino gifted class ratio. A 1% increase in the racial balance of a school district increases the ratio by 41%, moving the Latino gifted education ratio within 9 percentage points of parity in the average school district. However, when passive representation interacts with the external environment, it does not affect Latino students’ odds of being equitably grouped into gifted education.

Figure 2 allows for a more complete view of the interactive effect of teacher representation and the external environment on the Latino gifted class ratio. Despite the insignificant findings of Model 3, Figure 2 suggests that Latino teachers have a positive, but diminishing effect on Latino students’ assignment to gifted education, and this significant finding holds until the highest points of the racial balance scale. For example, a 1% increase in Latino teacher representation in the average racially imbalanced district is associated with a 0.90% increase in the Latino gifted class ratio, but when the district shifts to mirror the average racially balanced district, the ratio only increases by about 0.70%. In the perfectly balanced district, Latino teacher representation is insignificant.

Figure 2.

Marginal effect of Latino teacher representation on Latino students’ odds of gifted education assignment as policy environments change.

Overall, minority street-level bureaucrats are able to use their discretion to provide active representation for co-ethnic students, though on a much smaller scale than anticipated; however, the external environment does moderate the extent to which active representation occurs. Both African American and Latino bureaucrats appear to experience the greatest effects on their discretion and active representation in the most racially balanced districts. The differing outcomes of active representation across the two external environments and the decline in substantive active representation for minority students when the external environment is considered offers some evidence of this moderating effect; external environmental control influences the passive to active representation link.

Discussion and Conclusion

While the previous scholarship examining the relationship between passive and active representation is mixed, the current work explores the extent to which external environmental control and pressure explains the variation in findings. Furthermore, the research uses organizational theory and theories of bureaucratic control to guide expectations for the manner in which the environment may influence passive and active representation. Pfeffer and Salancik’s resource dependence theory offers one lens to examine the variation in passive to active representation linkages, while Lipsky and others’ work on bureaucratic discretion and control offer another. Pfeffer and Salancik’s work suggests that minority street-level bureaucrats may vary in their active representation because they depend on different external environments for resources, information, and cues of acceptable norms or practices. On the contrary, the research on bureaucratic discretion suggests that the external environment is less likely to influence or “control” bureaucrats because they hold significant amounts of discretion to implement policies using their own values, which may include values of actively representing their demographic group’s interests. Assuming most bureaucrats operate in this manner, we might not expect the external environment to influence the passive to active representation link.

Empirical results suggest that street-level bureaucrats do hold some level of discretion based on their active representation, but as Pfeffer and Salancik’s work suggests, this discretion can be externally influenced or moderated. The external environment seems to affect the relationship between passive and active representation; both African American and Latino teachers’ effect on grouping policies are minimized when the external environment is considered, and this leads to differing levels of active representation based on the level of racial balance in a district. Minority students experience relatively small gains via bureaucratic representation, but a shift in the level of racial balance negatively affects the gains of bureaucratic representation. The findings support the first hypothesis.

In general, these findings are also consistent with the findings of Roch and Pitts (2012). Although minority teachers may press to provide active representation in both environments, ultimately, the different settings lead to different levels of active representation and policy outcomes. This finding is possibly due to Pfeffer and Salncik’s (1978) point that organizations and their bureaucrats depend on the external environment for resources so the internal processes and decisions are also a reflection of the interests and demands of the environment. As Bryk and Driscoll (1988) point out, teachers often take cues from the community in their interactions and behaviors with students. School districts are communal, interdependent organizations that foster cooperative relationships in an effort to reach a common agenda of goals; teachers’ responsiveness to the community is a reflection of this communal ethos, so the extent to which their passive representation can actually yield benefits for clients is community-driven (Bryk & Driscoll, 1988). Greater research is required to explain the variation in the passive to active representation and conditions in which the external environment matters for representation and when it does not. However, this research contributes to the continual discussions on the factors that explain the success and process of representative bureaucracy. The transition from passive representation to active representation includes more than the actions of the actual bureaucratic agency and the bureaucratic agents. Environmental contexts and external mechanisms of control or influence also play a role in making passive representation a substantive benefit for clients. At the baseline, the results suggest that there is potential for the external environment to affect, and sometimes limit, representation, even in agencies that grant bureaucrats a significant level of autonomy and discretion and within the same policy field.

