Pre-Employment Testing and the ADA The Americans with Disabilities Act (ADA) requires rea- sonable accommodation in the hiring process, and the lan- guage of the ADA includes "appropriate adjustment or modifications of examinations" as a form of accommoda- tion. It is the goal of this article to acquaint employers with the legal requirements and measurement issues that must be considered in using selection tests under the ADA. There are three types of information that must be consid- ered in using selection tests under the ADA. First, em- ployers must be familiar with the legal requirements of the ADA and the rationale behind these requirements. Second, employers should be aware of the specific role of their tests in helping to select qualified employees (in other words, the validity basis for their tests). Third, employers should be familiar with the types of accommodations that are most likely to be effective in preserving the reliability and validity of the tests for people with various disabilities. Legal and Regulatory Requirements Under the ADA, it is discriminatory to use selection criteria that screen out or tend to screen out individuals with dis- abilities unless the criteria are shown to be job-related for the position in question and are consistent with business necessity. This is to ensure that tests do not act as barri- ers to the employment of persons with disabilities unless the person is unable to do the job, even with reasonable accommodation. Employers should design selection criteria for jobs to en- sure a close fit between the selection criteria and an indi- vidual's ability to do the job. A criterion that tends to screen out an individual with a disability must be shown to be job-related for the position and consistent with business necessity. To be consistent with business necessity a cri- terion must be related to an essential job function. The obligation to make reasonable accommodation means that an employer must make modifications or adjustments to the application process that would enable a qualified indi- vidual with a disability to be considered for the position he or she desires. Tests should not be given in formats that require use of the impaired skill, unless it is a job-related skill that the test is intended to measure. For example, it is unlawful to give a written test to a person who is unable to read because of dyslexia, unless the ability to read is the job-related skill that the test is designed to measure. If, instead, the test is designed to measure a factor such as verbal comprehen- sion or reasoning, the test should be given orally. Simi- larly, test time limits should be relaxed for applicants whose disabilities cause them to need more time to take a test, unless the test is specifically designed to test speed. However, the results of a test of speed could not be used to exclude an individual with a disability unless the skill was necessary to perform an essential function of the posi- tion that could not be performed by the individual with or without reasonable accommodation. If speed is necessary to perform an essential job function and there is no reason- able accommodation available to enable the applicant to demonstrate the skill or to perform the job, then the em- ployer is not required to employ the individual. An employer is obligated to make reasonable accommoda- tion only to the physical or mental limitations that result from the disability of a qualified individual with a disability, that are known to the employer. While an employer may inquire whether an employee is having difficulty perform- ing his or her job, it is generally the responsibility of the employee to inform the employer that accommodation for a disability is needed. Similarly, an employer is generally required to provide test- ing accommodations only if it knows in advance that an applicant has a disability that requires such accommoda- tions. Usually, it is the responsibility of the individual with a disability to request any accommodation for a test. The employer can be helpful by informing applicants in advance about any tests to be administered as part of the application process, so that they may request an accommodation, if needed. Employers may also ask applicants whether they will need an accommodation for the hiring process. It should be noted that the ADA and the Title I regulations prohibit pre-employment inquiry into a person's disability or the nature of the disability, with one narrow exception. The ADA permits employers to ask individuals with a hid- den disability who request accommodations at the applica- tion stage to provide reasonable documentation to verify the disability and the need for accommodation. However, the employer may not make further inquiries as to the na- ture or severity of the disability. For this reason, employ- ers can use psychological tests at the pre-offer state of the hiring process only if these tests are not medical-that is to say, only if the tests do not provide evidence that would lead to identifying a mental disorder. Test Validity Models The use of tests for making employment decisions is sup- ported by evidence of the tests' validity under one or more of the following models: * Content validity: the test is a representative sample of performance in some defined area of job-related knowledge, skill, ability, or other characteristic. * Construct validity: the test is demonstrated to be a measure of a job-relevant characteristic (e.g., reason- ing ability). * Criterion-related validity: the test is shown to be statis- tically related to some criterion of successful job per- formance. Test accommodations should be made so as to retain the validity of the test for selecting qualified employees. Types of Testing Accommodations Testing accommodations will be discussed under three broad categories: testing formats, time limits, and test con- tent. Testing format. A change in testing format refers to the use of a different medium or method to present the same information. Test information is usually presented in print in the English language. Therefore, Braille, large print, reader, and audiotape are simply different ways of present- ing the same information. In most cases, these formats could be interchanged without a change in the question content or the ability being tested. However, several prob- lem areas exist in the use of different formats: 1. Long reading passages may be more difficult when presented orally or in other formats for visually im- paired applicants. For oral presentation, the test-taker must try to keep the entire passage in memory. In Braille or large print, scanning through the passage is slower than it is with regular print. 