Computerized vs. written administration of the MMPI-A in clinical and non-clinical settings

The purpose of the present study was to see if comparisons of the computer and written administrations of the MMPI-A justify continuing to treat them as equivalent forms of the same test. Subjects for the present study included a total of eighty adolescents between the ages 14 to 18 years old. Using forty subjects from a clinical setting and forty from a non-clinical setting with equal numbers of males and females, the differences between the computer and written versions of the MMPI-A were examined using a test-retest design. Each subject was administered both versions of the MMPI-A, approximately one week apart, with the order of administration counter-balanced. After the second administration, the subjects were asked to compare the two forms of administration in eight areas of interest to investigate experiential differences and subject preference.;The present results suggest that the computer administration yields equivalent results to the traditional written administration for a general population of subjects taken from both clinical and non-clinical settings. For Validity, Clinical, Content, and Supplemental MMPI-A scales, the test-retest correlation coefficients were statistically significant and compared favorably to reliability data provided in the MMPI-A manual. Comparisons between T-scores for the computer and written scale scores revealed no significant difference and supported the use of identical normative data for both forms of administration. The only psychometric difference between the two forms was in administration time, which averaged twenty-eight minutes faster for the computer format. The subjective measures indicated a significant preference for the computer version on seven out of eight questionnaire items. The substantial time difference as well as the personal preference displayed by the sample suggests significant experiential differences between the two forms of administration. However, without significant psychometric differences between formats, the use of the identical normative data seems justified.