The research findings also offer some interesting implications for the research on bureaucratic discretion. While it is understood that street-level bureaucrats wield some level of discretion in policy implementation and outcomes, the results suggest that they may also use this discretion to reduce the influence that negative environmental factors may have on the organization or clients. The models on African American students reveal very limited evidence of substantive active representation independent of the external environment; Black students do not experience many substantive improvements to their odds of being assigned to gifted education as passive representation increased. However, when this representation was coupled with the environment, their substantive effect in reducing the odds of an even greater decline in equitable grouping emerged. Black students experience a 12.7% gifted class ratio decline when the level of racial balance increases, but this decline is only 0.70% when racial balance is coupled with Black teacher representation. Despite the limited substantive effects, Black teachers appear to use their discretion to actively stop more inequality than build toward more equality.

The findings add to the public administration research on bureaucratic control as well. Scholars acknowledge a few contexts in which bureaucrats are more likely to experience limits to their discretion such as political signals or economic pressures (Keiser, 1999; Keiser & Soss, 1998), managerial supervision and goals (Riccucci, 2005), or formal, structural rules of implementation (Buffat, 2015). However, the current findings suggest that perhaps the collective effect of the external environment can also work to restrict bureaucratic discretion. The current study lacks the capacity to test other environmental factors such as interactions with political elites or the community or the networked relationships that such actors may use to influence bureaucratic discretion; however, future research should consider the importance of macro-level controls on bureaucrats and the implications such factors have on the implementation and outcomes of public policy. Expanding our understanding of the role of bureaucratic discretion in public policy is important not only for studies of representation but also for the broader process of maintaining accountability in increasingly complex, interdependent organizations.

Finally, the null findings of upper level administrators also offer an interesting contrast to the findings for street-level bureaucrats. The models suggests that upper level administrators do not affect policy outcomes nor do they alter the active representation of street-level bureaucrats. These findings continue to support the contention among representative bureaucracy and bureaucratic scholars that the level of representation matters for examining passive to active representation linkages (Roch & Edwards, 2017; Roch & Pitts, 2012). They also support the contention that superiors often struggle to control street-level bureaucrats because they are semi-independent, boundary spanning, policy makers with unique policy implementation expertise and privileged access to information about both the agency and the client (Lipsky, 1980; Meyers & Nielsen, 2012; Prottas, 1978). Future work may consider the collaborative, networked relationships that upper level administrators make with the community or external environment to influence bureaucratic behavior.

The results, however, must be considered within the limitations of this research. First, a major assumption of the current study is that minority teachers have adequate levels of discretion in making ability grouping recommendations on which the environment can then act to influence bureaucratic behavior. However, the level of bureaucratic discretion teachers have in making ability grouping recommendations has eroded over time. Teachers are still the first source of information in students’ academic ability; however, some school districts opt to test entire student populations for special groupings. This practice reduces the level of influence that teachers have to offer in the grouping process. The lower predictive power of the models may be capturing this caveat.