2. Figural material is problematic for people with visual impairments. The embossing of figural material should not be viewed as a simple format change, because the tactile sense is quite different from the visual sense. 3. When readers are used, they should be people who read well and articulate clearly, and they should prac- tice reading the test in advance. They should be warned against inadvertently giving clues to the test- taker when they read. It should be noted that changing a test from a printed ver- sion into a sign language version is a translation into an- other language, rather than simply a change of format. It must be done with all of the care that would be taken in translating a test from English into, say, Japanese. Time limits. In most cases of accommodated testing it is necessary to change the test's time limits. Often the change in time limits causes a problem in interpreting test results. This problem arises because of the use of "speeded power" tests. In order to understand this prob- lem, it is necessary to learn a bit of testing terminology. A pure power test is a test in which everyone has an op- portunity to attempt to answer every question, and the scores are based on how many questions people can an- swer rather than on how fast they can work. The pure speed test, on the other hand, contains questions of trivial difficulty given with a very short time limit. Scores are based only on how fast people can work. Many tests that are intended to be power tests are actually somewhat speeded because a considerable number of people are un- able to attempt every question. On a speeded power test, a person who had unlimited time would have an advantage over people who took it with the regular time limit. How- ever, since many people with disabilities, e.g., Braille users, need extra time to take tests, there is the difficult problem of determining exactly how much extra time should be allotted so that the test-taker with a disability is at neither an advantage nor a disadvantage. The ideal solution to this problem would be to eliminate the use of speeded power tests. If a power test has a liberal time limit, with a completion rate of, say, ninety-five per- cent of all test-takers, then test-takers with disabilities can be given unlimited time without having an undue advan- tage. The difficult question of how much extra time to allow would no longer need to be answered. In the case of existing speed power tests in which the regular time limits cannot be changed, unlimited time may be inappropriate. One method of determining appropriate time limits is to conduct empirical studies. The U.S. Of- fice of Personnel Management conducted a study to set time limits for visually impaired and deaf applicants on one of its large volume examinations. It was found that at least double time was needed for visually impaired users of all media to answer questions that consisted of a short reading passage followed by five answer choices. Mathematical questions involving computation required considerably more time than that. Such empirical studies are only pos- sible in large-scale programs in which there are many test- takers. Pure speed tests are used in the employment context to test such skills as perceptual speed and clerical checking. Such tests are clearly inappropriate for use with visually impaired test-takers because all of the media for transmit- ting information are slower, and for some physically im- paired applicants, because the physical mechanism for responding (e.g., marking the answer sheet) is slower. The time limit cannot be adjusted on these tests because speed is the factor that is being tested. Therefore, the test user must decide if the speed test should be used with the regular time limit (which is permissible, as noted in the earlier section on legal and regulatory requirements, if the speed factor is necessary to perform an essential job func- tion and the applicant cannot meet the requirement even with reasonable accommodation) or if the test should be deleted from the battery and possibly be replaced with an- other type of assessment. This would be an instance of the last type of test accommodation-change of test con- tent. Test content. In the context of competitive testing for persons with disabilities, changes in test content are not made frequently. However, it is clear that this type of change is a form of accommodation that may be required for compliance with the ADA. Any change in test content would need to be consistent with the validity strategy on which the test was based. For example, substituting one test question for another is easily done under a construct validity model, but might be troublesome under a content validity model. Changes in test content can be divided for convenience into three types: change in individual test questions, change in the question-type, and change or deletion of a knowl- edge, skill, or ability (KSA) that is being measured. The first type of change, as mentioned above, is easily done in a construct-valid test. The second type of change-using a different type of question to test the same ability-is fea- sible if another question-type exists and if scoring compa- rability can be determined. The interpretive guidance to the EEOC's Title I ADA regu- lations describes some bold substitutions of methods for measuring the same KSA's, as the following excerpt shows: "Where it is not possible to test in an alternative format, the employer may be required, as a reasonable accommo- dation, to evaluate the skill to be tested in another manner (e.g., through an interview, or through education, license, or work experience requirements)." This excerpt does not reflect a concern for score compa- rability. In fact, it is difficult to see how this approach could be used if applicants needed to be rank-ordered on the basis of quantitative scores. If there is no effective way to test a person with a disabil- ity for a certain KSA, and if there is reason to believe that this KSA will not be required on the job by the person, the requirement for measuring this KSA should be changed or deleted. Accommodations for Specific Disabilities The following is a brief listing of the types of testing ac- commodations that are appropriate for test-takers with different disabilities. For test-takers with visual impairments, tests must be pre- sented in appropriate formats, such as braille, large print, and audiotape. Time limits