The United States Equal Employment Opportunity Commission (EEOC), together with state and local Fair Employment Practice Agencies, receives about 30,000 charges a year citing employment discrimination based on disability. Although the Supreme Court continues to shape the meaning of the Americans With Disabilities Act (ADA) in highly publicized decisions (1,2,3,4,5,6), policy and procedural changes at the EEOC may have an even greater effect on the experiences of people who file claims under Title I of the ADA, which contains the law's employment provisions.
In June 1995 the EEOC instituted a new three-tiered policy for processing charges to reduce its backlog and to more effectively fulfill its mission of redressing employment discrimination (7). Previously, EEOC policy required full investigation of all charges. Under the new policy, charges that meet the criteria of EEOC national or local enforcement plans and those for which an initial review detects a high probability that discrimination occurred are classified as category A and fully investigated. Charges in which evidence of discrimination is not compelling or in which the charging party or the employer is not covered under one of the laws enforced by the EEOC are classified as category C and slated for immediate dismissal. Charges for which additional evidence is needed to determine whether discrimination occurred are classified as category B. With category B cases, the EEOC routinely sends the employer a letter asking for its side of the story, but the claims of the two parties are verified by further independent investigation only as resources permit (7). Field office investigators categorize cases during or shortly after intake on the basis of the initial evidence provided by the charging party and their general impressions of the charge.
The charge priority policy has helped reduce the EEOC's backlog of charges (8). By prioritizing charges, however, the EEOC has made a deliberate decision to allocate the bulk of its resources to a small proportion of cases, and the selection of these cases is based to a considerable extent on the complaining party's credibility and the evidence the complaining party can provide during an intake interview. This paper focuses on how persons with psychiatric disabilities fare under this system, compared with persons who have other disabilities.
If EEOC field investigators lack experience and understanding of psychiatric disabilities, they may fail to perceive claimants with psychiatric disabilities as credible. Within a procedural framework that asks investigators to make quick judgments about the credibility of claims, including an assessment of whether the claimant is a "qualified individual with a disability," we hypothesized that persons with psychiatric disabilities may fare poorly under the charge priority policy.
In previous articles we presented findings about the outcomes of ADA employment discrimination charges, comparing benefits received by individuals with psychiatric and other disabilities (9,10,11,12,13,14). This paper focuses specifically on the charge priority policy and its consequences for persons with psychiatric disabilities.
We addressed three research questions in this study. First, are charges that receive a high priority assignment more likely than other charges to have outcomes that benefit claimants? Second, what factors are associated with charges that receive a high priority assignment? Third, are persons with psychiatric disabilities as likely as persons with other disabilities to have their charges assigned high priority? We addressed these questions by analyzing EEOC charge data from July 1, 1995, when the charge priority policy began to be implemented, through March 31, 1998.
Data were obtained from the EEOC's computerized charge data system. We analyzed data on all charges that were filed, categorized, and closed between July 1, 1995, when the charge priority policy was implemented, and March 31, 1998 (N=66,298). For our analysis of the effects of psychiatric disability we generated a random sample of individuals who made ADA claims that were prioritized and closed during that period (N=18,858).
The EEOC database includes claimant demographic information, the statute (or statutes) under which the charge was filed, issues alleged in the charge, the claimant's self-reported disability (or disabilities), the EEOC office handling the charge, the dates associated with receiving and closing the charge, the charge's priority assignment, and the final charge outcome (15).
Demographic variables in the database include the charging party's gender, race, ethnicity, and age. The EEOC database provides no information about income, so we included a variable specifying the average household income in the claimant's zip code. Because household incomes within zip codes are not homogeneous, we used this proxy as a reflection of economic opportunities or neighborhood effects that may influence the charging party. Because of skewness, the natural log of zip code area household income was used in the regression-based analysis.
To denote charges filed under multiple statutes, we incorporated separate dummy variables for charges filed simultaneously under the ADA and either the Age Discrimination in Employment Act (ADEA) or Title VII of the Civil Rights Act of 1964. Each of the 19 major issues under which charges can be filed were included as dummy variables. A variable denoting charges citing more than one issue was also included.
