Establishing Behavioral Correlates: The MMPI as a Case Study

Abstract

Two methods have been used to determine MMPI behavioral correlates: the whole-sample and split-sample techniques. The latter technique, which splits a sample into half samples, is currently in use because it ostensibly controls more adequately for Type I errors by requiring significance in both half samples. By simply adjusting the level of significance for the whole-sample approach, however, it can control for Type I errors as well as the split-sample technique. Furthermore, the whole-sample approach appears to control more adequately for Type II errors than the split-sample technique. Finally, data are presented suggesting that as an indicator of the average result obtained by repeated random splittings of a sample, results from the whole-sample approach are preferable to those from a single split of a sample. The whole-sample method was thus recommended; however, the probability of a Type I error for each correlate tested should be set at values smaller than those currently used.

Refining adolescent MMPI interpretations: Moderating effects of gender in prediction of descriptions from parents.

“…In the initial analyses, point-biserial correlations were calculated separately by sex between T scores for each MMPI profile scale and those items identified as potential correlates of that scale. Consistent with the recommendations of Green (1982) regarding the reduction of Type I and Type II errors in MMPI correlate research, correlates significant for only one sex were identified when the point-biserial correlation reached a significance level of ≤.005 (one-tailed). Those PIC items that were significant for both male and female subjects at the ≤.025 level (joint probability ≤.0001) were identified as replicated correlates.…”

Criteria associated with clinical scales and Harris-Lingoes subscales of the Minnesota Multiphasic Personality Inventory with adolescent inpatients.

“…Separate analyses were conducted for male and female participants (N. H. Wrobel & Lachar, 1992), and given the multiple analyses, statistical significance was set at the .01 level (Green, 1982).…”

An MMPI study of adolescents: I. Empirical validity of the standard scales.

“…Previous MMPI correlate studies (e.g., Graham & McCord, 1985;Lewandowski & Graham, 1972) and recommendations by Green (1982) were used to determine the analysis procedures for the present study, although we made some modifications. One was the use of MMPI raw scores, instead of the non-A!-corrected adolescent T scores (Marks et al, 1974), in the analyses, allowing for the study of gender differences.…”

“…We split the sample by gender at the third quartile of subjects for each clinical scale distribution to define high-scoring boys and girls. 2 Male and female subjects scoring in the top 25% of each clinical scale were compared with the remaining subjects in their gender subsample (i.e., the comparison sample) according to procedures used in previous studies (Graham & McCord, 1985;Lewandowski & Graham, 1972) and the recommendations of Green (1982). Either t tests or chi-square analyses were used to determine statistical significance.…”

“…One-tailed t tests were used when direction was predicted in our hypotheses, and two-tailed t tests were used when direction could not be predicted. A significance level of .01 was chosen for reporting empirical descriptors, based on Green's (1982) recommendations to control for both Type I and Type II errors in MMPI descriptor research. The decision to split the groups at the third quartile and perform significance tests, rather than using a dimensional analysis such as correlation or multiple regression, reflects our desire to report the results in a manner similar to the way the scales are used clinically, as a typology of problem behaviors.…”