The research reported here uses information from the admissions files of lawyers admitted to the Connecticut bar from 1989 to 1992 to compare those who were disciplined with those who were not disciplined. It analyzes information reported during the bar admissions process that may predict later lawyer misconduct including, inter alia, prior criminal history, problem credit history, prior employment history, academic misconduct, substance abuse, and psychological history. The study reveals that many of the responses on the admissions application are statistically associated with an elevated risk of future discipline. Nevertheless, these variables nevertheless make very poor predictors of subsequent misconduct. The explanation for this seeming paradox is that the overall baseline likelihood of discipline is so low (only about 2.5% of the 6,159 lawyers in our cohort). Thus, even if some variable (e.g., having defaulted on a student loan) doubles the likelihood of subsequent disciplinary action — a very strong effect — the probability of subsequent discipline for someone with a student loan default is still only 5%. It seems highly unlikely that any regulator would be comfortable denying admission to an applicant who had only a 5% chance of subsequent discipline. Put differently, even knowing that an applicant has a substantially elevated risk of future discipline is probably not sufficient to justify some kind of corrective or preventative action, given the low baseline risk.