The unit of analysis is also a limitation of the study. As Bradbury and Kellough (2011) point out, it is important to, when possible, “isolate the impact of active representation . . . in which individual bureaucrats are the units of analysis,” but school districts are the unit of analysis here (p. 161). Individual-level analyses would help me more closely pinpoint the true causal mechanisms of representative bureaucracy instead of observing the correlates of representation. I am unable to observe any individual behavior of minority teachers or students that may influence grouping outcomes. For example, teachers are likely to nominate for gifted education students who conform to classroom expectations and have fewer behavioral issues (Davis & Rimm, 2004). I am also unable to account for the potential effect that factors unique to the teachers, besides race and ethnicity, have on students’ grouping outcomes. For example, even teachers of gifted students frequently fail to recognize the broad spectrum of gifted characteristics that signal a students’ readiness for gifted education (Neumeister, Adams, Pierce, Cassady, & Dixon, 2007), so poor professional development or training in gifted education may affect grouping practices. On the contrary, teachers with more experience and knowledge in their substantive field and those with advanced degrees are more effective teachers to gifted students, and are potentially more adept in recommending an equitable mix of students for gifted or special education (Mills, 2003). Finally, the level of analysis prevents the research from considering the student-level factors that may also influence grouping decisions. For example, the models do not account for student achievement, parental support, and gender, which are noted student-level predictors of groupings (Hibel, Farkas, & Morgan, 2010). As previously noted, the models also fail to account for self-selection, that is, the research does not account for a student’s decision to pursue or accept a recommendation for gifted education. In this regard, omitted variable bias may in fact plague the study and affect the overall predictive power of the models. Future research should consider a deeper probe into the additional bureaucratic characteristics and factors that may influence their representative behavior.

Despite these issues, the research suggests that the external environment can influence the transition from passive representation to active representation. While active representation continues to matter in policy implementation, its effect is significantly diminished based on the external environmental context in which it operates. The empirical findings discussed hold implications for the governance of public organizations and the outcomes of public policies. Future work should seek to address some of the limitations of the current work and work to further explore the conditions in which the external environment is most salient to representative behaviors.

Footnotes

Appendix

Table A3.

Effect of Bureaucratic Representation on Academic Grouping Policy Outcomes for Latino Students Across Policy Environments (Gifted Education).

Variables	Coefficient (SE) (1)	Coefficient (SE) (2)	Coefficient (SE) (3)
Latino teacher representation^a (%)	0.007*** (0.001)	0.005*** (0.001)	0.011*** (0.003)
Latino–White Racial Balance Index		0.290*** (0.090)	0.357*** (0.107)
Representation × Racial Balance			−0.009*** (0.003)
Latino administrator representation (%)		0.001 (0.001)	0.002 (0.001)
Latino/White college graduate ratio	0.020 (0.027)	0.026 (0.029)	0.026 (0.029)
Latino/White income ratio	0.052 (0.075)	0.049 (0.078)	0.050 (0.078)
Homeownership (%)	0.001 (0.001)	0.001 (0.001)	0.001 (0.001)
Whites below poverty (%)	0.007** (0.003)	0.009*** (0.003)	0.009*** (0.003)
Level of residential segregation	0.034 (0.064)	−0.033 (0.073)	−0.039 (0.073)
District size (1,000s)	0.040** (0.017)	0.024 (0.018)	0.023 (0.018)
Latino populated region	0.007*** (0.001)	0.007*** (0.001)	0.007*** (0.001)
Black gifted (%)	0.000* (0.000)	0.000*** (0.000)	0.000** (0.000)
Constant	0.229*** (0.061)	0.058 (0.072)	0.011 (0.081)
Observations	2,098	1,983	1,983
R ²	.097	.102	.103
Adjusted R²	.093	.097	.098

Note. Robust standard errors in parentheses.

Dependent variable and main independent variable does not include imputed cases.

p < .10. **p < .05. ***p < .01.

Acknowledgements

The author gratefully acknowledges the helpful feedback of Kristie Seelman, Amanda Rutherford, Julie Dolan, and three helpful anonymous reviewers.

Author’s Note

The previous version of the paper was presented at the Midwest Political Science Association Annual Meeting, 2015, Chicago, IL.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The author gratefully acknowledges the support received for the preparation of this research from the American Educational Research Association, NSF Grant No. DRI-0941014.

Notes

Author Biography

K. Jurée Capers is an assistant professor in the Department of Public Management and Policy at the Andrew Young School of Policy Studies at Georgia State University. Her research focuses on public policy, representation, state and local government, and race and ethnic politics with a particular interest in education policy.

References

Baker

B. D.

McIntire

(2003). Evaluating state funding for gifted education programs. Roeper Review, 25, 173-179.

Bowman

K. L.

(2001). The new face of school desegregation. Duke Law Journal, 50, 1751-1808.

