7.5 pages body of supporting, cited information, Current issues and future trends related to the topic. Either way, this element should be addressed. Discuss your personal interest and takeaways regarding the topic and research (not the same as the closin

Work 47 (2014) 267–279 267 DOI 10.3233/WOR-121583 IOS Press

Allegation of ethnic minorities from 1993–2008: An equal employment opportunity commission (EEOC) study

Juan Carlos Arango-Lasprillaa,∗, Jessica M. Ketchumb, Jessica Hurleyc, Almaz. M. Getachewc and Kelli Williams Garyd aIKERBASQUE, Basque Foundation for Science, University of Deusto, Bilbao, Spain bDepartment of Biostatistics and Center for Rehabilitation Science and Engineering (CERSE), Virginia Commonwealth University, Richmond, VA, USA cDepartment of Rehabilitation Counseling, Virginia Commonwealth University, Richmond, VA, USA dDepartment of Occupational Therapy, Virginia Commonwealth University, Richmond, VA, USA

Received 29 March 2012

Accepted 25 September 2012

Abstract. BACKGROUND: Approximately 25% of working-aged Americans with disabilities work full or part time, yet still face dis- crimination despite the passing of the American’s with Disabilities Act (ADA) over 20 years ago. OBJECTIVES: To determine if the proportion of allegations of ADA Title I workplace discrimination with merit closed at any year between 1993 and 2008 differs among Whites, African Americans, Hispanics, and Asians; to determine if there was a change over time from 1993 to 2008 in merit closure rate within each race/ethnicity group; and to determine whether changes over time between 1993 and 2008 in the merit closure rate differ among the race/ethnicity groups. METHODS: Logistic regression was used for this cross-sectional panel study to model the merit closure rate for each ethnic group from 1993 to 2008 using 318,587 charging parties from the EEOC database. RESULTS: All ethnic groups exhibited significant changes over time in the merit closure rate. There were significant differences in the closure rates among the race/ethnicity groups specifically at closure years 1995–2000, 2002, 2003, and 2006. Finally, there was evidence that the trends in merit closure rates over time differed significantly among the race/ethnicity groups. CONCLUSIONS: There was significant evidence that the proportion of claims closed with merit was significantly different among the racial/ethnicity groups.

Keywords: Disability, employment, discrimination, minorities

1. Introduction

There are around 54 million Americans with disabil- ities, or approximately 10% of the population and, a

∗Corresponding author: Juan Carlos Arango-Lasprilla, Ikerbas- que Research, University of Deusto, Ikerbasque, Basque Foundation for Science, Avda de las Universidades 24, Bilbao 48007, Spain. Tel.: +34 94 413 90 03, ext 3261; E-mail: [email protected].

staggering 72% of working age Americans are not in the work force and 21% live below the poverty line [1]. Of all vulnerable groups in America, people with dis- abilities (PWD) are still the poorest [2]. The high un- employment rates of PWDs also have a major impact on their physical and psychological wellbeing. In addi- tion, employment has been identified as a vessel for fi- nancial independence [3], empowerment [4], and posi- tive self-worth [5]. Therefore, work is critical for inte- gration into society for people with disabilities [6].

1051-9815/14/$27.50 c© 2014 – IOS Press and the authors. All rights reserved

268 J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study

One of the identified causes as to why PWDs expe- rience employment discrimination is the disability it- self; however, other factors such as race and/or eth- nicity are also thought to have a major impact on em- ployability [9]. Persons from racial or ethnic minori- ties with disabilities experience double the discrimina- tion; researchers have identified this trend as “double jeopardy” when disability status is also combined with racial or ethnic minority status [10]. It is projected that in the United States racial and ethnic minorities will reach 45% of the total population by the year 2050 [1]. Furthermore, it is also expected that the proportion of people living with disabilities will also increase. Given the expected increase for both racial and ethnic mi- norities as well as people with disabilities, it is impor- tant to investigate circumstances surrounding this dou- ble jeopardy status in relation to disability employment discrimination.

Twenty years after the passing of the Americans with Disabilities Act (ADA), significant economic dis- parities still exist between PWDs and those without disabilities [11]. In 2008, the American Community Survey indicated that only 39.5% of PWDs were em- ployed compared to 79.9% of people without disabil- ities [12]. Although Title I of the ADA prohibits em- ployment discrimination against people with disabil- ities, PWDs continue to experience employment dis- crimination for multiple reasons [11]. Studies involv- ing ADA Title I claims of employment discrimination indicate that the most common charges have been un- lawful termination, failure to provide reasonable ac- commodation, discrimination in the terms/conditions of employment, disability harassment, and discrimina- tion in the hiring process. Employers have mistakenly pointed to the assumed high cost of accommodations on the job [13] and lower productivity due to functional limitations [14] as reasons for not employing people with disabilities. Even after overcoming discrimination and when PWDs actually do manage to become em- ployed and maintain employment, there are still large income gaps compared to people without disabilities. For example, the annual median income for PWDs is $39,600 compared to $61,200 for people without dis- abilities [15]. Furthermore, while Title I of the ADA has improved employee outcomes for PWDs; employ- ment discrimination continues to play a major role.

Today, according to our review of the literature, over 80 studies have investigated the characteristics of em- ployees (or charging parties) and employers involved in ADA Title I allegations of employment discrimi- nation based on disability as well as each allegation’s

merit status determination as found by the Equal Op- portunity Employment Commission (EEOC). These studies were conducted using the EEOC’s ADA Title I database which contains all Title I allegations filed for the lifetime of the ADA from 1993 to 2008 [16–37].

The majority of these EEOC studies have explored race/ethnicity allegations in the context of a larger al- legation type (e.g., disability type such as back impair- ment, multiple sclerosis, mental retardation, deafness, or depression; discrimination issue such as wrongful firing, disability harassment, or failure to provide a reasonable accommodation; or employer industry such as health care, manufacturing, or education). How- ever, there have only been two studies completed so far which have focused solely on race/ethnicity allega- tions. Lewis et al. [38] and Hurley et al. [39] exam- ined charging party and employer characteristics, re- spectively, between Title I allegations filed by Whites and African Americans. Lewis et al. [38] reported the following: Whites filed Title I allegations under more impairments than did African Americans; African Americans more often filed Title I allegations under chronic diseases such as asthma, diabetes, HIV/AIDS, and other blood disorders more than Whites; African Americans more often filed Title I allegations under more of a variety of employment discrimination issues than did Whites; African Americans more often filed Title I allegations under disciplinary issues such as terms and conditions of employment, suspension, dis- cipline, harassment, promotion, and assignment than did whites; and African Americans were typically 35 to 54 years old when they filed Title I allegations while Whites typically were much older when they filed. Hurley et al. [39] reported: African Americans and Whites filed Title I allegations under an equal number of employer industries; African Americans filed Title I allegations under more paraprofessional and profes- sional industries, while Whites filed more Title I alle- gations under more blue collar industries; Whites filed more Title I allegations under more small- to medium- sized employers, while African Americans filed more Title I allegations under more large to extra-large em- ployers; African Americans filed Title I allegations in the south while Whites filed more Title I allegations in the West, Northeast, and Midwest. Although these studies have added much to the race and employment discrimination literature, there is still a paucity of such literature. Therefore, there is a need for further explo- ration of race and disability employment discrimina- tion.

