Abstract
Using two administrative data sets from the Equal Employment Opportunity Commission (EEOC), this study examines the relationship between employer and environmental characteristics and Americans With Disability Act (ADA) discrimination charge rate. Results of a multiple regression analysis using a sample of mid- to large-sized private employers indicate that establishment size is negatively correlated with ADA charge rate, whereas several other employer characteristics are positively associated with charge rate, including parent organization size, federal contractor status, transportation or services industries, and relatively high minority representation. One of the main concerns of contemporary disability policy is reducing discrimination in employment, and our findings can inform employers, policymakers, and organizations working with employers to reduce perceived discrimination by identifying those employers most likely to receive charges. Further research is needed to better understand what specific behaviors, practices, and policies within these different types of establishments explain their differential charge rates.
In the past three decades, the employment rate for people with disabilities has been approximately half (or less) the employment rate of people without disabilities (Bjelland, Burkhauser, von Schrader, & Houtenville, 2011). This substantial gap remains even after adjusting for a wide range of individual- and job-related characteristics (Kaye, 2009). One explanation for this disparity is that discrimination against people with disabilities by employers and co-workers makes entry into the workforce and integration into the workplace challenging. Individuals with disabilities report being passed over for jobs because of their disabilities, while those who are employed report feeling marginalized and harassed, as well as receiving lower pay, less job security, and fewer opportunities for training compared with their non-disabled peers (Kennedy & Olney, 2001; Kessler/National Organization on Disability (NOD), 2010; Robert & Harlan, 2006; Schur, Kruse, Blasi, & Blanck, 2009). According to many individuals with disabilities, discrimination in the workplace is so pervasive that they choose not to disclose the presence of their disability out of fear of not being hired, being fired, being harassed, being treated differently, or being devalued in the workplace (Brohan et al., 2012; Dalgin & Gilbride, 2003; von Schrader, Malzer, & Bruyère, 2013).
Those who choose to disclose their disabilities are protected against such employment discrimination and guaranteed access to possible needed accommodations by Title I of the Americans With Disabilities Act (ADA) of 1990, as amended. In each year since 2005, there has been an increase in the number of employment discrimination charges filed under the ADA (U.S. Equal Employment Opportunity Commission [EEOC], 2012). This is a concern for employers who want to minimize the costs, stress, and the potential damage to their reputation associated with charges (Goldman, Gutek, Stein, & Lewis, 2006), and presumably want to create a workplace culture where individuals are not faced with discrimination. There are now several factors that may increase the number of individuals with disabilities in the workforce including an aging workforce, veterans returning to the civilian workforce, and federal initiatives such as the new regulations addressing the implementation of Section 503 of the Rehabilitation Act of 1973, as amended, suggesting that 7% of a federal contractor’s workforce across job groups should be composed of individuals with disabilities (Affirmative Action and Nondiscrimination Obligations of Contractors, 2013). With more individuals with disabilities in the workforce, there may be further increases in charges if employers are not ready to successfully accommodate disability.
In this study, we use two administrative data sets from the EEOC to examine employer characteristics associated with disability discrimination charge rates of mid- to large-sized employers using a multiple regression analysis. The study sheds light on what types of employers are being charged at higher rates with employment discrimination under the ADA, an area that has not been well explored before due to data limitations. By identifying employers most at risk for receiving charges, outreach can target limited resources to support these employers in developing good practices around disability inclusion. Our findings can support the EEOC and other organizations (e.g., ADA National Network regional centers, state protection and advocacy agencies, and others) that work with employers to reduce disability employment discrimination.