Bradbury

M. D.

Kellough

J. E.

(2008). Representative bureaucracy: Exploring the potential for active representation in local government. Journal of Public Administration Research and Theory, 18, 697-714.

Bradbury

M. D.

Kellough

J. E.

(2011). Representative bureaucracy: Assessing the evidence on active representation. American Review of Public Administration, 41, 157-167.

Bryk

A. S.

Driscoll

M. E.

(1988). The high school as community: Contextual influences and consequences for students and teachers. Madison, WI: National Center on Effective Secondary Schools.

Buffat

(2015). When and why discretion is weak or strong: The case of taxing officers in a public unemployment fund. In Hupe

Hill

Buffat

(Eds.), Understanding street-level bureaucracy (pp. 79-96). Bristol, UK: Policy Press.

Clotfelter

C. T.

(2004). After Brown: The rise and retreat of school desegregation. Princeton, NJ: Princeton University Press.

Clotfelter

C. T.

Ladd

H. F.

Vigdor

(2005). Who teaches whom? Race and the distribution of novice teachers. Economics of Education Review, 24, 377-392.

Crain

R. L.

Strauss

(1985). School desegregation and Black occupational attainments: Results from a long-term experiment. Baltimore, MD: Center for Social Organization of Schools, Johns Hopkins University.

10.

Davis

G. A.

Rimm

S. B.

(2004). Education of the gifted and talented (5th ed.). Needham Heights, MA: Allyn & Bacon.

11.

Dawkins

M. P.

Braddock

J. H.

(1994). The continuing significance of desegregation: School racial composition and African American inclusion in American society. The Journal of Negro Education, 63, 394-405.

12.

Dolan

J. I. I.

Rosenbloom

D. H.

(2003). Representative bureaucracy: Classic readings and continuing controversies. Armonk, NY: M.E. Sharpe.

13.

Esman

M. J.

(1997). Public administration, ethnic conflict, and economic development. Public Administration Review, 57, 527-533.

14.

Frankenberg

(2009). The demographic context of urban schools and districts. Equity & Excellence in Education, 42, 255-271.

15.

Freeman

D. J.

Brookhart

S. M.

Loadman

W. E.

(1999). Realities of teaching in racially/ethnically diverse schools: Feedback from entry-level teachers. Urban Education, 34, 89-114.

16.

Gamoran

(1992). Synthesis of research / Is ability grouping equitable? Educational Leadership, 50, 11-17.

17.

Goldsmith

P. R.

(2009). Schools or neighborhoods or both? Race and ethnic segregation and educational attainment. Social Forces, 87, 1913-1941.

18.

Goldsmith

P. R.

(2011). Coleman revisited: School segregation, peers, and frog ponds. American Educational Research Journal, 48, 508-535.

19.

Grissom

J. A.

Nicholson-Crotty

(2009). Race, region, and representative bureaucracy. Public Administration Review, 69, 911-919.

20.

Guryan

J. S.

(2004). Desegregation and Black dropout rates. The American Economic Review, 94, 919-943.

21.

Hibel

Farkas

Morgan

P. L.

(2010). Who is placed into special education? Sociology of Education, 83, 312-332.

22.

Hindera

J. J.

(1993). Representative bureaucracy: Further evidence of active representation in the EEOC district offices. Journal of Public Administration, Research and Theory, 3, 415-429.

23.

Hindera

J. J.

Young

C. D.

(1998). Representative bureaucracy: The Theoretical implications of statistical interaction. Political Research Quarterly, 51, 655-671.

24.

Hur

(2013). Racial diversity, is it a blessing to an organization? Examining its organizational consequences in municipal police departments. International Review of Administrative Sciences, 79, 149-164.

25.

Katz

Kahn

R. L.

(1966). The social psychology of organizations. New York, NY: John Wiley.

26.

Keiser

L. R.

(1999). State bureaucratic discretion and the administration of social welfare programs: The case of social security disability. Journal of Public Administration, Research and Theory, 9, 87-106.

27.

