Since the establishment of the United States as a nation, racial segregation has been an integral part of American culture and society. Racial groups have been separated from each other throughout American history structurally, spatially, and emotionally. Blacks endured de jure segregation from 1787 through 1865. It was reinstituted after Reconstruction in 1877 and remained legally supported until 1954. The Brown versus the Board of Topeka, Kansas Supreme Court decision eliminated segregation by law. As a result of this important court case, racial segregation evolved into one based on housing and spatial patterns. This is often characterized as de facto segregation  .1
Contemporary American society has this pervasive notion that racial equality has been achieved in the United States. Federal government programs aimed at ensuring equal access are under political attack. This has become even more problematic under the Trump presidency. De facto racial segregation patterns in the workplace heavily influenced by educational attainment disparity represent serious barriers to racial equality.
This research effort explores racial segregation in the workplace over an eighteen-year period. The comparison points of reference are years 2000 and 2018. More specifically, the race of the employee was examined as a predictor variable. Gender, age, educational attainment, personal income, work status, occupational prestige, and size of worksite were used as control variables. Conflict theory, in conjunction with race relations theory, framed the dynamics between the predictor and control variables and workplace racial segregation as the response variable.
2. Review of the Literature
Throughout United States history, physical separation of races, popularly known as racial segregation, was the most important form of discrimination experienced by Black Americans. From the establishment of racial slavery in Colonial America through the legal ending of Jim Crow America in 1954, residential segregation based on race bolstered other forms of discrimination in other social institutions such as the economy, religion, education, and government  . Although the workforce in the United States has become more racially diverse since the 1960’s, racial segregation in the workplace continues. Studies have examined racial segregation and they found that residential segregation is linked to educational inequality since public school attendance is predicated on housing patterns. In turn, differential job opportunities and occupational segregation occur, in large part, because of educational inequalities  .
Historically, the Federal government has played a substantial role in minimizing the impact of race discrimination in the workplace. For instance, it desegregated the military in 1948 using a series of executive orders aimed at eliminating discrimination in military units. This was quickly followed in the 1950’s and 1960’s by implementing similar executive actions minimizing employment barriers in Federal, state, and municipal organizations. Racial minorities historically encountered differential treatment resulting in restricted participation and position segregation in these organizations. The Civil Rights Act of 1964 along with the establishment of the Equal Employment Opportunity Commission (EEOC) supported earlier executive orders and focused on eliminating institutional practices that created differential treatment and segregation for a wide range of social groups within the Federal government. As a result, these initiatives provided a template for other organizations (governmental and non-governmental). It should be noted that the aim of the legislation and executive orders was to allow more equitable entrance into the workplace across a number of social dimensions. However, changing the power distribution, resources among groups, and ending workplace racial segregation were not objectives related to equal access opportunities  .
The study of workplace racial segregation begins with looking at the residential or spatial dimension. Rothstein  argues that racial housing segregation was encouraged and supported by the Federal government. It continues to play a strong role in the establishment and maintenance of racial inequality in the United States. Current research shows that racial segregation has increased across urban areas  . Concomitant issues such as violent crime were found to be related to residential segregation. These violent crime rates are centralized in communities with poverty, ethnic isolation and institutional decay. Additionally, the percentage of people living in extremely poor conditions had risen from 15% to 34% from 1970 to 2012 respectively  . The racial segregation of space tends to be related to racial inequalities associated with educational attainment, occupational skills, and job market placement. Some of the research indicates that workplace racial segregation is linked to differences in educational attainment and skill acquisition. Moreover, contrasting the wage differences between black and white employees can be explained by differential educational attainment  .
The differential availability of housing tends to result in de facto racial segregation in metropolitan areas. This may be attributed to generational predisposing, where the poor quality of education in a low-income area will inhibit a child’s abilities and skills. Therefore, family structure and school systems influence a child’s educational attainment. Adults with higher education tend to have a more vested interest in their child’s own education and schooling environment, which in turn enhances the quality of education locally. Urban developers may cater to more affluent families by building subdivisions with better access to higher quality amenities, such as nicer schools and safer neighborhoods. Indirectly, this contributes to continued racial and social class segregation. As a result, it becomes much more difficult for poorer families to break out of a cycle of poverty  .
