The use of multivariate statistical analysis of Minnesota Multiphasic Personality Inventory scores in the classification of delinquent and nondelinquent high school boys.

Abstract

The research worker in psychology and education is frequently confronted with the problem of analyzing differences between two or more groups of individuals with respect to several variables. The problem of differentiating certain selected groups in terms of MMPI scores, for instance, is a good case in point. However, the kinds of statistical procedures commonly used may be subject to criticism on several grounds. Briefly, the several empirical methods of approach which have been developed lack the rigidity of a strictly mathematical model, while such univariate procedures as the t test or analysis of variance, when applied to separate keys of the MMPI, fail to take into account the amount of correlation that exists between scores on these keys. These procedures, in short, have failed to provide methods for obtaining mathematically rigid statements of profile or pattern differences.It is one of the purposes of this paper to provide an illustration of the application of more recently developed statistical procedures, namely the generalized distance function (D 2 ) and Fisher's discriminant function (LDF) to the solution of problems involving the comparison of MMPI profiles. These techniques have been employed to find answers to the following questions:1. To what extent can groups of nondelinquent and delinquent-prone ninth-grade high school boys be differentiated in terms of

Manifest aggression: MMPI correlates in normal boys.

“…Major attempts at isolating personality characteristics of delinquents or delinquency-prone children have been made by Glueck and Glueck (1950) and Kvacareus (1953). Even more successful work by Hathaway and Monachesi (1953, 1957, Wirt and Briggs (1959), Rempel (1958), and Ball (1962) are closely related to the present study and will be reported in detail. Monachesi (1953, 1957) reported the results of a study in which about 4,000 ninth-grade children in Mineapolis were administered the Minnesota Multiphasic Personality Inventory (MMPI).…”

“…In a study by Rempel (1958), multivariate statistical procedures were used in the analysis of MMPI scores and school performance with ninth-grade boys to classify potential delinquents and nondelinquents. Application of this procedure to MMPI scores alone was effective in correctly classifying 62.5% of nondelinquents and 62.3% of delinquents.…”

Changes in Adolescent Response Patterns on the MMPI/MMPI-A Across Four Decades

“…From this large data set, Hathaway and Monachesi hoped to identify MMPI predictor variables associated with later development of delinquent behaviors. Analysis of their data by themselves and others (e.g., Briggs et al, 1961;Rempel, 1958;Wirt & Briggs, 1959) found that MMPI basic Scales F,4,8,and 9 were most useful in these predictive efforts, particularly when combined with data from psychosocial and educational records (Rempel, 1958) or incidents of severe illness or death among family members (Briggs et al, 1961). Adolescent norms for the MMPI were first published in the 1970s when Dahlstrom, Welsh, and Dahlstrom (1972) and Marks et al (1974) published T-score conversion tables developed by Marks and Briggs (1967).…”

Correlates of Delinquent Behavior

“…Indeed, the selection method eliminated the "drop-outs," only one boy in the sample had served a training school senrence, fewer than 25% of the delinquents had committed acts which could be rated four or more on the severity continuum, and fewer than 20% had a record of more than one offense in the 2 yr. preceding the study. Many efforts have been made to predict delinquency with the MMPI (Briggs, Wirt, & Johnson, 1961;Gottesman, 1959;Hathaway & Monachesi, 1953, 1957, 1963Rempel, 1958;Wirt & Briggs, 1959), the Rorschach (Taniguchi, DeVos, g: Murakami, 1958), the Szondi (Coulter, 1959), the Glueck and other special instruments. All of these studies have achieved comparable results, and all are implicitly based on the ass~imption that there is a trait, or set of traits, which is antecedent to all delinquent behavior.…”