Keiser

L. R.

Mueser

P. R.

Choi

S. W.

(2004). Race, bureaucratic discretion, and the implementation of welfare reform. American Journal of Political Science, 48, 314-327.

28.

Keiser

L. R.

Soss

(1998). With good cause: Bureaucratic discretion and the politics of child support enforcement. American Journal of Political Science, 42, 1133-1156.

29.

Keiser

L. R.

Wilkins

V. M.

Meier

K. J.

Holland

C. A.

(2002). Lipstick and logarithms: Gender, institutional context, and representative bureaucracy. American Political Science Review, 96, 553-564.

30.

Kingsley

J. D.

(1944). Representative bureaucracy: An interpretation of the British civil service. Yellow Springs, OH: The Antioch Press.

31.

Krislov

(1974). Representative bureaucracy. Englewood Cliffs, NJ: Prentice-Hall.

32.

Lim

(2006). Representative bureaucracy: Rethinking substantive effects and active representation. Public Administration Review, 66, 193-204.

33.

Lipsky

(1980). Street level bureaucracy: Dilemmas of the individual in public service. New York, NY: Russell Sage Foundation.

34.

Loeb

Darling-Hammond

Luczak

(2005). How teaching conditions predict teacher turnover in California schools. Peabody Journal of Education, 80, 44-70.

35.

Marvel

J. D.

Resh

W. G.

(2015). Bureaucratic discretion, client demographics, and representative bureaucracy. American Review of Public Administration, 45, 281-310.

36.

McBee

(2010). Examining the probability of identification for gifted programs for students in Georgia elementary schools: A multilevel path analysis study. Gifted Child Quarterly, 54, 283-297.

37.

Meier

K. J.

(1993). Representative bureaucracy: A theoretical and empirical exposition. In Perry

J. L.

(Ed.), Research in public administration (pp. 1-35). New Greenwich, CT: JAI Press.

38.

Meier

K. J.

Bohte

(2001). Structure and discretion: Missing links in representative bureaucracy. Journal of Public Administration Research and Theory, 11, 455-470.

39.

Meier

K. J.

Stewart

Jr. (1991). The politics of Hispanic education. Albany: State University of New York Press.

40.

Meier

K. J.

Stewart

Jr. England

R. E.

(1989). Race, class, and education: The politics of second-generation discrimination. Madison: University of Wisconsin Press.

41.

Meier

K. J.

Wrinkle

R. D.

Polinard

J. L.

(1999). Representative bureaucracy and distributional equity: Addressing the hard question. Journal of Politics, 61, 1025-1039.

42.

Meyers

M. K.

Nielsen

V. L.

(2012). Street-level bureaucrats and the implementation of public policy. In Peters

B. G.

Pierre

(Eds.), The SAGE handbook of public administration (pp. 264-277). Thousand Oaks, CA: SAGE.

43.

Milligan

Howley

(2015). Educational leadership in our peculiar institutions: Understandings of principals in segregated, White-staffed urban elementary schools in the United States. International Journal of Multicultural Education, 17, 43-61.

44.

Mills

C. J.

(2003). Characteristics of effective teachers of gifted students: Teacher background and personality styles of students. Gifted Child Quarterly, 47, 272-281.

45.

Mosher

F. C.

(1968). Democracy and the public service. New York, NY: Oxford University Press.

46.

Neumeister

K. L. S.

Adams

C. M.

Pierce

R. L.

Cassady

J. C.

Dixon

F. A.

(2007). Fourth-grade teachers’ perceptions of giftedness: Implications for identifying and serving diverse gifted students. Journal for the Education of the Gifted, 30, 479-499.

47.

Nicholson-Crotty

Grissom

J. A.

Nicholson-Crotty

(2011). Bureaucratic representation, distributional equity, and democratic values in the administration of public programs. Journal of Politics, 73, 582-596.

48.

Nielsen

L. B.

Wolf

P. J.

(2001). Representative bureaucracy and harder questions: A Response to Meier, Wrinkle, and Polinard. Journal of Politics, 63, 598-615.