Cases involving persons with psychiatric disabilities were defined as cases in which the charging party reported at least one psychiatric disability as a basis for discrimination. The EEOC database has categories for four specific psychiatric impairments—anxiety, depression, schizophrenia, and manic-depression—and a fifth category labeled other psychiatric illnesses. We created dummy variables for each of the five categories. In addition, because preliminary analysis showed HIV infection to be a categorization outlier compared with other nonpsychiatric disabilities, we included a dummy variable denoting a self-reported disability of HIV infection. Had we not included this variable, the model would not have been correctly specified. Because claimants may cite more than one statute and may report up to six disabilities and up to eight issues, dummy variables must be interpreted as nonexclusive. In the regression-based analyses, the disability reference category became complainants reporting neither psychiatric disability nor HIV infection.
In bivariate analyses, we included all three priority categories. In the multivariate analysis, we wanted to determine what factors predict whether a charge will be given the highest priority and thus receive a full investigation. Because category B was developed as a holding position until further information was gathered, we dichotomized the dependent variable, with charges classified as category A coded 1 and those classified as category B or C coded 0.
In the multivariate analysis, we used a time variable indicating the period when the charge was initially received as a control variable. We defined the first period as the initial implementation period of the ADA's employment provisions, from July 26, 1992, through June 30, 1995. This period preceded the implementation of the charge priority handling procedures, but it covers charges that were received during that time and then prioritized and closed later. Subsequent periods are divided into six-month blocks, until the final period, which is a three-month block running from January 1 through March 31, 1998. Periods were coded from 1 to 7, in chronological order. To account for the nonlinear relationship between time and assignment of priority category, we used the squared value of the time variable in the multivariate analysis.
Three of the six major categories of charge resolutions benefit the charging parties. "Withdrawal with benefits" and settlements are agreements between charging parties and employers that resolve charges before the EEOC has finished its investigation and determined the merits of the charge. In this outcome, the charges are resolved by the parties independently without the formal involvement of the EEOC. Settlements, by contrast, involve the EEOC as a party. Conciliation agreements are settlements achieved after the EEOC has established reasonable cause to believe that discrimination occurred (12).
Three other resolutions bring no direct benefits to charging parties. An outcome of "no cause" is an EEOC determination that there is not reasonable cause to believe that discrimination has occurred. An "unsuccessful conciliation" is a failure to obtain a settlement after an EEOC investigation established reasonable cause to believe that discrimination occurred. An "administrative closure" is a charge closed for other reasons without a cause determination (12).
t1 presents summary statistics on the variables used in the model.
We used SAS 6.12 for Windows (16) for all statistical calculations. In bivariate analyses, we used difference-of-proportion z tests to test for significant differences in outcome at each level of prioritization between cases involving persons with psychiatric impairments and cases involving persons with other impairments.
In our multivariate analysis, we used the generalized estimating equations (GEE) procedure to analyze priority assignment. The GEE procedure is a simple approach to accounting for within-cluster correlation in regression settings (17). We used it to better estimate the direct effects of our factors on priority assignment while controlling for clustering within EEOC offices because of the large number of claimants in each office. Within-office correlation in this study may derive from several sources—common procedures employed in offices, a similar economic climate, and drawing from a pool of claimants with similar demographic characteristics. We used the GEE-independent ordinary logistic regression method (PROC GENMOD), which employs a simple correction of the estimated standard errors to account for within-cluster variation. With GEE, parameter estimates can be interpreted as logistic regression coefficients. Odds ratios were calculated for each independent variable.
t2 details the relationship between priority assignment and charge outcome. Not surprisingly, category A cases had the highest total benefit rate, by a wide margin; category B cases followed, and few category C cases resulted in benefits for claimants. In addition, virtually all cases that reached conciliation, whether successful or unsuccessful, were category A cases. Few category B cases and virtually no category C cases reached conciliation. Because only category A cases tend to receive a full investigation, and because a full investigation tends to be a prerequisite for a reasonable-cause ruling and for the conciliation process to begin, an A priority assignment increases the chances for cases to reach the conciliation phase. Classifying a charge as category A also can influence settlement, such as when an ongoing EEOC investigation has uncovered damaging evidence or signals the employer that the agency regards the case as serious. Likewise, agency inactivity may signal the employer that it faces little risk by declining to settle.