ADA Charge Data
Like other federal employment non-discrimination legislations, Title I of the ADA is enforced by the EEOC. An individual who perceives discrimination in the employment process must first file a charge at the administrative level—with either an EEOC office or a state or local Fair Employment Practice Agency (FEPA)—prior to filing a lawsuit against the employer. Information on charges filed under Title I of the ADA and other statutes enforced by the EEOC is stored in the EEOC’s Integrated Mission System (IMS), which has been an invaluable data source to researchers. While not every (actual or perceived) discriminatory act leads to a charge being filed and not every charge represents actual discrimination, we believe that the charge data do provide guidance that can be helpful in designing and refining policy. As noted by scholars who have used the charge data extensively,
. . . given the complexity of identifying discrimination, researchers interested in systematically studying discrimination face a conundrum: efforts to operationalize the concept will inevitably be imperfect, but eschewing the study of discrimination because of its operational difficulties forestalls research that could have important implications for understanding ascriptive inequality and policies to alleviate it. (Hirsh & Kornrich, 2008, p. 1399)
Previous studies have shown that ADA discrimination charges most commonly cite issues of discharge, reasonable accommodation, terms/conditions, harassment, and hiring, while the most commonly cited specific bases are orthopedic and structural impairments of the back, non-paralytic orthopedic impairment, diabetes, depression, and other psychiatric disorders (Bjelland et al., 2010). The prevalence of these issues/bases among ADA charges differs according to employer characteristics (e.g., McMahon et al., 2008). Through 2010, ADA charges relative to the number of people with disabilities in the labor force have risen, while charge rates for other protected groups (i.e., older, female, or non-White) have remained stable over time or even decreased (Bruyère, von Schrader, Coduti, & Bjelland, 2010).
Currently, the literature lacks a comprehensive study on the relationships between employer characteristics and the ADA charge rate for at least two reasons. There are limitations to what the charge data alone can reveal about the characteristics of employers receiving charges. The quality of some employer data on the charge files may be limited as it must rely on the charging party’s knowledge of the firm, which might be incomplete (Ron Edwards, EEOC, personal communication, November 15, 2011). For example, more than 20% of charges are missing employer characteristics in data prior to 2007 (Bjelland et al., 2010), and industry information is missing for close to 50% of private business charges between 2008 and 2010 (personal communication, 2012). In addition, working with the charge data alone does not provide a comparison group of employers who did not receive charges to use in analysis.
Using the EEO-1 Employer Report to Supplement Charge Data
To address these issues, our study takes a different approach from previous work on ADA charges and examines employer factors related to an increased disability-related discrimination charge rate using two administrative data sets from the EEOC: ADA charge data from the IMS and the Employer Information Report (EEO-1 report). The EEO-1 report is a rich source of data for medium and large private sector employers (Cartwright, Edwards, & Wang, 2011). The EEO-1 report is a compliance survey report that is completed annually by establishments with 100 or more employees, establishments with fewer than 100 employees that are owned by or affiliated with another company that employs more than 100 employees, or federal contractors with 50 or more employees. Overall, EEO-1 reports cover about 40% of all private sector employment (Robinson, Taylor, Tomaskovic-Devey, Zimmer, & Irvin, 2005). The EEO-1 report includes information about the size of both the establishment and the parent organization. Establishments must describe the composition of their workforce by completing a matrix, placing all full- and part-time employees in cells by occupational group, Hispanic status, and gender; those who are non-Hispanic are further classified by race. The report is used by both the EEOC to enforce employment non-discrimination legislation (e.g., Title VII of the Civil Rights Act) and by the Office of Federal Contract Compliance (OFCCP) to regulate federal contractors by using the information to select employers for compliance reviews. The quality of information on private employers is more complete and accurate in the EEO-1 report because it is reported by the establishment. Unfortunately, the EEO-1 report does not contain any disability-related content.
Hirsh and Kornrich (2008), using the method of hand-merging, linked the EEO-1 and IMS data to study the manner in which workplace and institutional conditions affect race and sex charge rates from 1990 to 2002. In our study, we used a similar approach by linking these files to examine the characteristics of employers who received ADA charges in 2009.
Method
Sample and Construction of the Analysis File
A sample of 3,600 private employers was selected from approximately 230,000 individual establishments in the 2009 EEO-1 employer report file with establishment size of 50 employees or more. The sample was selected using probability proportional to size (PPS) sampling, similar to the approach used by Hirsh and Kornrich (2008). In this approach, an establishment’s probability of selection is proportional to its size, allowing us to have more establishments with charges in our sample, as larger establishments will generally have more employees eligible to file a charge. To account for this approach and get estimates for EEO-1 employers, we use sampling weights in our regression analyses.
