No abstract

Two basis assumptions of the Circumplex Model of family systems were evaluated: Healthy families evidence balanced degrees of cohesion and adaptability, whereas problem families evidence extreme degrees, and families with balanced degrees of cohesion and adaptability possess more positive communication skills than families with extreme degrees. Subjects were 58 mother-son dyads from father-absent families. In 29 of the families the adolescent was a juvenile offender, while in the remaining families there was no history of arrest or psychiatric referral. The dyads completed the Family Adaptability and Cohesion Scales (FACES) and an unrevealed differences interaction task. FACES was highly proficient at differentiating delinquent from nondelinquent families. Moreover, in families with balanced degrees of cohesion and adaptability, the mothers' communication was significantly more supportive and explicit, and the dyads evidenced significantly greater warmth and affection than dyads with extreme degrees.

Researchers of juvenile delinquency have related numerous demographic, individual, and family relationship variables to adolescent criminal activity. This study evaluated which of these variables most strongly predicted repeated and serious arrests among juvenile offenders and their siblings. Demographic, psychometric, self-report, and observational measures were collected on 74 father-present and 89 father-absent families, of which 62% and 77%, respectively, included an adolescent son who had been arrested. Multiple regression analyses were conducted separately for fatherpresent and father-absent families. Socialized-aggressive disorder, as rated by the parent(s), was the most consistent and powerful predictor of serious and repeated arrests among the male adolescents and their siblings. The second most powerful predictor was the adolescent's age at first arrest. Several family relationship measures were also significant predictors of adolescent and family arrest records. Combined, these measures accounted for a large percentage of the variance of each predicted variable. The implications of these findings for theory, research, and clinical practice are discussed.

No abstract

No abstract