The present study is a retrospective cross-sectional panel design and seeks to investigate all ADA Title I

J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study 269

allegations made and closed with merit by African Americans, Asian Americans, Hispanics, Native Amer- icans and Whites across the lifetime of the ADA (1992 to 2008). Investigating these changes over time will provide answers to emerging issues which may af- fect certain racial and ethnic minority groups and em- ployment discrimination. More specifically, this study will address the following three aims: (1) Does the proportion of ADA Title I workplace discrimination allegations closed with merit differ between Whites, African Americans, Hispanics, and Asians at specific time points between 1993 and 2008; (2) does the pro- portion of ADA Title I workplace discrimination alle- gations closed with merit change over time from 1992 to 2008 for each race/ethnic group; and (3) Do the changes over time between 1993 and 2008 in the pro- portion of ADA Title I allegations of workplace dis- crimination closed with merit differ among Whites, African Americans, Hispanics, and Asians?

2. Methods

2.1. Participants

The equal employment opportunity commission (EEOC) has a database, called the integrated mission system that holds allegations filed under all legislation in its jurisdiction, including the Americans with dis- abilities act (ADA). From this database, all ADA Ti- tle I allegations were placed into a separate database by researchers at Virginia Commonwealth University for use by the national EEOC, ADA research project which is housed at Virginia Commonwealth University (VCU). An interagency agreement between the EEOC and VCU allowed these data to be studied under the condition that they be de-identified.

Within the project database, the person who files the allegation of disability employment discrimination under Title I of the ADA (or the ADA’s employment provisions) is called the “charging party” and the em- ployer against whom the allegation is filed is called the “respondent”. In addition, the unit of measure within the project database is an allegation of employment discrimination and not the person with a person with a disability who filed it under Title I of the ADA. the project database contains 402,291 ADA Title 1 workplace allegations with the following categories: 44 charging party disability types (e.g., cancer, back impairment, cerebral palsy, alcoholism, or depression; charging party gender (i.e., m,f, or null); charging party

age (i.,e., 15–87 years with a mode of 44 years); charg- ing party race (i.e., white, African American, His- panic/Mexican, Asian, native American/Alaskan na- tive, mixed, other, and null); charging party discrim- ination issue (e.g., 41 types including hiring, firing, demotion, harassment, or reasonable accommodation); employer or respondent industry (e.g., 21 types based on the north American industry classification sys- tem including agriculture, construction, retail trades, transportation/warehousing, construction, and profes- sions/scientific/technical); employer or respondent size (i.e., 15–100 employees, 101–200 employees, 201– 500 employees, 501+ employees, or null); employer or respondent region (i.e., based on us census regions west, northeast, Midwest, south, foreign and territo- ries, or null); and allegation resolution type (i.e., closed by the EEOC with merit which means it determined discrimination to have occurred or closed by the EEOC without merit which means it determined no discrim- ination to have occurred. all ADA Title I allegations filed in acts of retaliation or that were the result of filing or duplication errors were excluded from the project database.

Because the current study seeks to explore the con- nection between race, disability, and employment dis- crimination under ADA Title I, data from each of the project database categories were evaluated descrip- tively as to its usefulness in shedding light on this focus. Due to a relatively high rate of nulls in the database category, employer region, it was eliminated.

In this sample, 69.4% (n = 221,191) of the charg- ing parties were White, 22.4% (n = 71,385) were African American, 6.9% (n = 21,840) were Hispanic, and 1.3% (n = 4,171) were Asian. The characteris- tics of the sample are summarized in Tables 1 and 2, overall and by race/ethnicity status. The percentage of claims was evenly distributed among male (51.7%) and female (48.3%) charging parties with the aver- age age of the sample 43.7 years (SD = 10.15). Then majority of charging party issues were classified as discharge (31.6%), other (28.7%), or reasonable ac- commodations (18.0%). Employer size was predomi- nantly small with 15 to 100 workers (32.2%) or large with more than 500 workers (44.6%) and employers tended to be located in the Southeast (23.7%), Mid- west (23.7%), or Southwest (16.6%). Most employer industry types were service producing (75%) with over 60% in manufacturing (20.7%), education and health services (19.1%), or trade, transportation, and utili- ties (19.0%). Most cases were closed without merit (77.0%), regardless of race/ethnicity.

270 J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study

Table 1 Description of nominal characteristics of the EEOC sample by charging party race/ethnicity

White African American Hispanic Asian Overall

Count Percent Count Percent Count Percent Count Percent Count Percent

Charging party sex Male 116114 52.58 33920 47.58 12289 56.32 2165 52.03 164488 51.71 Female 104703 47.42 37365 52.42 9530 43.68 1996 47.97 153594 48.29 (Missing) (374) (100) (21) (10) (505)

Charging party issue Discharge 70223 31.75 22962 32.17 6441 29.49 1101 26.40 100727 31.62 Reasonable accommodations 41015 18.54 11704 16.40 3743 17.13 779 18.68 57241 17.97 Terms/conditions 17894 8.09 6809 9.54 1910 8.75 441 10.57 27054 8.49 Harassment 16339 7.39 5874 8.23 2197 10.06 380 9.11 24790 7.78 Hiring 13136 5.94 3016 4.22 1034 4.73 228 5.47 17414 5.47 Other 62584 28.29 21020 29.45 6515 29.83 1242 29.78 91361 28.68

Employer size 15–100 workers 69695 32.91 20778 30.23 7021 33.35 1009 24.99 98503 32.24 101–200 workers 26031 12.29 8078 11.75 2388 11.34 389 9.64 36886 12.07 201–500 workers 23458 11.08 8135 11.84 1950 9.26 407 10.08 33950 11.11 501 or more workers 92565 43.71 31737 46.18 9696 46.05 2232 55.29 136230 44.58 (Missing) (9442) (2657) (785) (134) (13018)

Employer region New England 2315 1.05 281 0.39 90 0.41 43 1.03 2729 0.86 Northeast 9027 4.08 3132 4.39 1747 8.00 162 3.88 14068 4.42 Mid-Atlantic 23200 10.49 7891 11.05 399 1.83 214 5.13 31704 9.95 Southeast 51206 23.15 21801 30.54 2075 9.50 409 9.81 75491 23.70 Midwest 55660 25.16 17101 23.96 2108 9.65 654 15.68 75523 23.71 Southwest 32466 14.68 11516 16.13 8633 39.53 406 9.73 53021 16.64 Great plains 9384 4.24 2572 3.60 305 1.40 85 2.04 12346 3.88 Rocky mountain 7756 3.51 1021 1.43 1354 6.20 140 3.36 10271 3.22 Pacific 25160 11.37 5338 7.48 4895 22.41 1900 45.55 37293 11.71 Northwest 5002 2.26 722 1.01 233 1.07 158 3.79 6115 1.92 Foreign/military 15 0.01 10 0.01 1 < 0.01 0 0.00 26 0.01