We next merged our sample of establishments to the 2009 charge data, linking the two files by zip code. Then each record was reviewed by hand to identify establishments in the charge data that matched with an establishment in the sample. To increase the accuracy and speed of merging these files by hand, names and addresses of the employer record in both the EEO-1 report and the charge data were standardized using code developed by the authors. Calculating a match rate is not possible because many establishments in the charge database are not included in the EEO-1 report, either because they are too small or because they are not private businesses. As noted, the employer variables on the charge data are often missing or not reliable, so we used all charges in matching to the employer sample. To see whether we were matching charges to establishments at a reasonable rate, we compared percent of establishments with charges for sex and race with the results of Hirsh and Kornrich, our merging resulted in slightly lower but comparable rates of sex (female and pregnancy) and race (non-White race, color, and non-White national origin) charges under Title VII per establishment per year. More details on the matching method are available from the authors. Five establishments were dropped because of missing data on federal contractor status leading to a sample size of 3,595 for analysis.
Dependent and Independent Variables
In this study, we were interested in employer characteristics associated with our dependent variable, ADA charge rate. The choice of dependent and independent variables was partially informed by the work of Hirsh and Kornrich (2008). Specifically, we included some parallel independent variables such as establishment size and parent organization size, hypothesizing that larger establishments have more formal structures in place, and therefore are less likely to behave unlawfully and be charged with discrimination (Harcourt, Lam, & Harcourt, 2005). There is also evidence that employers in larger organizations express more positive attitudes toward workers with disabilities, although this effect seems to be smaller in more recent studies (Hernandez, Keys, & Balcazar, 2000). We also included the percentage of employees in the establishment who are managers, which Hirsh and Kornrich (2008) found was associated with more charges on the basis of race and sex, suggesting that with more close supervision there may be more opportunity for “contentious interaction with supervisors” (p. 1402).
One of our interests in this study was to understand the relationship between diversity (in terms of race and gender) and ADA charges. While having more female and racial/ethnic minorities may indicate an environment that is more inclusive, we hypothesized that having more protected group members would lead to more opportunities for charges that can be filed citing multiple statutes. As suggested by Barrington and Troske (2001), we included two measures of diversity: occupational diversity that indicates the diversity along the occupational ladder (how evenly women/minorities are distributed across occupations) and payroll diversity that measures how diverse the organization is compared with the workforce in the industry. We calculated occupational segregation indices for both race and sex (Massey & Denton, 1988) as was done in Hirsh and Kornrich (2008) and two separate binary indicators that compared minority and female representation within the establishment to the state level for the given industry.
In addition, we included variables that described the institutional environment, including federal contractor status, establishment type (single-unit, multi-unit headquarters, and multi-unit individual establishment), and industry. Federal contractors are required to have affirmative action plans as well as to maintain and analyze data on the sex and race/ethnicity of their applicants and employees. Ideally, this sort of plan would reduce perceived discrimination in the workplace; in reality, however, there may actually be more charges, as federal contractors are required to prominently post information about protections against discrimination under federal law. There seems to be little evidence that these plans reduce race and sex discrimination charges (Hirsh & Kmec, 2009; Hirsh & Kornrich, 2008). Private businesses completing the EEO-1 report identify themselves as a single establishment, multi-unit headquarters, or a multi-unit individual establishment, and these were indicators included in the model. We anticipated that the headquarters may have more charges, as the workforce may be more highly educated and aware of their rights. We included six industry groups: manual, transportation, sales, professional, health services, and other services. Our decision to break down the analytical sample into six industrial groups was based on patterns that we observed in the data. Specifically, in the data, we identified a natural grouping of the two-digit North American Industry Classification System (NAICS) groups into six major industrial groups by examining the number of establishments and ADA charges received by establishments within the groups.
Our model also included local labor-market-specific variables that we anticipated might be related to charges. These included county prevalence of disability among the working-age population, county unemployment rate, and disability charges per 100,000 workers 18 to 64 years old in the state. We anticipated that establishments in counties with higher rates of disability might have fewer charges, as exposure and positive experience with disability might lead to less perceived discrimination. At least at an organizational level, this seems to be supported (Hernandez et al., 2000). We included the county unemployment rate in the model to control for differences in the labor market conditions across counties. We expected that in areas with a higher unemployment rate, establishments would be at higher risk of receiving ADA charges. We also anticipated that establishments in states with higher rates of disability charges would experience more charges, as there is more familiarity with the process among the labor force.
