A COMPARISON OF THE DIAGNOSTIC ACCURACY OF LONG AND SHORT MMPI FORMS
The diagnostic accuracies of the standard MMPI and several short forms were compared by determining the degree to which each test version predicted actual clinical diagnoses in a mixed sample of 100 adult psychiatric patients and normals. Five short forms of the MMPI (the FAM, Hugo, and three scoring versions of the MMPI-168) were extracted from the standard MMPI answer sheets for the entire sample, to obtain six MMPI profiles (one long form and five short forms) for each subject. Without concurrent knowledge of these profiles, subjects were given criterion diagnoses of Neurotic, Psychotic, Sociopathic or No Significant Psychopathology based on the Goldberg (1972) hierarchical diagnostic classification system. Explicit procedures for making the criterion diagnoses were obtained by making a simultaneous diagnosis according to DSM-III. Diagnoses were made by psychiatrists or other mental health professionals based on multiple sources of information about each subject.^ In group data analyses, each of the short forms was compared to the standard MMPI using traditional group measures. For most comparisons, the short forms showed basic psychometric similarity to the long form, although less-than-perfect matches were found using statistical procedures. In individual data analyses, diagnostic predictions by both clinical and actuarial methods from all six test versions were compared for accuracy to the criterion diagnoses. For both prediction methods, no significant differences were found between the accuracies of the test forms, with all versions predicting the target diagnoses about half of the time. For both methods, technically invalid profiles performed no worse than technically valid profiles in making diagnostic predictions for all test versions, with one exception. A breakdown by diagnostic category did reveal some statistical differences between and within the test forms.^ Conclusions of this research were that the MMPI, regardless of version, is a poor predictor of actual clinical diagnoses; within this context, short forms perform no worse than the long form; neither profile validity status nor profile diagnostic method significantly affects diagnostic accuracy; and, all the short forms studied, with the possible exception of the 168-74 and Hugo are probably adequate substitutes for the long form MMPI for gross-category diagnosis.^
KROSSNER, RHONDA PARRELLA, "A COMPARISON OF THE DIAGNOSTIC ACCURACY OF LONG AND SHORT MMPI FORMS" (1983). ETD Collection for Fordham University. AAI8323537.