This is a report of new research on suicide attempts, based on an analysis of data from the Epidemiologic Catchment Area surveys in the United States. Risk of making a suicide attempt during a 1-2 year observation interval in the early 1980s was estimated in relation to selected personal and behavioral attributes of 13,673 study participants who completed baseline and follow-up interviews for these surveys. Being an active case of Major Depression was associated with increased risk of suicide attempt (estimated relative odds, RO = 41; 95% CL = 6.46-262), as was active alcoholism (RO = 18; 95% CL = 2.75-118) and being separated or divorced (RO = 11; 95% CL = 1.64-77). Being a user of cocaine was associated with increased risk of making a suicide attempt (RO = 62; 95% CL = 2.51-1528), but illicit use of marijuana, sedative-hypnotics, or sympathomimetic stimulants was not (P greater than 0.30). Educational achievement was inversely associated with risk of suicide attempt at a marginal level of statistical significance (P = 0.068). These multivariable conditional logistic regression results were obtained by applying a conventional epidemiologic strategy with post-stratification of subjects into homogeneous risk sets. Limitations of the study data and the analytic strategy are discussed in relation to directions for future epidemiologic field surveys.