Employer industry type Natural resources and mining 2957 1.55 770 1.24 462 2.49 65 1.82 4254 1.55 Construction 4652 2.44 1070 1.72 439 2.37 54 1.52 6215 2.26 Manufacturing 40406 21.18 12307 19.82 3560 19.19 721 20.24 56994 20.72 Trade, transportation, and utilities 36787 19.28 11013 17.73 3478 18.75 566 15.89 51844 18.85 Information 8885 4.66 3085 4.97 745 4.02 148 4.15 12863 4.68 Financial activities 11204 5.87 3473 5.59 839 4.52 210 5.89 15726 5.72 Professional and business services 16331 8.56 5654 9.10 1503 8.10 276 7.75 23767 10.64 Education and health services 35978 18.86 12406 19.98 3254 17.54 765 21.47 52403 19.05 Leisure and hospitality 6659 3.49 2039 3.28 530 2.86 128 3.59 9356 3.40 Public administration 19299 10.12 7437 11.98 2680 14.45 420 11.79 29836 10.85 Other services 7631 4.00 2850 4.59 1061 5.72 210 5.89 11752 4.27 (Missing) (30399) (9281) (3289) (608) (43577)

Employer industry type II Goods producing 48015 25.17 14147 22.78 4461 24.05 840 23.58 67463 24.53 Service producing 142777 74.83 47957 77.22 14090 75.95 2723 76.42 207547 75.47 (Missing) (30399) (9281) (3289) (608) (43577)

Merit closure Yes 52896 23.91 15154 21.23 4234 19.39 910 21.82 73194 22.97 No 168295 76.09 56231 78.77 17606 80.61 3261 78.18 245393 77.03

2.2. Measures

The primary outcome variable of interest was a di- chotomous variable indicating whether or not the ADA workplace discrimination claim by the charging party

was closed with merit. Demographic characteristics in- cluded in the database regarding the charging party in- clude age (in years), gender, and race/ethnicity. Char- acteristics regarding the claim filed include issue of claim and the year the claim was closed. Characteris-

J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study 271

Table 2 Description of age of the sample by charging party race/ethnicity

Charging party age (years) N Mean SD Median IQR Range Missing

White 206688 44.23 10.29 44.00 37 to 51 15 to 79 14503 African American 65372 42.32 9.46 42.00 36 to 49 16 to 77 6013 Hispanic 20108 42.97 10.19 43.00 36 to 50 16 to 75 1732 Asian 3873 45.50 10.72 45.00 38 to 53 18 to 75 289 Overall 296041 43.74 10.15 44.00 37 to 51 15 to 79 22546

SD = Standard Deviation.

tics regarding the employer include size of company, Department of Energy (DOE) region, and North Amer- ican Industry Classification System (NAICS) industry type. The primary predictor variables of interest were race/ethnicity and year the case was closed.

Age was analyzed as a continuous variable, while gender and race/ethnicity were modeled as nominal variables. Issue of claim represents the reason the charging party was filing a claim and was catego- rized into 6 groups: Discharge, Reasonable Accommo- dations, Terms/Conditions, Harassment, Hiring, and Other. This categorization was chosen to represent the 5 most frequently occurring issues, with all other is- sues grouped together. Employer size was as a nomi- nal variable with 4 levels: 15 to 100 employees, 101 to 200 employees, 201 to 500 employees, and more than 500 employees. Employer NAICS industry type was re-categorized into 11 groups using the NAICS Al- ternate Aggregation Structure (see http://www.census. gov/eos/www/naics/history/docs/cm_2.pdf). Closure year was modeled as a nominal variable.

2.3. Statistical analyses

The characteristics of the sample were described overall and separately for the groups of claims with and without merit using standard summary measures (counts/percentages for nominal variables and means/ standard deviations [SD] or medians/interquartile ran- ges [IQRs] for continuous variables. A single logistic regression model was used to model the proportion of claims closed with merit as a function of race/ethnicity, year of closure, and the race/ethnicity by year inter- action effect. In addition, the model adjusted for the aforementioned charging party, claim, and company characteristics. This model was used to address all of the aims for this study.

For aim 1, 14 overall comparisons were tested using 3 degree of freedom chi-square tests within the logistic regression model to determine if the merit closure rate differed between the race/ethnicity groups at each year between 1993 and 2008, except for 2004 and 2005. A

Bonferroni adjustment of α = 0.05/14 = 0.0036 was used to control for multiple comparisons and maintain an overall Type I error rate of α = 0.05. For years where the overall 3 degree of freedom chi-square test comparing the race/ethnicity groups was significant (p- value < 0.0036) 6 pairwise comparisons of the merit closure rate among the race/ethnicity groups were fur- ther tested using single degree of freedom chi-square tests within the logistic regression model. A Bonfer- roni adjustment of α = 0.05/6 = 0.0083 was used to control the Type I error rate. Odds ratios and 95% con- fidence intervals were used to compare the odds of a claim closed with merit versus a claim closed with no merit among the four racial/ethnic groups.

For aim 2, four overall comparisons were conducted using 13 degree of freedom chi-square tests within the logistic regression model to determine if there were changes in the merit closure rate over time for each of the racial/ethnicity groups. A Bonferroni adjustment of α = 0.05/4 = 0.0125. If the overall hypothesis test was significant for any race/ethnicity group then the trends were described over time within the specific race/ethnicity group by quantifying the changes in the odds of closure rate between different years. The gen- eral trend in the unadjusted merit closure rate will be used to generate 4 or 5 time segments to more easily describe the trend.

For aim 3, the race/ethnicity by year interaction ef- fect in the logistic regression model was examined at a significance level of α = 0.05. A significant interaction effect would be indicative that the changes in the merit closure rate over time differed significantly among the race/ethnicity groups.

3. Results

3.1. Preliminary analysis

Initially, unadjusted preliminary analyses were con- ducted to assess the changes merit closure rates over time for each race/ethnicity group, without controlling

272 J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study

Table 3 Unadjusted merit closure rates across charging party race/ethnicity by time

White African American Hispanic Asian Overall

Year Count Percent Count Percent Count Percent Count Percent Count Percent

1993 1477 32.16 417 31.52 76 23.10 9 20.00 1979 31.46 1994 2489 25.18 625 22.45 206 21.82 28 22.22 3348 24.37 1995 2783 16.46 711 14.60 272 17.46 51 18.81 3817 16.17 1996 3269 14.67 722 11.03 322 14.70 42 11.60 4355 13.88 1997 3679 16.11 941 13.84 276 13.24 53 12.93 4949 15.40 1998 4458 18.99 1073 14.51 314 14.46 113 27.70 5958 17.81 1999 5406 24.99 1493 18.58 445 19.33 122 22.80 7466 22.97 2000 6544 29.92 1986 25.90 559 22.01 142 31.77 9231 28.38 2001 6330 29.44 2436 31.12 711 26.84 129 26.11 9606 29.58 2002 5904 26.85 1823 23.72 573 19.20 121 24.49 8421 25.40 2003 3905 26.05 1186 23.46 457 23.06 43 12.76 5591 25.00 2006 2102 31.38 667 33.65 11 40.74 17 16.83 2797 31.76 2007 2421 37.26 577 31.44 10 27.78 28 33.33 3036 35.92 2008 2129 35.27 497 31.46 2 4.76 12 21.05 2640 34.22