Measures of cognitive skills and academic performance have been used to assess neighborhoods. The findings were that living in poor neighborhoods resulted in substantial declines in reading and language skills. Violence in African American children’s environments had a direct negative influence on tests of reading, language, and problem solving by more than a third. Exposure to nearby homicides impacted children’s vocabulary assessments as well as impulse control and attention  .
Workplace racial segregation is influenced by both race distribution difference within occupational categories and between occupational categories. Historically, government approaches for reducing segregation focused on racial disparities within occupational categories. Efforts, such as Affirmative Action policies, were aimed at minimizing disparities over a prescribed time-period. These seemed to have a positive impact on racial distribution within occupational categories from the late 1960’s through the middle 1980’s. Addressing racial participation between occupational categories has been more problematic. Racial minority individuals continue to be over-represented in lower, less professional occupations in comparison to their white counterparts  .
Segregation in the workplace appears to be influenced by educational attainment and racial background. Several studies suggest that organizations separate workers by skills related to educational attainment. In the United States, job skills are related to race and influence racial segregation in the workplace. In addition, position segregation based on job skills found among white employees. This supports the fact workplace separation takes place within as well as across racial groups with educational attainment playing a role in this type of separation. Finally, Hellerstein and Neumark  found that work skills were determined by educational attainment and differences in educational attainment between blacks and white led to wage gaps in the economy.
The current composition of the United States workforce seems to support the prior discussion of racial segregation in the workplace. As of 2016, roughly 78% of the workforce was white (including Hispanic and non-Hispanic). Hispanics (across all racial categories) were responsible for 17% of the workforce. Asian Americans comprised approximately 6% of the workforce and blacks were 6%  .
3. Theoretical Approach and Research Hypothesis
The sociological approach from a conflict theoretical perspective can be employed for understanding the social dynamics impacting racial segregation in the contemporary workplace. Conflict theory provides a foundation for the development of a conceptual race relations framework. Functional conflict as presented by Lewis Coser makes four assumptions concerning society  . One suggests that society is comprised of interrelated parts. Secondly, these parts influence the others. Thirdly, a change in one part causes a change in another. Lastly, these parts often display conflict, instability, and social change. Conflict is not inherently dysfunctional but can be integrative with a potential for improving the adjustment of society. Therefore, social conflict is explained through the manifest and latent interests of loosely formed groups  . Loosely formed groups can be extrapolated to include race.
3.1. Establishing Dominant and Subordinate Groups
The ability to assign individuals to a group is a critical factor for creating social inequality. Social attributes, either physically tangible or socially created, can be used to determine group membership. Within society there must be agreement regarding these differences and the importance associated with them. As a result, group membership creates a social perception of an in-group versus an out-group orientation between people  . Social attributes are major components through which individual as well as group interactions occur. Access to power, authority, and resources is differential.
Relations between different social groups begin with social differentiation. It is the distance or separation between groups in a society as well as the degree of internal separation within organizations  . This is a fundamental step which establishes boundaries between groups of individuals. Therefore, social differentiation is a process whereby social units are horizontally separated according to socially defined criteria. It is important to note that no substantial social meanings are assigned to the groups created based on the socially defined criteria. Using race as an example, a socially defined criterion is established for categorizing different groupings of individuals based on social definitions of skin color. Groups are identified ranging from lighter to darker skinned individuals. No social meanings are assigned to the social distinctions made between groups.
Once social meanings are assigned to the criteria, such as whites are more intelligent than blacks or blacks are physically superior to whites and better suited for labor occupations, a critical conceptual movement is made from differentiation to ranking. Social ranking is a process where groups are vertically arranged based on the assignment of meaning to group distinctions. The group at the top of the vertical arrangement is considered more valuable than those located below. Social differentiation is a prerequisite for ranking. Social ranking represents the initial development of stratified relationships between groups based on differential access to social power  .
Social differentiation and ranking are the underpinnings for viewing inter-group relations. For example, racial groups can be horizontally separated based on physical appearance. In the abstract, the social perception of physical appearance distinctions (largely skin-color) is vague. Van den Berghe  argues when differences are given moral or social value, distinctions become important for social ranking groups of individuals. Once meanings are assigned to social distinctions, groups become unequal and eventually this inequality becomes institutionalized within society. This process is known as social stratification. Institutionalized inequality creates differential access between groups with respect to economic, educational, political, and social opportunities.
3.2. Conceptual Framework
The inequality that precipitates from social differentiation and ranking must be accomplished through a legitimization process. Members of both the dominant group and the subordinate group participate in a process that rationalizes the existence of racial inequality. A conceptual framework is suggested to demonstrate how various concepts impact the legitimization of inequality.
