The researchers hypothesized that clients admitted to treatment with any legal involvement or those who choose a treatment program as a result of incarceration would be less likely to leave treatment early than those with no legal factors at treatment entry. Concern about legal problems was reported by 21 percent of subjects in response to an open-ended question about factors that led to admission; 42 percent had other legal circumstances that were determined either from documented evidence of legal involvement from chart review or from self- report through structured questioning about legal status. For those with "concern about legal problems," the early attrition rate was 16 percent compared with 33 percent for those with no legal factors [odds ratio (OR) .38 with 95 percent confidence interval (CI) (.20, .73)]. The later attrition rate was 25 percent, OR .76, 95 percent CI (.39, 1.48). Those subjects with "legal circumstances" only had an early attrition rate of 29 percent -- OR .86, 95 percent CI (.54, 1.37) and a later attrition rate of 19 percent, OR .53, 95 percent CI (.29, .98) compared to those with no legal factors. There was little evidence of confounding in any of these results from adjusted analyses. Findings suggest that only self-reported concern about legal status as a problem mentioned at entry to treatment is significantly associated with lower early attrition, and there is no significant relationship of legal factors with later attrition. 3 tables and 15 references