Fig. 1. Unadjusted trends in merit closure rates over time by race/ ethnicity group. (Colours are visible in the online version of the arti- cle; http://dx.doi.org/10.3233/WOR-121583)

for any other characteristics. Irrespective of year of clo- sure, there was significant evidence that the proportion of claims closed with merit was significantly different among the racial/ethnicity groups (χ23 = 403.0, p < 0.0001; Cramer’s V = 0.0352). As can be seen from the bottom of Table 1, the trend was such that Whites had the highest merit closure rate (24.0%), followed by Asians (22.1%), and African Americans (21.3%), with Hispanics having the lowest merit closure rate (19.4%). The unadjusted merit closure rates by year for each race/ethnicity and overall are plotted in Fig. 1 and summarized in Table 3. In general, there was a decreas- ing trend in merit closure rates from 1993 through 1996 followed by an increasing trend from 1996 through 2001. After 2001, Whites and African Americans ex- hibit slight deceases in merit closure rates through 2003 followed by increases through 2008. Hispanic and Asians had rather different trends after 2001 than Whites and African Americans and showed greater de-

viations from the overall trend over time than the other race/ethnicity groups. Asians had considerably lower merit closure rates than other groups in 1993, 2003, 2006, and 2008 and higher rates in 1998. Hispanics, on the other hand, were notably lower than the other groups in 1993, 2001, 2008, and 2009 and higher in 2006.

3.2. Aim 1

The tests for differences in the closure rates among the race/ethnicity groups at each year are summarized in Table 4. Using a Bonferroni adjusted significance level of α = 0.0036 there were significant differences in the closure rates among the race/ethnicity groups for closure years 1995 to 2000, and 2002, 2003, and 2006. Pairwise comparisons of the odds of closure with merit (versus closure without merit) were tested between the race/ethnicity groups at these significant years using a Bonferroni adjusted significance level α = 0.0083; p- values for these tests can be seen in Table 4. The odds ratios comparing the odds of closure with merit (ver- sus closure without merit) among the race/ethnicity groups that were statistically significant are summa- rized with 95% confidence intervals in Table 5. Odds ratios greater than 1 indicate that the odds of closure with merit are greater for the first group as compared to the second group, while odds ratios smaller than one indicate the odds of closure with merit are smaller for the first group as compared to the second group; confi- dence intervals not including 1 are statistically signif- icant at the 5% level of significance. For example, the odds of closure with merit were 14% to 49% greater for Whites as compared to African Americans during the years of 1996 to 2000.

J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study 273

Table 4 (Aim 1) Global and individual comparisons for differences in merit closure rates among race/ethnicity by time

W vs. AA W vs. H W vs. A AA vs. H AA vs. A H vs. A

Year F * p-value p-value p-value p-value p-value p-value p-value

1993 3.59 0.0130 1994 2.88 0.0347 1995 4.73 0.0027† 0.0754 0.0197 0.0411 0.0018‡ 0.0117 0.3760 1996 14.16 < 0.0001† < 0.0001‡ 0.4290 0.5397 < 0.0001‡ 0.3315 0.3907 1997 6.65 0.0002† 0.0039‡ 0.0022‡ 0.0221 0.2194 0.1308 0.3774 1998 36.83 < 0.0001† < 0.0001‡ < 0.0001‡ < 0.0001‡ 0.0962 < 0.0001‡ < 0.0001‡ 1999 41.27 < 0.0001† < 0.0001‡ < 0.0001‡ 0.2605 0.2452 0.0213 0.1363 2000 24.53 < 0.0001† < 0.0001‡ < 0.0001‡ 0.0510 0.0002‡ 0.0008‡ < 0.0001‡ 2001 2.27 0.0783 2002 18.32 < 0.0001† 0.0446 < 0.0001‡ 0.0012‡ < 0.0001‡ 0.0077‡ 0.7979 2003 10.39 < 0.0001† 0.0104 0.7640 < 0.0001‡ 0.2093 < 0.0001‡ < 0.0001‡ 2006 5.17 0.0014† 0.0368 0.0676 0.0063‡ 0.1228 0.0019‡ 0.0023‡ 2007 4.21 0.0055 2008 1.86 0.1333

*F -statistics have 3, 247×103 degrees of freedom. † Indicates statistically significant using Bonferroni adjustment of α = 0.0036. ‡ Indicates statistically significant using Bonferroni adjustment of α = 0.0083.

Table 5 (Aim 1) Significant odds ratios for merit closure between race/ethnicity groups by time

W vs. AA W vs. H W vs. A AA vs. H AA vs. A H vs. A

Year OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI

1995 0.764 (0.646, 0.904) 1996 1.360 (1.236, 1.496) 0.696 (0.594, 0.814) 1997 1.136 (1.042, 1.238) 1.256 (1.085, 1.453) 1998 1.341 (1.236, 1.454) 1.528 (1.327, 1.761) 0.528 (0.415, 0.672) 0.394 (0.307, 0.505) 0.345 (0.262, 0.454) 1999 1.487 (1.381, 1.601) 1.373 (1.214, 1.553) 2000 1.195 (1.116, 1.280) 1.512 (1.349, 1.695) 1.265 (1.117, 1.433) 0.663 (0.522, 0.842) 0.524 (0.406, 0.677) 2002 1.501 (1.334, 1.690) 1.559 (1.191, 2.042) 1.398 (1.230, 1.588) 1.452 (1.103, 1.910) 2003 2.677 (1.832, 3.911) 2.384 (1.623, 3.502) 2.623 (1.763, 3.902) 2006 2.248 (1.257, 4.022) 2.539 (1.411, 4.570) 5.016 (1.780, 14.136)

W = White; AA = African American; H = Hispanic; A = Asian; OR = Odds Ratio; CI = Confidence Interval.

3.3. Aim 2

There were significant changes in the proportion of cases closed with merit over time for each for the race/ethnicity groups: White F(13, 247×103) 308.34, p < 0.0001; African American F(13, 247×103) = 120.0, p < 0.0001, Hispanic F(13, 247×103) = 19.49, p < 0.0001; and Asian: F(13, 247×103) = 7.98, p < 0.0001. Based on the general trend of the plot, odds ratios were computed comparing the following years: 1996 to 1993, 2001 to 1996, 2003 to 2001, and 2009 to 2003. The estimated odds ratios and 95% confi- dences are summarized by race/ethnicity and overall in Table 6. Odds ratios greater than 1 indicate increases in the odds of merit closure rates over time and odds ratios smaller than 1 are indicative of decreases over time.