The attitudinal component is initially formed in support of racial inequality and this is illustrated in Figure 1. Ethnocentrism, a key conceptual aspect of the attitudinal component, represents the way an individual interprets situations he/she encounters. Individuals and the groups they represent tend to view themselves as different from others and in-group and out-group categorizations emerge. The meanings placed on physical characteristics are shaped by ethnocentrism. Geschwender  emphasizes that inequality between blacks and whites was framed by the initial contact Europeans had with Western Africans. Being technologically advanced, Europeans attributed their superiority to a blend of Biblical interpretation and economic greed. Racism and prejudice are predictable by-products of ethnocentrism.
Once a dominant-subordinate relationship is formed in a society, racism is used to legitimate and rationalize it. Racism can be defined as “any set of beliefs that organic, genetically transmitted differences (real or imagined) between groups are intrinsically associated with the presence or the absence of certain socially relevant abilities or characteristics  .” Racism culturally sanctions the dominant group occupation of important positions in society. This contributes to the social power associated with dominant group members. As a belief system, racism spawns two other components: prejudice and stereotypes.
Prejudice is comprised of attitudes that a person has concerning members of
Figure 1. Components of contemporary intergroup relations in united states.
another group. It entails an unfavorable attitude toward members of a specific racial or ethnic group. Prejudice is influenced by cultural transmission, ethnocentrism, and racism. Prejudice represents a personal socialization process wherein racism provides beliefs about subordinate groups in society. Racism is a belief system while prejudice is an attitude at the individual level.
Stereotyping often accompanies prejudice. These over-generalized pictures of members of a rival racial group provide images that assist in legitimizing negative attitudes and subsequent differential behavior. Racism, prejudice, and stereotyping together form the attitudinal component that rationalizes the use of discriminatory behavior for the maintenance of inequality.
The behavioral component of the conceptual framework, as presented in Figure 1, is composed of three types of discrimination: personal, institutional, and interlocking. Discrimination refers to actions or methods used by members of the dominant group that differentially affect members of the subordinate group. Personal discrimination refers to differential treatment by an individual against members of a subordinate group  . This can be extended to include any injurious or differential treatment motivated by racial group membership. Therefore, anyone can exhibit individual discrimination irrespective of group membership.
Institutional discrimination plays a critical role in differential distributional patterns between groups of people within organizations. Broadly, institutional discrimination refers to organizational policies, guidelines, or actions that adversely affect racial minorities. Institutional discrimination can be conceptually separated into two types: intentional and unintentional.
Intentional institutional discrimination refers to organizational policies, actions, or guidelines that are intentionally designed to adversely affect racial minorities. This definition is a derivative of one offered by Feagin & Feagin  , who termed it “direct institutional discrimination” and described it as “socially prescribed actions that, by design, have a differential and adverse impact on members of subordinate groups.” Jim Crow America is replete with examples of this type of discrimination to include racially separate washrooms, water fountains, public schools and transportation. A system of racial etiquette rationalized the racism, prejudice, and stereotyping which supported the differential institutional behavior.
Unintentional institutional discrimination refers to organizational policies, actions, or guidelines that adversely affect racial minorities although they were not designed to harm. Feagin  presents a similar term called “indirect institutional discrimination” and it refers to “actions that have a differential impact on members of subordinate groups even though they may not be intended to harm”. Unintentional institutional discrimination is often the barrier remaining after society has removed intentional forms of discrimination  .
Institutional discrimination is directly and indirectly influenced by racism, prejudicial attitudes and profit motives stemming from racism. This leads to various organizations being interlocked and influencing each other. Therefore, discrimination occurring in one organization impacts activities in other organizations  . For instance, discrimination that creates educational attainment inequality influences job placement in economic organizations. Institutional discrimination is complex and cumulative. It is linked to the cultural value climate in society. The assumption in American society is that a meritocracy exists, and this is guided by the cultural value climate. For instance, standardized testing for college entrance is seen by many decision-makers as a fair, colorblind screening approach. But standardized testing represents unintentional institutional discrimination reflected by racially segregated school districts at the secondary education level. Differential placement in the economic sector is result of discrimination in the educational system.