To calculate local (county-level) conditions, such as disability prevalence rates and unemployment rates, we used the American Community Survey estimates from 2008 to 2010 available from American FactFinder. The choice of this range of years was determined by the desire to get county-level estimates, which are not available for less populated counties for single years. We computed disability charges per 100,000 workers between 18 and 64 years old in the state using two separate data sources. First, we calculated the number of disability charges filed with EEOC or FEPA offices for each state during fiscal year 2009 using the IMS database. Then, using the Current Population Survey–Annual Social and Economic Supplement (CPS-ASEC), we estimated the number of workers between 18 and 64 years old for each state.
Empirical Model
In the empirical specification of charges, we assumed that the expected number of ADA charges for the ith establishment follows a Poisson distribution.
In Equation (1), the number of ADA charges in the ith establishment is given by Yi and is the product of the exposure parameter, ci, the number of individuals with disability in the ith establishment and intensity parameter, λ i , and the number of ADA charges per individual with a disability in the given establishment. The exposure parameter takes into account that establishments with more members of the protected group will have more individuals who can file a charge. Unfortunately, we do not observe in the data the number of people with disclosed disabilities in the establishment. Therefore, we estimated the exposure parameter, ci, by taking the number of people with disabilities who were employed in the county and dividing that count by the total number of people who were employed in the county, and then multiplying that ratio by the size of the establishment to approximate the number of individuals with disabilities in the ith establishment. While not a perfect measure, with the limitations of the data, we believe this is a reasonable approach. We refer to our dependent variable as a “charge rate” because it adjusts the number of charges for the number in individuals in the protected group in the establishment.
We further assume that the intensity parameter is an exponential function of the location-specific (Z) and employer-related (X) variables. A simple algebraic manipulation provides us the final statistical specification for the relationship between the expected number of ADA charges for the ith establishment and its employer and location characteristics
A key assumption of the Poisson regression is equi-dispersion of the dependent variable (Cameron & Trivedi, 2005) such that the mean of the dependent variable should equal its variance. The weighted mean of ADA charges was 0.05 charges per establishment while the variance was 0.2842 = 0.08 indicating a moderate level of over-dispersion in the main dependent variable. The over-dispersion problem can be handled in two ways: either by using a more flexible specification for the count process, for example, negative binomial, or by simply estimating robust standard errors for the Poisson estimates (Cameron & Trivedi, 2010). In this study, we followed the second option.
Discrimination charges may be jointly filed; that is, a single charge can cite more than one anti-discrimination statute. Because nearly one third of ADA charges are jointly filed (Bruyère et al., 2010), we controlled for the number of Title VII–Race charges received by the establishment in 2009 in our analyses. By controlling for this variable, we also attempted to eliminate the potential for spurious correlations between minority-related variables and the main dependent variable of the analysis. That is, if the number of race charges increased with the proportion of minorities within the establishment, and at the same time, the number of race charges is positively correlated with the number of ADA charges received by the establishment, then the coefficients for minority-related variables in the absence of the number of race charges in the regression analysis would not have a clear interpretation. Our preliminary analysis showed that the numbers of race and sex charge rates were strongly correlated; therefore, we expected that by introducing the number of race charges into the regression analysis, we would eliminate any biases in female-related employer variables in our final specification of ADA charges. Due to concerns about over-specification of the model, we decided to control only for the number of charges that have the strongest association with ADA charges, which was the number of race charges.
Multi-collinearity, or high correlations between independent variables, can be another issue that may hinder the ability to correctly interpret estimates from the regression analysis. To evaluate the level of multi-collinearity in our empirical inquiry, we computed variance inflation factors (VIF) for each independent variable, a method commonly used in linear regression. The variance inflation factors in this analysis did not exceed 3 for any covariate, implying that the level of correlation across variables in our analysis has a statistically acceptable level (the conventional rule of thumb is that any variable with the VIF above 10 would merit further investigation).