Whites, African Americans, and Hispanics all show- ed significant decreases in the odds of closure with merit from 1993 to 1996 (ORs between 0.269 and 0.580; p’s � 0.0010) while Asians showed nominal but non-significant decreases (OR = 0.634; p = 0.2993). All race/ethnicity groups showed significant increases in the odds of closure with merit from 1996 to 2001 (ORs between 2.167 and 3.499; p’s � 0.0003). Whites, African Americans, and Asians all showed significant decreases in the odds of closure with merit from 2001 to 2003 (ORs between 0.395 and 0.862; p’s � 0.0001) while Hispanics showed non-significant decreases (OR = 0.915; p = 0.2806). Whites, African Americans, and Asians all showed significant increases in the odds of closure with merit from 2003 to 2009 (ORs be- tween 1.477 and 2.218; p’s � 0.0376) while Hispanics showed non-significant decreases (OR = 0.780; p = 0.7576).

274 J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study

Table 6 (Aim 2) Changes in the odds ratios for merit closure over time by race/ethnicity

1996 vs. 1993 2001 vs. 1996 2003 vs. 2001 2009 to 2003

OR 95% CI OR 95% CI OR 95% CI OR 95% CI

Whites 0.355 (0.328, 0.385)† 2.481 (2.351, 2.619)† 0.862 (0.816, 0.911)† 1.477 (1.374, 1.589)† African/Americans 0.269 (0.230, 0.315)† 3.499 (3.156, 3.879)† 0.740 (0.672, 0.815)† 1.507 (1.312, 1.731)† Hispanics 0.580 (0.420, 0.802)† 2.167 (1.836, 2.558)† 0.915 (0.778, 1.076) 0.780 (0.161, 3.787) Asians 0.634 (0.269, 1.498) 2.278 (1.466, 3.538)† 0.395 (0.253, 0.615)† 2.218 (1.047, 4.699)† Overall 0.433 (0.343, 0.547)† 2.554 (2.266, 2.888)† 0.693 (0.614, 0.782)† 1.401 (0.903, 2.173) † Indicates statistically significant using Bonferroni adjustment of α = 0.05.

Table 7 (Aim 3) Pairwise comparisons in trends over time between race/eth- nicity groups

1996 vs. 1993 2001 vs. 1996 2003 vs. 2001 2009 to 2003

Year p-value p-value p-value p-value

W vs. AA 0.0020† < 0.0001† 0.0070† 0.8033 W vs. H 0.0039† 0.1279 0.4995 0.4286 W vs. A 0.1882 0.7055 0.0006† 0.2911 AA vs. H < 0.0001† < 0.0001† 0.0278† 0.4157 AA vs. A 0.0545 0.0629 0.0066† 0.3212 H vs. A 0.8493 0.8357 0.0005† 0.2416 † Indicates statistically significant using Bonferroni adjustment of α = 0.05. W = White; AA = African American; H = Hispanic; A = Asian.

3.4. Aim 3

There was evidence that the trends in merit closure rates over time were significantly different across the four race/ethnicity groups (F(39, 247×103) = 8.29, p < 0.0001). The changes over time using the 4 time segments examined in the aim 2 were compared be- tween the race/ethnicity groups, after controlling for charging party, claim, and company characteristics; the p-values from these tests are summarized in Table 7.

From 1993 to 1996, Whites (OR = 0.36) and African Americans (OR = 0.27) showed significantly greater decreases in the odds of merit closure over time as compared to Hispanics (OR = 0.58), and African Americans showed significantly greater decreases in the odds of merit closure as compared to Whites. From 1996 to 2001, African Americans (OR = 3.50) showed significantly greater increases in the odds of merit clo- sure over time as compared to Whites (OR = 2.48) and Hispanics (OR = 2.17). From 2001 to 2003, Asians showed significantly greater decreases in the odds of merit closure over time as compared to Whites (OR = 0.86), African Americans (OR = 0.74), and Hispan- ics (OR = 0.92) and African Americans showed sig- nificantly greater decreases in the odds of merit clo- sure over time as compared to Whites and Hispanics.

Finally, from 2003 to 2009, there were no significant differences in the increases in the odds of merit closure over time between the race/ethnicity groups.

3.5. Other characteristics

Looking at the covariates that were adjusted for in the model, all were significantly associated with merit closure rate: charging party age (p < 0.0001), charging party sex (p = 0.0021), issue (p < 0.0001), employer size (p < 0.0001), employer region (p < 0.0001), and NAICS (p < 0.0001).

4. Discussion

The purpose of the present study was to determine if Whites, African Americans, Hispanics, and Asians exhibited differences in the merit closure rate of alle- gations of ADA Title I workplace discrimination with merit closed between 1993 and 2008. Based on the re- sults of this study we found the following: There were significant differences in merit closure rates among the race/ethnicity groups at closure years 1995–2000, 2002, 2003, and 2006, (2) all race/ethnic groups ex- hibited significant changes over time from 1993–2008, and (3) the changes over time from 1993–2008 were significantly different among the race ethnicity groups. These results highlight discernible differences between the majority and minority subgroups with regard to employment discrimination for people with disabili- ties.

In regard to Aim 1, the current study found sta- tistically significant differences in the proportion of ADA Title I workplace discrimination allegations that were closed with merit at specific years between 1993 and 2008 among Whites, African Americans, Hispan- ics, and Asians. The highest unadjusted merit clo- sure rate was among Whites (24%; Table 1), regard- less of closure year. The potential to have a higher

J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study 275

merit closure rate may be related to a higher amount of filing allegations overall. McMahon, Roessler, Rum- rill et al. [32] highlighted that hiring allegations were most frequently filed by charging parties who identi- fied themselves as White. These results included non- hiring allegations, as well. Likewise, Lewis et al. [38] found that Whites filed more allegations of Title 1 employment discrimination across more impairments compared to any other race/ethnicity group. Therefore, a higher proportion of allegations initially filed can lead to a higher percentage of claims that were closed with merit.

Another salient feature regarding these findings is, regardless of closure year, Asians and African Ameri- cans had the second and third, respectively highest un- adjusted merit closure rate (22% and 21%) and His- panics having the lowest (19%). There are a few plau- sible explanations for these findings. Although evi- dence has shown that Asians Americans are less likely to file employment discrimination allegations [30], the higher proportion of merit closures may be indica- tive of the type of occupational characteristics that are more prevalent among the minority group compared to African Americans and Hispanics. Asian Americans have been perceived as the ‘model minority group’ by many in the US due to their overall educational and ca- reer status despite being a minority in the US [40] and as compare to other minority groups many have ob- tained careers, or as immigrants have entered the US, as physicians, dentists, pharmacists, scientists, and en- gineers representing jobs at the high end of the occu- pational spectrum [41]. Speculating that many of the allegations filed by Asian Americans may come from those in higher profile careers, the likelihood that the case would fare in favor of the charging party is highly probable. Hispanics, however, who had the lowest rate of merit closures, are plagued by lower educational attainment and a higher concentration of lower wage jobs [42]. Although the type of job does not determine whether a merit case is awarded or not, it may influence a lesser amount of allegations filed. Additionally, the language barrier that is more often experienced by His- panics as compared to African Americans can poten- tially lead to less filing of discrimination allegations; thus, less allegations closed with merit.