Institutional discrimination reflects the removal of attitudes from the behavior. The impersonal nature of this type of discrimination separates the individual perpetrator from the racial ideas and attitudes that initially created the institutional policy  . The conceptual framework indicates that discriminatory behavior may exist in society regardless of the presence or absence of racial attitudes. Over time, the elimination of racial attitudes may not eliminate all forms of discrimination, especially unintentional institutional discrimination. There tends to be a natural tension between attitudes and behaviors. The existence of unintentional institutional discrimination (i.e., housing segregation, school segregation, and workplace segregation) may in fact influence people’s perceptions of racial differences. This could result in an increase in negative racial imaging and continued racial segregation.
Using institutional discrimination as a frame of reference, indicators of differential treatment based on race can be established. If nothing were impacting workplace composition, both black and white workers should be found in work settings that are composed mostly of white employees. This follows the rationale that the majority of those in the workforce are white. From the discussion of the conceptual framework and how attitudinal and behavior elements impact structural dynamics between racial groupings, the following research hypotheses were delineated.
H1: White employees are more likely to be found in work settings comprised of mostly white employees while black employees are more likely to be found in workplaces comprised of racially diverse employees.
H2: Educational attainment and personal income influence the perception of the racial composition of work settings. Individuals located in workplaces which are mostly white will exhibit higher educational attainment and personal income in comparison to those workplaces that are more diverse.
H3: Occupational prestige and the number of employees at the worksite influence the racial composition of work settings. Work settings perceived to be mostly white will exhibit higher occupational prestige scores. In addition, work settings perceived to be mostly white will be associated with smaller numbers of employees at the overall worksite.
The data used to examine perceptions of racial inequality in the workplace were obtained from the General Social Survey 2000 and 2018 files (Smith, 2018)  . These data provide a great deal of information on core demographic, behavioral, and attitudinal questions including those related to racial inequality (GSS 2018). Statistical analyses were conducted using the IBM Statistical Package for the Social Sciences (SPSS).
This study employed multiple logistic regression analysis utilizing a number of research variables. The sociological theory discussed earlier provided the foundation for the statistical analysis.
Dependent Variables. The database included one variable related to perceptions of racial segregation in the workplace. It was categorical and used as the dependent variable. How respondents were queried is displayed below.
Question. “Which best characterizes your current workplace?”
The response categories are 1 = All white, 2 = Mostly white, 3 = Half white and half black, 4 = Mostly black, 5 = All black.
Independent Variables. The analysis considered eight possible predictors to one’s attitudes regarding racial segregation in the workplace. There are five demographic variables which includes race (dummied into two categories; 0 = White, 1 = Black), gender (dummied into two categories; 0 = men, 1 = women), age (spanning 18 - 89 years of age), highest years of formal education completed (ranging from 0 - 20 years), and annual personal income. There are three variables associated with the workforce; labor force status (1 = working full-time, 2 = working part-time, 3 = not working), occupational prestige score (0 to 100), and number of employees at the respondent’s work site (1 to more than 500 employees).
Two multiple binary logistic regression models, guided by our theoretical underpinnings, were created for responses linked to workplace racial segregation. It should be noted that the dependent (response) variable was recorded into a dummy variable to facilitate the regression analysis. Workplace composition was collapsed into two categories (Mostly white = 0 and Not mostly white = 1).
Control Variable. The effects of the independent variables were explored over an eighteen-year period. Therefore, year was employed as a control variable. Comparisons of independent variables effects were conducted for 2000 and 2018.
5. Study Findings
The sample was comprised on individuals who were working at the time of the survey data collection. For 2000, 26% of the respondents indicated their workplaces were made up of all white employees and 48% said their workplaces were made of mostly white employees. In 2018, 18% indicated their workplace was composed of all white employees and 43% said their workplaces were mostly white (see Table 1). Over the eighteen-year period, it appears to workplace became
Table 1. Summary of study predictor variables associated with perceptions of racial segregation in the workplace, 2000 and 2018.
slightly more diverse. It should be noted that since the majority of the workforce is white, the research expectations would be to discover that most of the respondents should indicate their workplace composition is mostly white.
Regarding traditional demographic variables of race, gender, and age, the sample distribution remained relatively stable between 2000 and 2018. In 2000, about 85% of the sample was white and 15% black. Approximately half were female, and half were male. The average age of the respondent was 41 years. For 2018, 81% of the sample was white and 19% black. With respect to gender, 53% were females and 47% were males. The average age was nearly 43 years (see Table 1).