A final important consideration in the selection of the appropriate empirical specification is related to the fact that establishments in the sample can be part of the same organization. If a relatively high proportion of establishments are part of the same parent organization, then the assumption that ADA charge rate across establishments is independent and identically distributed would be violated, compromising efficiency and consistency of estimates. To deal with the hierarchical feature of the sample, we chose to also fit Equation (1) with a population-averaged Poisson regression allowing correlation across establishments within the same organization. By comparing this approach to the conventional Poisson regression, we are able to see whether the potential correlation across observations in the sample affects the parameter estimates and their implications.
Results
Descriptive Statistics
Table 1 presents the frequency of receiving one or more ADA charges in the sample. Among our sample of 3,600, about 14% received at least one ADA charge; less than 5% of the establishments received more than one charge. When examining the weighted descriptive statistics in Table 2, the mean number of charges was 0.05; a lower value as the sample was drawn with probability of selection proportional to the size of the establishment, and larger establishments were more likely to have received charges. In our weighted analytical data set, the mean establishment size was 214 employees, and the mean parent organization size was nearly 64,000. Fourteen percent were single establishments, with the remainder being multi-unit establishments; nearly 40% were federal contractors. The most common industries among the establishments were manual and sales, each accounting for approximately 23%.
Number of ADA Charges Among Sample Establishments in 2009.
Note. ADA = Americans with Disabilities Act.
Descriptive Statistics of Key Establishment and Environmental Variables, 2009.
Note. The statistics presented are weighted to account for sample selection. ADA = Americans with Disabilities Act.
Manual industry includes manufacturing, agriculture, forestry, fishing, hunting, mining, utilities, and construction; transportation includes transportation and warehousing; sales includes wholesale and retail trade; professional includes information, finance and insurance, real estate and rental and leasing, professional, scientific, and technical services, and management of companies and enterprises; health services includes health care and social assistance; other services includes arts, entertainment, recreation, accommodation, food, administrative, support, waste management, remediation, educational, and other services (except public administration).
Results reported in Table 3 show that the gain from accounting for the potential correlation between establishments representing the same organization was insubstantial. Whereas the population-averaged Poisson provided more robust evidence of the potential positive association between the proportion of managers and ADA charge rate, other parameters from two models were comparable in terms of magnitude and statistical significance. However, the more parsimonious model showed the higher statistic for the Wald test suggesting the higher explanatory power. Therefore, in the remainder of the article, we present the results from the Poisson model (rather than the population-averaged Poisson) to make appropriate implications for policy and practices. In the following paragraphs, rather than reporting the more difficult to interpret Poisson regression coefficients from Table 3, we transform the coefficients to highlight the percentage change in ADA charges rate associated with each independent variable. Specifically, we use the following formula:
where β is the parameter of interest and δ is the unit of change.
Associations Between Employer and Environmental Characteristics and ADA Discrimination Charge Rate in 2009: Poisson vs. Population Averaged Poisson Regressions.
Note. The statistics are weighted to account for sample selection. In the Poisson regression, the standard errors are clustered at the state level, in the population-averaged Poisson regression, the standard errors are clustered at the individual level.
p < .10. **p < .05. ***p < .01.
Results show that the number of race charges and ADA charge rate were positively associated. The point estimate for the number of race charges can be interpreted as one additional Title VII–Race charge led to a 27% increase in ADA charge rate. As was expected, the correlation between the number of race charges and minority-related variables led to reductions in the magnitudes of the parameters of both the race occupational segregation index and the indicator of whether the proportion of minorities is above the state-industry mean. The ADA charge rate was 36% higher for establishments that employ a higher proportion of minorities than the state-industry level compared with their counterparts. Similarly, a one standard deviation increase in occupation segregation by race was associated with a (non-significant) 9% increase in the ADA charge rate.
Results also showed establishments with a proportion of females above the state-industry level were not more likely to receive ADA charges, and sex segregation was not significantly related to ADA charge rate. Similar to Hirsh and Kornrich (2008), we found that the establishment size was negatively correlated with the ADA charge rate while organization size was positively associated. Specifically, the point estimates for these variables indicated that a one log increase in establishment size decreased the ADA charge rate by 23%, while a similar increase in organization size increased the ADA charge rate by 16%.
Being either the headquarters of a multi-unit organization or a federal contractor was also positively associated with ADA charge rate. The rate of ADA charges was 83% higher for multi-unit headquarters compared with multi-unit individual establishments. Federal contractors, on average, had a 50% higher ADA charge rate than non-contractors.
