The MMPI and MCMI were administered to 163 former opiate addicts who were being maintained in a methadone program affiliated with an urban hospital. Highest group mean MMPI scores were found for Psychopathic Deviate, Depression, Hypomania, and Hysteria. For the MCMI, highest group mean clinical syndrome scores were found for Drug Abuse, Alcohol Abuse, Anxiety, and Dysthymia; highest personality disorder scores were found for Antisocial, Narcissistic, Histrionic, and Paranoid. The MCMI Drug Abuse Scale identified only 49% of subjects as having a recurrent or recent history of drug abuse. Frequency and factor analyses documented the heterogeneity of the population with respect to clinical syndromes, as well as the prevalence of personality disorders (86% had elevations on MCMI Personality Scales). Factor and correlational analyses did not provide strong evidence of similar factor structure or convergent validity of the MMPI and MCMI with this population.

Two hundred patients who chose to continue in treatment at a community mental health center were compared with 200 patients who chose to discontinue treatment after a limited number of visits. Cross‐tabulations and loglinear analysis were used to assess group differences on 11 demographic and 8 clinical variables. Cross‐tabulations provided evidence of significant group differences on 10 variables, and loglinear analysis did so for 9 variables. As a group, patients who chose not to continue in treatment tended to be younger and less dysfunctional and had less prior mental health treatment.

Patterns of juvenile criminal activity were investigated with respect to differences in a range of demographic, family, and individual variables. Subbjects were 89 juveniles who were referred to a county juvenile services center (70 males, 19 females; 78 Caucasians, 10 Blacks, 1 Oriental; mean age = 15.7). Discriminant function analyses were used to assess the discriminating power of the variables across crime categories (felony vs. misdemeanor), crime subcategories (person vs. property), number of prior referrals (none vs. one or more), and grade level (junior vs. senior high). All groups could be discriminated effectively on the basis of Wilks' lambda and associated chi‐square tests of statistical significance.

Effects of rater characteristics (role, gender, age, and parental status) on ratings of childhood problems were investigated. Four groups of raters were included: clinical psychologists, school psychologists, teachers, and parents. Ail raters evaluated the psychological importance of the 118 items comprising the Child Behavior Checklist developed by Achenbach and Edelbrock. A discriminant function analysis provided strong evidence that role groups could be differentiated on the basis of ratings. One-way analyses of variance, Scheffe post hoc comparisons, and Spearman rank-order correlations also yielded many significant results. These included role group differences for 78 items, gender differences for 12 items, age differences for 59 items, and parental status differences for 15 items.