In order to explain delinquency, it is imperative to establish the causes of delinquency. This study investigated causation using a method in criminology. It studied the causes of delinquency by comparing within-individual and between-individual correlations of risk factors with delinquency. The analysis required repeated measures of delinquency and of presumed causal factors. The study was based on analyses of the oldest sample of the Pittsburgh Youth Study comprised of 506 boys. Delinquency was measured according to self-reports of the frequency of committing 25 types of acts in the specified time period, such as carrying a weapon, fire-setting, breaking and entering, stealing items of various value, check fraud, and gang fighting. Ten variable factors were chosen to correlate with delinquency: (1) Hyperactivity, impulsivity or attention deficit problems; (2) low academic achievements; (3) depressed mood; (4) poor parental supervision; (5) low parental reinforcement; (6) poor parent-boy communication; (7) low involvement of the boy in family activities; (8) low socio-economic status (SES); (9) poor housing; and (10) peer delinquency scores. According to these correlations, only three variables were significantly related to delinquency: poor parental supervision, low parental reinforcement, and the boy’s low involvement in family activities. The backward-lagged within-individual correlations indicated that delinquency did not cause poor supervision, low reinforcement, or low involvement. Peer delinquency was the strongest correlate of delinquency according to between-individual correlations. This suggests that peer delinquency is not a cause of delinquency. Different categories of boys had different types of relationships between risk factors and delinquency. In both simultaneous and forward-lagged correlations, poor housing was positively correlated with delinquency for boys living in bad neighborhoods. This suggests that poor housing might cause delinquency only for boys living in bad neighborhoods. Generally between-individual risk scores were better predictors of delinquency than were within-individual risk scores. Study limitations are discussed. References