The study used data from the National Longitudinal Survey Youth-Geocode data set for 1979 and included a sample of 3,581 individuals. The research used a technique derived by Mouton in 1990 to account for the bias introduced by using aggregate data in conjunction with individual-level data. Results revealed that 2,015 of the individuals were involved in a criminal act in 1979. Results from a logit regression model indicated that race, sex, and peer pressure have statistically significant effects on the probability that a young adult will commit a crime. Thus, black males who are heavily influenced by their peers are more likely to commit crimes than are other young adults. Results also suggested that police presence, as measured by police expenditures per capita at the county level does not deter young adults from committing crimes. Tables and 39 references (Author abstract modified)