FAKABILITY AND VALIDITY INDICATOR IN THE MMPI

Abstract

The Japanese edition of MMPI wasadministered twice to a group of 214 Japanese cadets. The first administration was under the usual instructional set. Two weeks later, the Ss were divided into three subgroups consisting of approximately an equal number of Ss which differed according totest-taking instructional set. Two experimental groups (favorable set group and unfavorable set group) and one control group (neutral set group) were utilized.
The introduction of favorable or unfavorable set produced significant changes in the MMPI scores of testees. The K correction has not been found sufficiently valuable to in dicate the faking bad profiles.
In order to test the efficiency of validity indicators, it has been planned to make a comparison of the validating scores obtained from the three different set groups. The results reported in the present study indicated that the three keys (L, F, K), considered singly, are able to identify faking sets to some extent and that Hs raw score, utilizedsingly, was more successful in discriminating the faking good from the normal scores, but theirmaximum efficacy is realized in combination.
The results of this study confirmed Gough'sresults that the F-K index has a superior value to detect a normal person who has tried to fake. In the present data, the F-K cutting score for the faking bad profiles was plus seven. This cutting score would correctly classify 97 per cent of the normal cases and 83 per cent of the faking bad cases.
A new index-L raw score plus K raw score minus Hs (Hypochondriasis scale) raw score- was presented to detect the faking good set. The highest phi coefficient was given by a cutting score of plus 21. This cutting score would correctly classify 97 per cent of the normal cases and 90 per cent of faking good cases.
The F-K index and the L+K-Hs index may be considered as efficient measures to detectmalingering and test dissimulation in the MMPI.