49.

Oakes

(1985). Keeping track: How schools structure inequality. New Haven, CT: Yale University Press.

50.

Orfield

Eaton

S. E.

(1996). Dismantling desegregation: The quiet reversal of Brown v. Board of Education. New York, NY: The New Press.

51.

Pfeffer

Salancik

G. R.

(1978). The external control of organizations: A resource dependence perspective. New York, NY: Harper & Row.

52.

Pitts

(2007). Representative bureaucracy, ethnicity, and public schools: Examining the link between representation and performance. Administration & Society, 39, 497-526.

53.

Prottas

J. M.

(1978). The power of the street-level bureaucrat in public service bureaucracies. Urban Affairs Quarterly, 13, 285-312.

54.

Rehfuss

J. A.

(1986). A representative bureaucracy? Women and minority executives in California career service. Public Administration Review, 46, 454-460.

55.

Riccucci

N. M.

(2005). How management matters: Street-level bureaucrats and welfare reform. Washington, DC: Georgetown University Press.

56.

Roch

C. H.

Edwards

(2017). Representative bureaucracy and school discipline: The influence of schools’ racial contexts. The American Review of Public Administration, 47, 58-78. doi:10.1177/0275074015589126

57.

Roch

C. H.

Pitts

(2012). Differing effects of representative bureaucracy in charter schools and traditional public schools. The American Review of Public Administration, 42, 282-302.

58.

Rodgers

H. R.

Jr. Bullock

C. S.

III . (1976). Coercion to compliance. Lexington, MA: Lexington Books.

59.

Rosenthal

C. S.

Bell

L. C.

(2003). From passive to active representation: The case of women congressional staff. Journal of Public Administration Research and Theory, 13, 65-82.

60.

Sanders

J. M.

(1984). Faculty desegregation and student achievement. American Educational Research Journal, 21, 605-616.

61.

Selden

S. C.

(1997). The promise of representative bureaucracy: Diversity and responsiveness in a government agency. Armonk, NY: M.E. Sharpe.

62.

Shafritz

J. M.

Ott

J. S.

Jang

Y. S.

(2005). Classics of organization theory. Belmont, CA: Thomson Wadsworth.

63.

Southworth

Mickelson

R. A.

(2007). The interactive effects of race, gender and school composition on college track placement. Social Forces, 86, 497-522.

64.

Taeuber

(1964). Negro residential segregation: Trends and measurement. Social Problems, 1, 42-50.

65.

Thompson

F. J.

(1976). Minority groups in public bureaucracies: Are passive and active representation linked? Administration & Society, 8, 201-226.

66.

U.S. Department of Commerce, U.S. Census Bureau. (2010). Census regions and divisons of the United States. Retrieved from https://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf

67.

U.S. Department of Education, Office for Civil Rights. (2010). 2009-10 civil rights data collection: Part 1 and part 2. Retrieved from http://www2.ed.gov/about/offices/list/ocr/data.html

68.

Watkins-Haynes

(2009). Race-ing the bootstrap climb: Black and Latino bureaucrats in post-reform welfare offices. Social Problems, 56, 285-310.

69.

Wells

A. S.

Crain

R. L.

(1994). Perpetuation theory and the long-term effects of school desegregation. Review of Educational Research, 64, 531-555.

70.

White

M. J.

Kim

A. H.

(2005). Residential segregation. In Kempf-Leonard

(Ed.), Encyclopedia of social measurement (pp. 403-409). Miamisburg, OH: Elsevier Science & Technology Books.

71.

Wilkins

V. M.

Keiser

L. R.

(2006). Linking passive and active representation by gender: The case of child support agencies. Journal of Public Administration Research and Theory, 16, 87-102.

72.

Wilkins

V. M.

Williams

B. N.

(2008). Black or blue: Racial profiling and representative bureaucracy. Public Administration Review, 68, 654-664.

73.

Wilkins

V. M.

Williams

B. N.

(2009). Representing blue: Representative bureaucracy and racial profiling in the Latino community. Administration & Society, 40, 775-798.