When controlling for other predictor variables in the model (Tables 4 and 5) differences among the groups were noted as significant at specific years, but not all years. While lack of a difference count be indicative of no difference, it could also be due to low power. No- tably, for the years where differences were identified,

Whites consistently had significantly higher closure rates than African Americans (1996–2000), Hispanics (1997–2002), and Asians (2002–2006), but did have significantly lower rates than Asians in 1998. In addi- tion, African Americans had significantly lower merit closure rates than Asians in 1995 and 1996, but signif- icantly higher rates in 2000 and 2002. When compared to Asians, African Americans had significantly lower closure rates in 1998 and 2000, but significantly higher rates in 2002, 2003 and 2006. Finally, when Hispanics were compared to Asians, Hispanics had significantly lower merit closure rates in years 1998 and 2000 but significantly higher rates in 2003 and 2006.

With respect to Aim 2, the current study extended the findings related to the difference in proportion of merit closures for the various racial/ethnic groups a step further to examine the change over time in the pro- portion of ADA Title I allegations closed with merit. As noted, there were significant changes over time for each of the race/ethnicity groups examined. Specifi- cally, there were precipitous changes in the Asian sub- group between the years 1997 and 1998, as well as 2002 and 2006. The first spike was between the years 1997 and 1998; whereas, there was a notable increase in merit closures. From 2002 to 2003 a dramatic de- crease of merit closures was indicated, followed by the most drastic decline in 2006. The instability in the merit closure rates may be attributed to many fac- tors. Most notably are cultural factors related to Asian Americans; whereas, they are less likely to file formal complaints [30]. Therefore, if there was an increase in merit closures by the Asian American group, then the discrimination claims were probably more of a grave nature.

There was a gradual decline of merit closure rates over time for Hispanics, until 1998 when the percent- age started to increase slightly from 1998 to 2001. Af- ter 2001, there was some substantial variability from 2002 and 2008. An extremely notable decrease of merit closures was indicated in 2008. Clearly, something odd transpired during those times. One plausible explana- tion could be the substantial changes in classification of Hispanics on federal documents [43]. This can be problematic when filling out federal forms such as in the U.S. Census or filing an allegation of employment discrimination with the EEOC. For instance, a charg- ing party who is originally from Honduras who also happens to be African American, would be unable to choose between the Hispanic or African American cat- egories. Likewise, the term, “Mixed”, does not apply to this person, as he or she is African American only.

276 J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study

Because the term, “Hispanic”, is applied so broadly to people from Spanish cultures, there may be some un- reliability in charging parties who did or did not iden- tify themselves as, “Hispanic” and based their choice on their own best estimate. Hispanics are also another subgroup of minorities that have been known to be least likely to file allegation complaints similar to the Asian American subpopulation. McMahon et al. [32] suggested Hispanics may have higher expectation of negative outcome regarding the utility of their allega- tions of discrimination in the hiring process. Thus, it can be surmised that Hispanics believe that it is not worth the trouble to jeopardize their job and go through the arduous process of filing a complaint with a higher likelihood of a negative outcome. Moss et al. [44] re- ported that for psychiatric disabilities in particular, be- ing Hispanic significantly reduces the likelihood of re- ceiving a benefit from an employment discrimination case. Specifically, the odds ratios indicate that Hispan- ics were nearly four times less likely to receive a ben- efit than non-Hispanics.

For the African American community, there was a significant decline of merit closures from 1993 to 1996. After 1996, the trend of merit closures for African Americans steadily increased until 2001. Then there was a relatively sharp decline beginning in 2002. There was a sharp rise in rates again in 2006 and rates remained steady until 2008. It is difficult to determine if there were any distinct historical events that may have influenced the trends of allegations closing with merit among African Americans. However, Lewis et al. [38] and Hurley et al. [39] suggested that any de- cline in allegations may be related to African Amer- icans propensity to file under the Civil Rights Act of 1964 due to historical significance as opposed to filing under ADA Title 1. Potentially, the impairment char- acteristics of charging parties who are African Amer- ican may affect the trend of allegations for this mi- nority group. Conyer, Boomer, and McMahon [22] noted that a disproportionate number of African Amer- icans filed HIV/AIDS employment discrimination alle- gations. As time progressed from 1993 through 2008, medical advances allowed more workers to return to work with the diagnoses of HIV/AIDS. This surely cre- ated a ripe environment for employment discrimination allegations to be filed.

Results indicate that merit closure rates for Whites dramatically decreased between 1993 and 1996, and then started to gradually increase until 2008. There were a relatively high percentage of merit closures for Whites in 2005 and rates fluctuated from that year un-

til 2008. The gradual increase for Whites closer to beginning years that allegations were being filed and awarded can be attributed to the growing knowledge about the law and accessibility to detailed informa- tion about filing. It is well known that Whites dispro- portionately benefitted from filing ADA Title 1 hiring- related allegations [45] and this group was probably some of the first to take advantage of the new law and continued to increasingly file due to more optimistic in the outcome of the process. In addition, a few land- mark court cases that clarified ADA Title 1 law and possibly influenced an influx of filing allegations in the 21st century.

Variability of merit closure rates overall could be influenced due to backlog issues for ADA Title 1 al- legations. Moss et al. [44] noted that the process for many ADA Title 1 allegations resulting in closure of merit or non-merit status could have taken an excep- tionally long period of time due to backlog (but the exact years of backlog are difficult to determine be- cause the dataset collects closure year and not filing year). The fact that backlog exists which could have slowed the charge process for anyone filing and possi- bility lead to meritorious claims of discrimination be- ing dismissed without an adequate review by EEOC personnel. Additionally, any variability in cases may be due to them possibly being dully file under both Ti- tle IV and Title 1 if racial discrimination and disability discrimination both took place, particularly for Asians, Hispanics, and African Americans.

In 1999, a series of three Supreme Court cases called “The Sutton Trilogy” narrowed the ADA definition of disability, thereby making it more difficult for a charg- ing party to be covered under the Act and to then re- ceive a merit resolution [46]. In regard to Aim 3 of the current study, by looking at the combined number of race allegations with a merit resolution for each year (See Table 2), one can see that after 2001, the num- ber of cases with a merit resolution steadily declines. Because case closure backlog exists at the EEOC, Sut- ton’s effects showed up more and more every year after 2001. When the ADA Amendments Act of 2008 went into effect in January of 2009, the narrow ADA def- inition of disability was alleviated and is expected to make the number of cases with a merit resolution revert toward the mean in the future iterations of the EEOC ADA Title I database.

There are similar observations for broad application of parties who have dual identities for race and ethnic- ity. As noted earlier, the fact that the EEOC did not have a classification for charging parties that would

J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study 277

identify as Hispanic ethnicity and African American or White race, can be problematic when filling out federal forms for filing allegations. This could lead to skewed results for those in Hispanic, African American, and White category due to charging parties who chose to select one category or the other based on their choice of what to be identified as.

As Table 2 displays, there is a relatively consistent number of missing race data across groups with the overall percent missing being 7.7%, and within the race groups the percent missing ranged from 7.0% to 9.2%, a very narrow range. The missing data could be from an error in the collection process on the side of the EEOC staff members who entered the data into the database or it could be the result of charging parties potentially being unwilling to reveal their race due to fear of government repercussions (e.g. a person of His- panic origin afraid that he or she will be deported even though he or she has already become a citizen). Of course, those groups with smaller counts (e.g., Asians) could be more greatly affected by even the average or standard amount of missing race data in the EEOC database.

