Practitioner Guidelines for Psychometric Testing

Arrangements for testing people with disabilities or impairments

Considerable care and expertise is needed when the mode of administration of a test has to be changed to accommodate the needs of people with disabilities. As always, local and national law and practice* needs to be considered, and the individual’s rights to privacy must be respected. In seeking information regarding types and levels of disability, inquiries should only seek information relating to each person’s ability to undertake the activities required to complete the test. Particular care needs to be exercised in relation to employment testing**.

There is no simple rule of thumb that can be used to ensure that a test is administered fairly for people with all types of disability. It is a matter of professional judgement as to whether it is better to use some alternative form of assessment, or to modify the test or its mode of administration. In practice, it is rarely possible to norm modified tests on sufficient samples of people with equivalent disability in order to ensure comparability of the test with the standardised version. However, where data exist on, for example, the effects of changing time limits, use of Braille or audiotape spoken versions of tests, such data should guide the user in making the necessary accommodations. While full standardization of a modified version may not be possible, pilot testing on small samples of individuals should be carried out whenever practical.

Given the dearth of information about the performance of people with disabilities on tests (whether modified or not), it is often more appropriate for test result to be used in a more qualitative manner. They can be used to give an indication of the characteristic being assessed (ability, motivation, personality, etc.), which can be supplemented and supported by information gathered using other methods.

For individual assessment, the assessor can usually tailor the assessment procedures to the capabilities of the person who is being assessed. However, particular issues arise in group testing (e.g. for selection into employment). Here there may be practical difficulties involved in varying the mode of administration for particular individuals within a group administration setting. Furthermore, all parties may see differences in treatment as being unfair. For example if more time is given for test completion, those with the disability may be conscious that they are being treated ‘differently’, and those without the disability may feel that the extra time provides an unfair advantage.

Advice on special needs can usually be obtained from relevant disability organisations as well as the individual test takers. It is generally helpful (where the law permits) to ask the individual directly in a non-threatening and supportive way if there are any considerations that need to be taken into account***. In many cases such consultation will enable suitable modifications to be made to the test taking environment without requiring changes to the test itself.

The following outline protocol provides a general guide to the process of deciding whether to modify testing and how to carry out the modification. Essentially, disability may contribute no variance to test scores, contribute construct relevant variance or construct irrelevant variance. In the first case, no modifications are necessary. In the final case, modifications should be aimed at removing the irrelevant source of variance (by suitable modification of the test conditions or substitution of a more suitable test). For the second case (construct relevant variance), however, modification to the test will affect the relevance of the test scores.

1. Is the disability likely to have an effect upon test performance? Many people have disabilities that would not affect test performance. In such cases, it would be inappropriate to make accommodations for them.
   
2. If the disability is likely to affect test performance, then is the effect on performance incidental to the construct being measured? For example, a person with an arthritic hand may have trouble with a speeded test which involves writing. If the ability to perform manual tasks rapidly is part of the construct being measured, then the test should not be changed. However, if the purpose is to assess visual checking speed, then an alternative mode of response would be appropriate.
   
3. When the particular disability is incidental to the construct being measured but is likely to affect the individual’s performance on the test, then modification of the procedure may be considered.
   
4. Users should always consult the test manual and the publisher for guidance on modification and for information regarded alternative formats and procedures.
   
5. Users should also consult relevant disability organisations for advice and guidance on the possible implications of a specific disability, relevant literature or documentation, and the sort of adaptations or accommodations that may prove helpful.
   
6. Any modifications made to the test or test administration procedures should be carefully documented along with the rationale behind the modification.

The essential point here is that Ciba must not treat people with disabilities less favourably than others and this applies not only to recruitment and selection but also to training, promotion and redundancies. Any assessment technique used must, as far as is reasonably possible, be free of any requirement that places a disabled person at a substantial disadvantage, unless that requirement can be justified.

In terms of justifying a requirement, it is not enough simply to prove that the requirement is essential for the job; Ciba must also have considered whether the person with disabilities could meet the job-related criterion if a ‘reasonable adjustment’ were made either to the selection process or to the job itself.

Summary

Psychometric tests and questionnaires are a potential source of adverse impact if they are not used with proper care. It is the responsibility of every test-user to ensure that the tests selected for use are relevant to the assessment issue and sufficiently robust psychometrically to ensure that they discriminate between people only on the basis of the attributes that are pertinent to the assessment.

All test-takers are unique individuals who will have various concerns and special needs relating to the testing process. It is the responsibility of the test-user to give test-takers every opportunity to voice concerns or give information about their special needs. This is best done by asking the test taker, prior to the test session, to communicate any concerns or special needs. Where the test-taker is unsure about how to address the special needs, the individual concerned should generally be consulted. Where the test-taker is unsure as to how any adjustments might impact on the interpretation of the test, the test publisher should be consulted or other suitable qualified professional, such as a business psychologist. The priority in any assessment context is to give every individual test-taker equal opportunity to demonstrate the attributes that are relevant to the assessment issue.