Can student-perpetrated college crime be predicted based on precollege misconduct?

Objectives Many colleges assess criminal histories during the admissions process, in part, to address violence on campus. This study sought to examine the utility of screening as a means of reducing violence. Methods Using cohort and case-control analyses, we identified college misconduct through college records and self-reports on a confidential survey of graduating seniors, and examined precollege behaviour as indicated on admissions records, a survey and criminal background checks. Results One hundred and twenty students met our case definition of college misconduct, with an estimated OR of 5.28 (95% CI 1.92 to 14.48) associated with precollege misconduct revealed on the college application. However, only 3.3% (95% CI 1.0% to 8.0%) of college seniors engaging in college misconduct had reported precollege criminal behaviours on their applications and 8.5% (95% CI 2.4% to 20.4%) of applicants with a criminal history engaged in misconduct during college. Discussion Though precollege behaviour is a risk factor for college misconduct, screening questions on the application are not adequate to detect which students will engage in college misconduct. This pilot work would benefit from replication to determine the utility of criminal background investigations as part of admissions.