Our research highlights the significant challenges for racial and ethnic minorities encountering employ- ment discrimination. These findings have direct im- plications for policies, practices, and program devel- opment. Apparently, racial and ethnic minorities are prone to underreporting ADA Title I allegations. Hu- man resource departments must issue policies tailored to explicitly describe EEOC discrimination allegation procedures in laymen terms upon hire instead of a brief, insufficient overview at orientation. New hires should know exactly where to report and EEOC of- ficials should be in a visible location and this infor- mation should be updated on an interval basis. Poli- cies should be extremely clear about the requirement to keep allegations anonymous to ameliorate the fear of those most vulnerable of job loss to encourage re- porting of discriminatory actions. Additionally, it has been elucidated that language barriers may limit ac- cess to information, which can also negatively influ- ence reporting of allegations. EEOC regulations must be available in other languages and/or interpreters to provide clarification related to language deficits. Be- yond the aforementioned implications, increase edu- cation continues to be needed to distinguish ADA Ti- tle I discrimination from other types of discrimination to target filing allegations under most accurate legisla- tion.

These results should be interpreted in light of var- ious limitations. The limitations are as follows: (1)

There may be multiple cases filed by the same individ- ual thus all cases may not be independent; there was no way of identifying if there were multiple cases from one single individual. (2) Secondary data could poten- tially have errors from the collection process such as from the EEOC staff members who filled out the paper- work to be entered into the database when the charg- ing party came in to file a complaint or from the charg- ing party who might have felt too afraid of government repercussions to reveal his or her race (e.g. Hispanics who fear somehow being deported even though they have obtained citizenship). This could have contributed to all or some of the missing data and potentially skew results even if these missing cases have been removed. (3) Sample size was small in early years, and espe- cially in later years (likely because not all cases had been closed) and sample size for Hispanics and Asians is much smaller relative to Whites and Hispanics, again especially in the later years. There may be unknown bias as to why there are fewer than expected cases for these race/ethnicity groups. (4) There are certain vari- ables specifically related to race and ethnicity, such as acculturation, income, etc., that could potentially act as confounders. However, because the EEOC database did not collect these variables, they could not be con- trolled for in the analyses. (5) Potentially, there is miss- ing data because all categories that were labeled null for each variable were deleted prior to analyses. Addi- tionally, if people chose to opt out of providing their race to EEOC, then we would have less race-related null data or missing data. (6) Also, with regards to missing or inaccurate data, the fact that the EEOC database relies on self-report data is a substantial lim- itation. EEOC staff members must take charging par- ties at their word as far as how they classify themselves or if they choose not to do so. When a charging party comes in to make a claim, things like disability type are taken at face value from the charging party’s inter- view. They might have several disabilities but only re- port one since they would have to choose at the point in order to file the claim.

The present study provides direction for future re- search. Some research should include comparison and contrast of each race/ethnic group in a study between specific disability types (e.g., back impairment, cere- bral palsy, multiple sclerosis, muscular dystrophy, de- pression, schizophrenia, alcoholism, drug abuse, etc.). Additionally, studies that explore whether race and/or ethnicity would predict merit or non-merit outcome of ADA Title I allegations would potentially reveal inter- esting findings. Although it has been mentioned that

278 J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study

the EEOC did not identify individuals that have mul- tiple claims, research that could examine differences among racial and ethnic groups who file multiple al- legations may lead to further evidence related to merit and non-merit closure outcomes for those groups. Since the ADA have experienced dramatic changes due to the ADA amendment of 2008, cross-sectional panel outcome data among racial and ethnic groups that be- gins with time periods after 2008 are needed. Quali- tative and mixed methodology should also be consid- ered to explore the unreported side of disability em- ployment discrimination and, more specifically, how it is affected by race/ethnicity.

5. Conclusions

In conclusion, this study responds to future rec- ommendations cited in Lewis et al. [38] that call for follow-up of the influence of race on the merit out- come of ADA Title I employment discrimination alle- gations and investigating the same with more than one racial or ethnic group. This study substantially con- tributes to emerging research that explores distinct dif- ferences in ways employment discrimination is experi- enced by minority subgroups of the disability popula- tion. Given the paucity of previous literature, this study will further elucidate the importance of publishing em- pirical evidence about the difference and trends over time of ADA Title I employment discrimination alle- gations, particularly as it relates to merit closure in a heterogeneous sample that clearly represents a diverse population in the United States.

Acknowledgments

This work was supported by the National Insti- tute of Disability and Rehabilitation Research and the United States Department of Education (grant H133A080060).

References

[1] Rehabilitation research and training center on disability demographics and statistics. [Disability Statistics Report]. Ithaca: Cornell University; 2008.

[2] Risher P, Amorosi S. The 1998 NOD/Harris Survey of Ameri- cans with Disabilities. Washington, DC: Louis Harris and As- sociates Author; 1998.

[3] Partch-Davies T, Rivera J. Everday Heros: How Taxpayers with Significant Disabilities are Building Assets. Building a Better Economic Future. 2008; 45.

[4] Hillborg H, Svensson T, Danermark B. Expectations, visions and sense of empowerment: in the face of a vocational rehabil- itation process for people with psychiatric disabilities. Scandi- navian Journal of Disability Research. 2010; 12(2): 109-124.

[5] Burge P, Ouellette-Kuntz H, Lysaght R. Public views on em- ployment of people with intellectual disabilities. Journal of Vocational Rehabilitation. 2007; 26 (1): 29-37.

[6] Smits S. Disability and employment in the USA: the quest for best practices. Disability & Society. 2004; 19 (6): 647-662.

[7] Kreider B, Pepper JV. Disability and employment. Journal of the American Statistical Association. 2007; 102 (478): 432- 441.

[8] Andelic N, Stevens LF, Sigurdardottir S, Arango-Lasprilla JC, Roe C. Associations between disability and employment 1 year after traumatic brain injury in a working age population. Brain Injury. 2012; 26(3): 261-269.

[9] Ficke RC, Digest of data on persons with disabilities. [Statis- tical Data Report]. Washington, DC: National Institution on Disability Rehabilitation Research (ED/OSERS). 1992.

[10] Alston RJ, Bell TJ. Cultural mistrust and the rehabilitation enigma for African Americans. J Rehabil. 1996; 62(2): 11.

[11] Acemoglu D, Angrist JD. Consequences of employment pro- tection? The case of the Americans with Disabilities Act. J Polit Econ. 2001; 109(2): 1.

[12] Turner J, Bourdreax M, Lynch V. A preliminary evaluation of health insurance coverage in the 2008 American Community Survey. [Report]. US Census Bureau; 2009.

[13] Kemp Jr EJ. Disability in our society. In: Weaver CL, edi- tor. Disability and work: Incentives, rights and opportunities. Washington, DC: AEI Press; 1991. p. 55.

