Potential Limitations of Computer-Assisted Testing

There are numerous benefits of computer-assisted testing. They can enhance test administration, scoring, interpretation, and integration. Test administration and scoring may be enhanced due to the standardization that is built in to computers. Another benefit is that each test taker receives the same presentation of test items and response sets. The availability of computerized testing devices allows people with a disabilities to complete tests with minimal assistance. This allows the test results to be more valid since there is less enteraction between takers and givers. Test scoring can also be simplified and enhanced due to reduced computational errors.

Test interpretation may be enhanced by providing the counselor with an expanded and consistent knowledge base to assist in the interpretation of test data. Computer-based test interpretation (CBTI) is typically based on research data and clinical experience. Roid and Gorsuch (1984) described four approaches to CBTI: 1) descriptive interpretations; 2) clinician-modeled interpretations (renowned clinician type); 3) clinician-modeled interpretations (statistical model type); and 4) clinical actuarial interpretations. Counselors can use CBTI to support or challenge their judgments about the nature of client problems and potentially effective intervention strategies.

Test integration may be enhanced by including computer-assisted instruction as part of CAT. Clients can be better prepared to use their test results by being more aware of basic concepts and the general nature of their scores. Relieved of presenting repetitive test interpretation information, counselors have more time to explore clients' perceptions of their test data and the implications of the test data for behavior change. The computer can be used to deliver both text-based and interactive video-based instruction (Sampson, 1990a).

POTENTIAL LIMITATIONS OF COMPUTER-ASSISTED TESTING

Computer-assisted testing can limit, as well as enhance, test administration and interpretation. Although paper-and-pencil and computer administration of tests often produce equivalent results, variations in results have sometimes been found to exist. French (1986) recommended that the equivalency of results from different types of administration modes needs to be established for each instrument. Establishing equivalency will reduce the likelihood that computer administration is influencing the