The socio-economic variables provided some interesting distributions. In 2000, approximately 82% of the respondents earned less than $50,000 annually. This percentage declined to 62% in 2018. The average annual personal income was $22,873 for 2000 and $25,626 in 2018. The average number of years of formal education was 13.7 years in 2000 and 14.1 years in 2018. About 28% of the respondents had a college degree in 2000 and that percentage increased to 34% in 2018 (see Table 1).
Work-related variables were informative as well. The sample was comprised overwhelmingly of individuals who worked full time. For both 2000 and 2018 more than 80% were full-time workers.
Respondent perceptions of work environment composition indicated that the contemporary workplace tends to be racially segregated. Although the company or organization may have both black and white employees, respondents stated these group members were generally segregated from each other (see Table 2). Most white employees were located in work settings comprised mostly of other white workers. Conversely, black employees were found in workplaces where workers were more evenly distributed by race. More specifically, Table 2 shows that in 2000, 80% of white employees stated their workplace was mostly white while 44% of blacks said their workplace was mostly white. In 2018, white employees that described their workplace as mostly white dropped to approximately 71%. Black employees who said their workplace was mostly white declined to 31% and those that indicated their workplace was not mostly white increased to 69%. It can be concluded that workplace distributional racial differences lessened modestly between 2000 and 2018.
Employee age distributions were similar for black and white employees over the eighteen-year time-period. In 2000, the majority of worker for both groups were between the ages of 25 and 54. This remained the same for 2018 although white employees tended to be older in comparison to black employees (see Table 2).
Educational attainment differences between black and white employees diminished between 2000 and 2018. However, differences in personal income remained the same. In 2000, whites had higher educational attainment in comparison to blacks (see Table 2). In 2018, there was no statistical difference between the two groups. Most white and black employees had a high school diploma with
Table 2. Perceptions of workplace racial composition by race, 2000 and 2018.
*p < 0.05, **p < 0.01, ***p < 0.001, ns-not statistically significant.
some college experience. Slightly over 30% of both groups earned a college degree. With respect to personal income, white employees earned more than their black counterparts. Larger percentages of blacks earned less than $50,000 annually in comparison to whites. Conversely, larger percentages of white employees earned more than $50,000 per year. This trend was found for both 2000 and 2018 (see Table 2).
Workplace variable comparisons were interesting. Table 2 shows that whites tended to have significantly higher occupational prestige scores than blacks in 2000 and 2018. It can be concluded that white employees occupied positions that were considered more important or valuable to the organization. There was little to no difference between the two groups with respect to labor force status and the number of employees at the work site.
The cross-tabulation distributions were further analyzed using simple binary logistic regression analysis. The findings show that employees generally perceive a racially segregated workplace in their places of employment (see Table 3). Using a bivariate model and excluding all of the predictor variables with the exception of race, employees indicated they work in a setting comprised mostly of white employees. However, black employees were nearly five times more likely to say they work in settings where racial composition is either half white and half black or mostly black. These trends remained stable from 2000 to 2018. In 2000, race of the employee was responsible for 11.3% of the perception of workplace segregation. In 2018, it was responsible for 12.5% of the workplace segregation perception.
Multiple binary logistic regression was employed to determine the potential impact of other variables on variation in the perception of workplace segregation. The findings are displayed in Table 3. For 2000, race of the employee,
Table 3. Summary of binary logistic regression models of selected independent variables on workplace racial composition, 2000 and 2018.
*p < 0.05, **p < 0.01, and ***p < 0.001.
number of employees at the worksite, and occupational prestige scores impacted perception of workplace segregation. Race of the employee had the strongest effect on perceptions of workplace segregation. Black employees were more likely to work in diverse workplaces while white employees were more likely to work in mostly white settings. Blacks were four times more likely to work in a diverse setting in comparison to their white counterparts. As the number of employees increased at the worksite, individuals were more likely to indicate they worked in racially diverse workplace. Conversely, as the number of employees decreased at the worksite, respondents were more likely to state they worked in a mostly white workplace. Employees in positions with higher occupational prestige scores tended to work in mostly white workplaces while those in positions with lower occupational prestige scores worked in mostly diverse workplaces. This multivariate model explained 16.3% of the variation in workplace segregation perception. The additional predictor variables explained an additional 5% of the variation in comparison to the bi-variate logistic regression model.