ADA charge rates varied across industries. Compared with the reference group of “other services,” industries such as manual and sales had lower ADA charge rates, holding all else equal. Specifically, the ADA charge rates were 46% and 42% lower, respectively, in manual and sales industries than in other services. Some county- and state-specific factors were associated with the ADA charge rate. A 1% increase in the county prevalence rate of disability decreased the ADA charge rate by 12%. A 1% increase in county unemployment rate increased the ADA charge rate by 6%.
Analysis by Federal Contractor Status
As noted, the overall model estimates reveal that holding all else constant, federal contractors had a higher ADA charge rate compared with non-contractors. Because federal contractors are a group of particular interest, we did a sub-analysis by federal contractor status with results available on request from the authors and summarized below.
The number of Title VII–Race charges was positively associated with ADA charge rate, regardless of federal contractor status. However, the number of Title VII–Race charges had a more profound association in establishments with federal contracts, compared with their counterparts. Specifically, an additional Title VII–Race charge increased the ADA charge rate for federal contractors by 38%—more than twice the increase for non-federal contractors (18%).
The ADA charge rate decreased with increasing establishment size only among federal contractors. A one log increase in establishment size decreased the ADA charge rate by 34%. Although the same association for non-federal contractors had the opposite sign, the point estimate was not significant. The association between parent organization size and the ADA charge rate was similar across two groups of establishments; a one log increase in the organization size increased the ADA charge rate by about 14%.
Interestingly, the proportion of managers among non-federal contractors was significantly and positively associated with the ADA charge rate with a 1% increase in the proportion of managers increasing the ADA charge rate by 2%. Only for non-federal contractors did the headquarters have a higher ADA charge rate compared with multi-unit individual establishments. Specifically, compared with multi-unit individual establishments, the ADA charge rate was higher by 94% for headquarters without any federal contracts.
Federal contractors with a proportion of minority employees above the state/industry level had an 82% higher ADA charge rate compared with federal contractors with a lower proportion of minority employees. A similar association was not observed for non-federal contractors. Interestingly, however, a high level of sex occupational segregation within establishments without federal contracts led to a significantly higher ADA charge rate. Specifically, a one standard deviation increase in the sex occupational segregation index increased the ADA charge rate by 39%. In contrast, establishments with federal contracts lacked any association between ADA charge rate and the level of sex occupation segregation.
Discussion
The findings of this study are a first step in understanding which employers are more likely to receive ADA charges of discrimination. Informed by these findings, employers, policymakers, and those tasked with supporting employers to help limit perceived discrimination in the workplace can proactively address workplace policies and practices that might contribute to perceived discrimination. It appears that larger establishments receive fewer charges relative to smaller establishments; therefore, outreach efforts may want to target smaller employers, who may have fewer human resource structures in place to limit perceived discrimination.
The overall results also show that service industries (excluding health care) have more charges than manual or sales sectors. Therefore, service industry establishments may particularly value targeted support around decreasing charges. One possible explanation may be that there are differences in organizational practices and cultures in these industries, or that individuals employed in these industries differ in their knowledge of and/or attitudes toward charge filing. Another explanation could be a potentially higher rate of disability in service industries as compared with manual or sales. However, this seems unlikely given the relatively similar disability prevalence rates in these industries—according to estimates from the 2009 American Community Survey, the proportion of the workforce with disabilities in manual is 5.7% and in sales is 6.0%, not much different from the rate of 5.9% in services (author calculations from U.S. Census Bureau, American Community Survey, 2009).
We found that diversity, in terms of minority representation in the establishment as compared with a state and industry standard was related to ADA charges for at least some employer groups. While the overall model demonstrated that organizations that receive race charges were more likely to receive disability charges, we still saw that organizations with minority representation greater than the state/industry mean receive more ADA charges. This is an important finding that indicates that even after controlling for the number of race charges, organizations with more minorities receive more ADA charges. This is consistent with the finding of Balser (2000)—racial/ethnic minorities with disabilities are more likely to perceive disability discrimination, as compared with their non-minority peers. It was enlightening to look at federal contractor status to understand the effect of the factors on ADA charges for these specific employer groups. Through these analyses, we see that for federal contractors, having relatively high minority representation was related to a greater ADA charge rate, whereas this was not the case for employers without federal contracts.