must be extended for all of these media, and speed tests are inappropriate. Within the context of changing test materials into different formats, certain types of test material may be problematic, as noted earlier. In addition, the test-taker will probably need ac- commodation or assistance in marking answers. For test-takers who have physical impairments that affect use of the hands, the principal test accommodation is the adjustment of test time limits and the avoidance of speed tests. In addition, accessible test sites, the assistance of a test administrator in turning pages and marking answers, and extra rest breaks may be required. Among hearing impaired test-takers, only those who are deaf need extensive testing accommodations. For the ma- jority of prelingually deaf persons, that is, persons who lost their hearing before acquiring speech, verbal tests are not good measures of any ability. For most pre-lingual deaf people, English is a second language and the native lan- guage is sign language. (Of course, there are exceptions to this rule; some prelingually deaf people have very good English skills.) Therefore, as a general rule, verbal tests cannot be used effectively with most deaf test-takers to test anything except verbal ability. Tests that are com- pletely nonverbal, however, do not pose a problem. Test instructions should be given very carefully, with the use of sign language or demonstration, and time limits should be explained clearly. Extra time should be allowed on power tests that include verbal material. Individuals with specific learning disabilities now constitute the largest group that requires testing accommodations. The specific tasks that are affected by learning disabilities vary widely, so it is difficult to generalize about testing accommodations. Accommodations will need to be ar- ranged on a case-by-case basis for applicants with specific learning disabilities. The most frequently used accommo- dations are the allowance of additional time for power tests and reconsideration of speed tests in areas of specific weakness. For example, a test- taker who had a specific learning disability that affected numerical computation might be screened out by a speeded test of computation. Under the ADA, it would be inappropriate to use that test unless it tested an essential job function that the test-taker could not perform with or without reasonable accommoda- tion. Resources ADA Regional Disability and Business Technical As- sistance Center Hotline (800) 949-4232 (voice/TTY) Equal Employment Opportunity Commission, 1801 L Street, NW, Washington, DC 20507, (800) 669-4000 (Voice) to reach EEOC field offices; for publications call (800) 800-3302 or (800) 669-EEOC (voice/TTY). This brochure is one of a series on human resources practices and workplace accommodations for persons with disabilities edited by Susanne M. Bruyère, Ph.D., CRC, SPHR, Director, Program on Employment and Disability, School of Industrial and Labor Relations - Extension Division, Cornell Univer- sity. It was updated in July, 2000, by Mary Anne Nester, Ph.D., U.S. Immigration and Naturalization Service, Washing- ton D.C., from the original, which she wrote in 1997. Cornell University was funded in the early 1990's by the U.S. Department of Education National Institute on Disability and Rehabilitation Research as a National Materials Develop- ment Project on the employment provisions (Title I) of the ADA (Grant #H133D10155). These updates, and the devel- opment of new brochures, have been funded by Cornell's Program on Employment and Disability, the Pacific Disabil- ity and Business Technical Assistance Center, and other sup- porters. Cornell University currently serves as the Northeast Disabil- ity and Business Technical Assistance Center. Cornell is also conducting employment policy and practices research, exam- ining private and federal sector employer responses to dis- ability civil rights legislation. This research has been funded by the U.S. Department of Education National Institute on Disability and Rehabilitation Research (Grant #H133A70005) and the Presidential Task Force on Employ- ment of Adults with Disabilities. The full text of this brochure, and others in this series, can be found at: www.ilr.cornell.edu/ped/ada. Research reports relating to employment practices and policies on disability civil rights legislation, are available at: www.ilr.cornell.edu/ped/surveyresults.html For further information, contact the Program on Employment and Disability, Cornell University, 102 ILR Extension, Ithaca, New York 14853-3901; 607/255-2906 (Voice), 607/255-2891 (TTY), or 607/255-2763 (Fax). More information is also available from the ADA Technical Assistance Program and Regional Disability and Business Technical Assistance Centers, (800) 949-4232 (voice/TTY), www.adata.org Disclaimer This material was produced by the Program on Employment and Disability, School of Industrial and Labor Relations- Extension Division, Cornell University, and funded by a grant from the National Institute on Disability and Rehabilitation and Rehabilitation Research (grant #H133D10155). The U.S. Equal Employment Opportunity Commission has reviewed it for accu- racy. However, opinions about the Americans with Disabilities Act (ADA) expressed in this material are those of the author, and do not necessarily reflect the viewpoint of the Equal Em- ployment Opportunity Commission or the publisher. The Commission's interpretations of the ADA are reflected in its ADA regulations (29 CFR Part 1630), Technical Assistance Manual for Title I of the Act, and EEOC Enforcement Guidance. Cornell University is authorized by the National Institute on Disability and Rehabilitation Research (NIDRR) to provide information, materials, and technical assistance to individuals and entities that are covered by the Americans with Disabilities Act (ADA). However, you should be aware that NIDRR is not responsible for enforcement of the ADA. The information, materials, and/or technical assistance are intended solely as informal guidance, and are neither a determination of your legal rights or responsibilities under the Act, nor binding on any agency with enforcement responsibility under the ADA. The Equal Employment Opportunity Commission has issued enforcement guidance which provides additional clarification of various elements of the Title I provisions under the ADA. Copies of the guidance documents are available for viewing and downloading from the EEOC web site at: http://www.eeoc.gov . Pre-Employment Testing and The ADA July, 2000 CORNELL UNIVERSITY