The multiple binary regression analysis findings for 2018 are presented in Table 3. Race of the employee, educational attainment, and the number of employees at the worksite influenced employee perception of workplace segregation. Race of the employee had the strongest impact on perceptions of workplace segregation. Blacks were four times more likely to work in a diverse setting in comparison to their white counterparts (unchanged from 2000). Educational attainment influenced perceptions of workplace segregation. For each year of educational attainment increase, an employee was more likely to work in a mostly white workplace. Those will less educational attainment tended to work in a more diverse workplace. Individuals located in larger worksites perceived themselves as working in more diverse workplace settings. Conversely, those in smaller worksites perceived themselves as working in mostly white workplace settings. This multivariate model explained 16.8% of the variation in workplace segregation perception. The additional predictor variables explained an additional 4.3% of the variation in comparison to the bi-variate logistic regression model.
6. Discussion and Summary
The findings clearly demonstrate that from a perceptional standpoint, black and white employees generally work in different racial compositional environments. Black employees were found to work more in diverse workplaces in comparison to whites who tended to work in environments where most of their counterparts were also white. For 2000, racial background of the employee explained 11.3% of the variation in perceptions. For 2018, it explained 12.5% of the variation in perceptions of workplace racial segregation. Other variables had only a small influence on this overall relationship. In 2000, occupational prestige and the number of employees at the work site had a minimal influence on the overall relationship between the race of the employee and the composition of the workplace. In 2018, educational attainment and the number of employees at the worksite influenced the overall relationship between the race of the employee and the composition of the workplace. For both comparative years, age, personal income, work status, and sex had no influence on perceptions of workplace segregation. It should be noted that each of the multivariate statistical models for 2000 and 2018 explained about 16% of the change in workplace segregation perceptions.
Based on the study results in Table 2 and Table 3, Hypothesis 1 is accepted. Therefore, it can be concluded that white employees are more likely to be found in work settings comprised of mostly white employees while black employees are more likely to be found in workplaces comprised of racially diverse employees.
Regarding Hypothesis 2, the findings support partial acceptance. Educational attainment was found to influence perceptions of workplace segregation in 2018 but not in 2000. As a result, it can be concluded that educational attainment influences racial composition of the workplace. Those who work in mostly white workplace environments tend to have higher educational attainment in comparison to those with lower educational attainment. In addition, personal income has not influenced perceptions of workplace segregation.
The findings support acceptance of Hypothesis 3. For both 2000 and 2018, workplaces with more employees tended to exhibit larger numbers of more employees indicating their workplaces were more diverse. In contrast, smaller worksites tended to have more employees who perceived their workplaces to be mostly white. Additionally, for 2000, higher occupational prestige scores were associated with workplaces that were composed of mostly white employees.
This research effort demonstrates that employee perceptions of workplace racial segregation have remained constant over an 18-year period. The findings show that black employees view diversity in the workplace much differently than their white counterparts. Blacks see themselves in more diverse work settings. Whites, on the other hand, view their work settings as mostly white. Other variables such as educational attainment, occupation prestige, personal income, and size of the work setting have minimal impact on racial segregation perceptions.
Future studies on racial segregation in the workplace should employ case study or focus group methods to obtain more perceptional information. Although limited with respect to generalizations, empirical data from these techniques would add to the overall understanding of workplace racial segregation. In addition, types of organization should be targeted for future analysis as well. For instance, employees in non-profit or governmental organizations may display different racial segregation perceptions than those working in for-profit organizations.
1The terms black and African American are used interchangeably.
 Settles, I.H., Buchanan, N.T. and Yap, S.C.Y. (2010) Race Discrimination in the Workplace. In: Paludi, M.A., Paludi Jr., C.A. and DeSouza, E., Eds., Handbook on Understanding and Preventing Workplace Discrimination, Praeger Publishers, Westport.
 Hellerstein, J.K. and Neumark, D. (2008) Workplace Segregation in the United States: Race, Ethnicity, and Skill. Review of Economics and Statistics, 90, 459-477.
 Galster, G. and Sharkey, P. (2017) Spatial Foundations of Inequality: A Conceptual Model and Empirical Overview. RSF: The Russell Sage Foundation Journal of the Social Sciences, 3, 1-33.
 Smith, T.W., Marsden, P.V. and Hout, M. (2018) General Social Surveys (GSS), 1972-2018. National Opinion Research Center, Chicago, The Roper Center for Public Opinion Research, University of Connecticut, Storrs.