Our Other Measures of Diversity Focus on Occupational Segregation by Race and Sex. Among non-federal contractors, higher occupational segregation by sex—that is, women being clustered in certain occupational groups (rather than being distributed evenly across occupational groups)—was associated with more ADA charges. In non-contractor establishments, perhaps, when women are represented more evenly across the occupational ladder, there is a greater perception of organizational fairness among employees of the establishment, which may carry over to limiting the perceptions of disability discrimination. It should be noted that non-federal contractors do not have the same affirmative action requirements as contractors.
These findings suggest that outreach to employers should possibly focus on organizations with relatively high minority representation and should emphasize the importance of considering the intersection of protected groups (e.g., minority employees with disabilities). This is consistent with findings of other scholars who have noted that “disability discrimination is disproportionately directed against women and minorities with disabilities” (Robert & Harlan, 2006, p. 621). Although increasing the diversity of the workplace in terms of disability representation is an important goal, simply increasing diversity without managing it across the employment process may not necessarily lead to positive outcomes, as has been demonstrated in racial and gender diversity studies (Prieto, Phipps, & Osiri, 2011).
Beyond diversity, there are other factors that appear to be related to ADA charges for some employer groups. A higher proportion of managers in the establishment was associated with more charges among non-federal contractors. In these types of establishments, it may be worthwhile to consider the management structure and offer additional training to managers on disability awareness, effective supervision/management, the accommodation process, and conflict resolution. Based on preliminary analyses that demonstrated differences in ADA charge rates by establishment type, we differentiated multi-unit headquarters and multi-unit individual establishments. Although the difference in the ADA charge rate across these two types of establishments was substantially reduced after controlling for employer-, state-, and county-specific characteristics in the regression analysis, the difference remained statistically significant in the overall model. In our sub-analysis, only among non-federal contractors did headquarters have a significantly higher rate of ADA charges. This may reflect greater rights awareness among more highly educated or experienced employees at headquarters, or may be an artifact of the data whereby the charging party may identify the headquarters rather than their unit of employment.
Conclusion
In summary, despite the passage of the ADA more than 20 years ago, people with disabilities are still more likely to be underemployed and unemployed, at least in part as a result of discrimination in the workplace (e.g., Kessler/NOD, 2010; Robert & Harlan, 2006). An aging workforce and policies aimed at increasing disability employment, such as the recently published revised regulations for Section 503 of the Rehabilitation Act, have the potential to greatly increase the number of people with disabilities in the workplace. However, it is not clear that the majority of employers are prepared to successfully integrate qualified workers with disabilities into the workplace. Research has suggested that organizational mechanisms that encourage and tolerate discrimination in the workplace can make disability discrimination “a patterned feature of the workplace” (p. 621) rather than just an occasional isolated interpersonal incidence (Robert & Harlan, 2006). Disability discrimination seems to be disproportionally experienced by women and minorities (Balser, 2000; Robert & Harlan, 2006), and the prevalence of charges filed jointly under bases of Title VII of the Civil Rights Act and the ADA further supports this finding. It is essential that employers are supported in educating their workforces and creating an inclusive environment for all workers, including individuals with disabilities, if the goal of reducing perceived discrimination is to be achieved.
Although this study identifies certain employer characteristics associated with higher rates of ADA charges, we cannot ascertain from these data how the culture, practices, and policies within these establishments are associated with perceived discrimination and bias in the workplace. Establishments similar on certain characteristics, for example, size or industry, vary widely, and further research is needed to understand how systematic behaviors and practices are related to the perception of discrimination. Previous research has highlighted practices and policies that can lead to greater inclusion of people with disabilities and other protected groups, including implementing diversity/disability training programs, mentoring, and setting up an office or individual responsible for diversity/inclusion management programs (Kalev, Dobbin, & Kelly, 2006; Lengnick-Hall, Gaunt, & Kulkarni, 2008; Schur, Kruse, & Blanck, 2005). Unfortunately, these sorts of activities may not decrease charges, as greater awareness of rights through training and other mechanisms may lead to more charges (Hirsh & Kmec, 2009). Nevertheless, workplace culture has an important impact on individuals with disabilities. The creation of an inclusive, fair, and open workplace culture for all groups may well decrease perceptions of bias against people with disabilities (Disability Case Study Research Consortium, 2008). Organizations in which employees in general report high levels of fairness and responsiveness are more likely to eliminate the gap between people with disabilities and their non-disabled peers in areas such as job satisfaction and turnover intentions (Schur et al., 2009).