[14] LaPlante MP. State estimates of disability in America: Dis- ability statistics report 3. [Statistical Data Report]. Washing- ton, DC: National Institution on Disability Rehabilitation Re- search (ED/OSERS). 1993.

[15] Erickson W, Lee C, von Schrader S. Disability statistics from the 2008 American Community Survey (ACS) and Current Population Survey (CPS). Ithaca, NY: Cornell University Re- habilitation Research and Training Center on Disability De- mographics and Statistics (StatsRRTC); 2010. Retrieved from www.disabilitystatistics.org.

[16] Unger DD, Campbell LR, McMahon BT. Workplace discrim- ination and mental retardation: The nutritional EEOC ADA research project. J Vocat Rehabil. 2005; 23(3): 145.

[17] Lowman DJ, West SL, McMahon BT. Workplace discrimina- tion and Americans with cerebral palsy: The national EEOC ADA research project. J Vocat Rehabil. 2005; 23(3): 171.

[18] Mitchell PR, McMahon BT, McKee D. Speech impairment and workplace discrimination: The national EEOC ADA re- search project. J Vocat Rehabil. 2005; 23(3): 163.

[19] Lewis AN, McMahon BT, West SL, Armstrong AJ, Belongia L. Workplace discrimination and asthma: The national EEOC ADA research project. J Vocat Rehabil. 2005; 23(3): 189.

[20] Bowe FG, McMahon BT, Chang T, Louvie, I. Workplace dis- crimination, deafness and hearing impairment: The National EEOC ADA Research Project. Work. 2005; 25(1): 19.

[21] West SL, McMahon BT, Monasterio E, Belongia L, Kramer K. Workplace discrimination and missing limbs: The National EEOC ADA Research Project. Work. 2005; 25(1): 27.

[22] Conyers L, Boomer KB, McMahon BT. Workplace discrim- ination and HIV/AIDS: The national EEOC ADA research project. Work. 2005; 25(1): 37.

[23] Armstrong AJ, McMahon BT, West SL, Lewis AN. Work- place discrimination and cumulative trauma disorders: The

J.C. Arango-Lasprilla et al. / Allegation of ethnic minorities from 1993–2008: An EEOC study 279

National EEOC ADA Research Project. Work. 2005; 25(1): 49.

[24] McMahon BT, Edwards R, Rumrill P, Hursh N. An overview of the National EEOC ADA Research Project, Phase One. Work. 2005; 25(1): 1.

[25] McKenna M, Fabian E, Hurley JE, McMahon BT, West SL. Workplace discrimination and cancer: The National EEOC ADA Research Project. Work. 2007; 31(8): 8.

[26] Neath J, Roessler R, McMahon B, Rumrill P. Patterns in per- ceived employment discrimination for adults with multiple sclerosis. Work. 2007; 29(3): 255.

[27] McMahon BT, West SL, Shaw LR, Waid-Ebbs K, Belongia L. Workplace discrimination and traumatic brain injury: The National EEOC ADA Research Project. Work. 2005; 25(1): 67.

[28] West MD, Dye AN, McMahon BT. Epilepsy and work- place discrimination: Population characteristics and trends. Epilepsy Behav. 2006; 9(1): 101.

[29] Roessler RT, Neath J, McMahon BT, Rumrill PD. Workplace discrimination outcomes and their predictive factors for adults with multiple sclerosis. Rehabil Couns Bull. 2007; 50(3): 139.

[30] Tartaglia A, McMahon BT, West SL, Belognia L, Beach LS. Workplace discrimination and healthcare: The national EEOC ADA research project. J Vocat Rehabil. 2007; 27(3): 163.

[31] McMahon BT, Hurley JE. Discrimination in hiring under the Americans with Disabilities Act: An overview of the national EEOC ADA research project. J Occup Rehabil. 2008; 18(2): 103.

[32] McMahon BT, Roessler R, Rumrill Jr PD, Hurley JE, West SL, Chan F, Carlson L. Hiring discrimination against people with disabilities under the ADA: Characteristics of charging parties. J Occup Rehabil. 2008; 18(2): 122.

[33] West SL. Reasonable accommodation and employment dis- crimination: An introduction to the special issue. Rehabil Pro- fess. 2008; 16(4): 187.

[34] West SL, Roessler R, Rumrill PD, McMahon BT, Hurley JE, Carlson L, Chan F. Employer characteristics and reasonable accommodation discrimination against people with disabili- ties under the ADA. Rehabil Profess. 2008; 16(4): 209.

[35] Van Wieren T, Reid C, McMahon BT. Workplace discrimina- tion and autism related disorders: The National EEOC ADA Research Project. Work. 2009; 31(3): 299.

[36] Roessler RT, Hurley JE, McMahon BT. A comparison of al- legations and resolutions involving issues of discharge versus

constructive discharge. Adv Dev Hum Resour. 2010; 12(4): 407.

[37] Hurley JE. Merit determinants of ADA Title I allegations in- volving discharge: Implications for human resources manage- ment and development. Adv Dev Hum Resour. 2010; 12(4): 466.

[38] Lewis AN, Hurley JE, Armstrong AJ, Koch LC, Gary KW, McMahon BT. The influence of race on allegations of em- ployment discrimination due to disability. J Min Disabil Res Prac. 2010; 1(3): 30.

[39] Hurley JE, Lewis AN, Koch LC, Armstrong AJ, Gary KW, McMahon BT. An examination of the relationship between the race of ADA Title 1 claimants and attributes of their em- ployers. J Min Disabil Res Prac. 2010; 1(3): 119.

[40] Rim KH. Model, victim, or problem minority? Examining the socially constructed identities of Asian-Origin ethnic groups in California’s media. Asian American Policy Review. 2007; 16: 37.

[41] Xie Y, Goyette K. Social mobility and the educational choices of Asian Americans. Social Science Research. 2003; 32(3): 467-498.

[42] langHirschman C, Massey DS. Places and peoples: The new American mosaic. In D.S. Massey New faces in new places: The changing geography of American immigration. 2008: 1- 22. NY: Russell Sage Foundation.

[43] Office of Management and Budget [Webpage on the Inter- net]. Washington, DC: Revisions to the Standards for the Classification of Federal Data on Race and Ethnicity, Fed- eral Register Notice; 1997 [updated 1997 October 30; cited 2011 November 30]. Available from: http://www.whitehouse. gov/omb/fedreg_1997standards.

[44] Moss K, Burris S, Ullman M, Johnsen S, Swanson J. Un- funded mandate: An empirical study of the implementation of the Americans with Disabilities Act by the Equal Employ- ment Opportunities Commission. University of Kansas Law Review. 2001; 50(1): 1.

[45] McMahon BT. Hurley JE. Discrimination in hiring under the Americans with Disabilities Act: An overview of the National EEOC ADA Research Project. J Occupat Rehabil. 2008; 18(2): 103.

[46] McMahon BT, West SL, Hurley JE. Who is a person with a disability under the ADA? Mitigating circumstances, the U.S. Supreme Court and the case of diabetes. J Vocat Rehabil. 2006; 24(3): 177.

Copyright of Work is the property of IOS Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.