We used responses of two large samples of court-ordered participants from a statewide alcohol/driving safety program to investigate factor structure, score reliability, and criterion-related validity of the Short Inventory of Problems (SIP). Exploratory and confirmatory factor analyses, using both item-level and subscore-level data, support a one-factor structure for the SIP. Internal consistency score reliability estimates were consistent across samples and high enough to warrant use for making decisions about individuals. Item response theory model calibration of the scale, using a two-parameter logistic model, yielded consistent estimates of location and discrimination (slope) across samples. Estimated scale scores correlated moderately with an independent indicator of alcohol problems and poorly with an indicator of risky driving behavior, lending evidence of convergent and discriminant validity. We judge the SIP as adequately described by a single factor, that the joint person-item scale is coherent, and scores behave consistently across samples.