This study provides a snapshot of employer characteristics related to perceived disability discrimination. Using these findings, those who work to enhance the implementation of the ADA can target employers most at risk of receiving charges, including smaller establishments, federal contractors, those in service industries, and those with greater racial diversity. While there is not an easy solution to employment discrimination, educating these employers about ways to effectively promote inclusion in the workplace, including diversity training, offering mentoring and networking opportunities for protected groups, and establishing organizational responsibility for diversity (e.g., affirmative action plans, diversity committees, and diversity staff positions) may be particularly effective.
Limitations and Opportunities for Future Research
While our study results improve the understanding of employer characteristics associated with ADA charges, there are several limitations to the study that should be noted. At this time, the EEO-1 report contains no information about disability within the establishment. Our findings would be strengthened if we could pinpoint the actual prevalence of disability in the establishment. However, unlike race and gender, which are usually apparent, disability is often hidden and not shared with co-workers or employers. It is not trivial to accurately count the number of people with disabilities. However, tracking disability is part of the revised regulations for the implementation of Section 503 of the Rehabilitation Act of 1973, as amended (Affirmative Action and Non-discrimination Obligations of Contractors, 2013). While there are challenges to collecting disability information, the federal government has been collecting information on disability for many years and has had significant success; in a recent report, 11% of individuals employed by the federal government were individuals with disabilities (U.S. Office of Personnel Management, 2012). Incorporating disability into the EEO-1 report would provide invaluable information to understand disability discrimination in the workplace, and help to understand how initiatives either at the organizational, state, or federal level affect the disability representation, building on work that has been done for race and gender (e.g., Stainback & Tomaskovic-Devey, 2012).
We focused only on charges from 2009, a single point in time. A time series extension would provide more information on how changes in macroeconomic conditions over time relate to the ADA charge rate within the establishment; this is a planned next step in our inquiry. This type of analysis will also enable us to understand whether there exists any time dependency in the rate or propensity of ADA charges as time progresses. Furthermore, we would have a better understanding of whether certain structural changes within the establishment affected the likelihood and the number of ADA charges. Another extension would be to compare employers who do and do not receive charges at different points in the employment process. While most charges cite termination as the perceived discriminatory act, it would be informative to examine which employer-related factors are associated with the receipt of charges for hiring, harassment, or other issues. While not a part of this study, the examination of employers who receive charges found to have reasonable cause (only a small portion of charges filed overall) is another area for study.
Another planned extension is to compare the establishment characteristics that are associated with charge receipt under different statutes. While this article highlights characteristics of employers who are receiving ADA charges, these characteristics may be the same or different from characteristics associated with charges under different statutes. The results of this further analysis can inform approaches reaching out to employers related to employment discrimination more generally.
Footnotes
Acknowledgements
The authors would like to acknowledge Ronald Edwards and Georgianna Hawkins of the U.S. Equal Employment Opportunity Commission (EEOC) for assistance in obtaining and working with these data as well as Susanne Bruyère, Sara VanLooy, Melissa Bjelland, and our anonymous reviewers for their insightful comments and suggestions.
Authors’ Note
Sarah von Schrader obtained an Intergovernmental Personnel Act (IPA) position at the U.S. Equal Employment Opportunity Commission (EEOC), affording her access to data from the EEOC’s annual EEO-1 Employer Report and Integrated Mission System, which includes detailed information on every charge the EEOC receives, as well as those dually filed with Fair Employment Practice Agencies (FEPAs). Further questions about the IPA or study methodology should be directed to Sarah von Schrader (Email:
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support for this article was provided by the U.S. Department of Education National Institute on Disability and Rehabilitation Research. This research is part of a larger project titled Employer Practices Related to Employment Outcomes Among Individuals With Disabilities Rehabilitation Research and Training Center (Grant #H133B100017).