Our previous research showed a modestly lower overall benefit rate for charges involving psychiatric disabilities compared with charges involving other disabilities (9). t2 shows that at each prioritization level, cases involving individuals with psychiatric disabilities had a lower overall benefit rate than cases involving individuals with other disabilities. For categories A and B, significant differences were observed in benefit rates between cases involving psychiatric disabilities (z=2.14, p<.05) and those involving other disabilities (z=2.33, p<.05). Among category A cases involving individuals with nonpsychiatric disabilities, the rate of successful conciliation was significantly higher (9.7 percent) than among category A cases involving individuals with psychiatric disabilities (6.8 percent) (z=2.69, p<.01).
t3 shows that each of the four psychiatric disabilities and the category "other psychiatric illnesses" received a category A assignment at a rate significantly lower than the average rate for cases involving other disabilities. The rate was lowest for charges in which schizophrenia was the claimant's disability.
t4 summarizes the results of the GEE procedure for predicting whether or not a case will receive a category A assignment. Within the GEE procedure, because the EEOC office handling the case has been defined as a class variable, office parameter estimates were not computed.
The analysis controlled for demographic characteristics, issues over which cases were filed, and statutes under which cases were filed as well as the EEOC office location and the period during which the charge was filed. Charges by claimants with the three most prevalent psychiatric impairments—anxiety, depression, and manic-depression—were significantly less likely than charges by claimants with other disabilities to receive a category A assignment. Although charges by claimants with schizophrenia had the lowest rate of category A assignment, the small number of cases may have prevented the relationship from reaching significance in the analysis. The category "other psychiatric illnesses" was not statistically significant in the model.
Claimants with HIV infection were more likely than others to receive a category A assignment, and the relationship was highly significant. We are currently exploring why cases involving HIV infection often receive a category A assignment and result in benefits for claimants. The most likely explanation, we think, is that HIV claimants benefit from a long line of court cases recognizing the disease as a disability covered by the ADA (6), whereas coverage of many other disabilities is still being vigorously disputed in litigation (2,3,4).
Among the demographic variables, only the claimants' age was statistically significant. Age was negatively associated with the likelihood of receiving a category A assignment.
Filing simultaneously under the ADEA or Title VII in addition to the ADA was negatively associated with receiving a high priority assignment, and the relationship was statistically significant. Whether filing under multiple statutes actually harms one's chances of receiving a category A assignment in the ADA process is unclear. Although association does not indicate causation, we suspect that filing under more than one statute reflects uncertainty about the basis of discrimination. This uncertainty may reduce the charging party's credibility and thereby decrease the likelihood that the charge will receive a category A assignment.
t4 also provides data about issues under which charges are filed. Charges involving discharge, harassment, and wages were less likely than others to receive a category A assignment, and the relationship was significant in each case. Discharge cases, along with many harassment and wage dispute cases, involve claimants no longer employed. When individuals are no longer working for the employer, settlements are often more difficult to achieve. It is likely that many of these individuals will find new jobs and stop pursuing their discrimination case.
Hiring and insurance were the only issues over which cases were filed that were both significantly and positively related to receiving a category A assignment. The relationship between failure to hire and higher prioritization is particularly noteworthy given a recent statement by the National Council on Disability that "the EEOC has not engaged in any proactive strategies to address discrimination in the hiring process, a problem that charge processing does not address well" (18). Although filing under multiple statutes is less likely to result in high-priority assignment, the presence of multiple issues in a given case is positively associated with receiving a category A assignment.
The time control variable displays two significant trends, suggesting that, all else being equal, in the initial period of implemention of the charge priority policy, cases were more likely to have been classified as category A, whereas more recent cases were less likely to have received a category A classification. However, this trend may reflect the disproportionate number of category C cases in more recent periods; category A cases take longer to close, and a larger proportion of category A cases from the recent periods are still open.
In an effort to more effectively combat workplace discrimination, the EEOC has taken major steps to streamline the administrative charge process. The charge priority policy implemented in 1995 is the most recent attempt to reduce the EEOC's charge backlog and help workers achieve a quick resolution of their complaints.
We found a highly significant relationship between the assignment of high priority and an outcome with benefits for the claimant. We also found that benefit rates for charges assigned the two top priority levels involving persons with psychiatric disabilities were significantly lower than rates for charges involving other types of complainants. When other case characteristics were controlled for, charges involving anxiety, depression, and manic-depression had a significantly lower likelihood of receiving a category A classification. Although differences in rates of priority assignment between cases involving psychiatric and nonpsychiatric disabilities were modest, the fact that category A charges have the best chance by far of receiving a full investigation makes any difference important.
Because we could not evaluate the merits of each charge, we suggest several alternative explanations of why cases involving psychiatric issues may receive fewer category A assignments and outcomes with benefits for claimants. These findings may reflect real differences in the merits of the two types of cases. Alternatively, if we assume no baseline differences in the merits or complexity of psychiatric versus other cases, our findings suggest unequal application of the 1995 charge handling policy. Given widespread misunderstanding of psychiatric illness, the possibility that stereotypes or stigma lead investigators to assign fewer psychiatric cases the highest priority cannot be ruled out. Inadequately trained staff may not understand how psychiatric impairments affect individual functioning in the workplace.
The assigning of priority level occurs early in the process—during or immediately after intake—and it determines whether an investigation occurs. The charge priority policy thus places a premium on a complainant's ability to articulate a clear case of discrimination to the investigator. Many individuals with psychiatric impairments have great difficulty articulating their thoughts and presenting their case. If one function of the EEOC is to help victims of discrimination receive fair hearings on their claims, our results raise the concern that the charge priority policy may fail some of the complainants who are most in need of such assistance.
The EEOC is to be commended for issuing guidelines in 1997 that focus specifically on charges involving people with psychiatric disabilities (19). However, we find cause for concern about how the charge priority policy affects certain ADA claimants, because the EEOC has never validated how accurately its staff prioritizes charges. Given the differences found in this study, the EEOC should broaden its efforts to train investigators to assist individuals with psychiatric disabilities in fully articulating the merits of their case. Furthermore, the disparity in benefit rates of cases involving individuals with psychiatric disabilities, even after priority assignment has been controlled for, suggests the need for ongoing monitoring of the way psychiatric cases are investigated and resolved.
This study was funded by grant R01-MH-57077 from the National Institute of Mental Health. The authors are indebted to the U.S. Equal Employment Opportunity Commission for arranging an Intergovernmental Personnel Act agreement allowing the project's principal investigator, Dr. Moss, access to the EEOC's confidential charge data system. The authors also thank Leo Sanchez for assistance with the EEOC data.
Mr. Ullman is a research associate at the Cecil G. Sheps Center for Health Services Research at the University of North Carolina at Chapel Hill, 725 Airport Road, Chapel Hill, North Carolina 27514 (e-mail, firstname.lastname@example.org). Dr. Johnsen is associate professor at the Center for Mental Health Services Research at the University of Massachusetts Medical School in Worcester. Dr. Moss is a senior research fellow at the Jordan Institute for Families in the School of Social Work at the University of North Carolina at Chapel Hill and a research fellow at the Sheps Center. Mr. Burris is professor of law at Temple University School of Law in Philadelphia and senior associate at the Johns Hopkins School of Public Health in Baltimore.
Demographic and case characteristics of 66,298 complainants filing Americans With Disabilities Act Title I charges with the US. Equal Employment Opportunity Commission
Charge outcome of psychiatric and other disability claims, by priority category, for all EEOC cases categorized and closed between July 1, 1995, and March 31, 1998, as percentages1
1 Category A, high probability that discrimination occurred; category B, additional evidence of discrimination needed; category C, evidence not compelling
Distribution of priority categories, by disability, for all EEOC charges categorized and closed between July 1, 1995, and March 31, 1998
Generalized estimating equations model predicting category A priority assignment for all EEOC cases closed between July 26, 1992, and March 31, 1998 (N=18,858)