This paper evaluates judgments of driver crash culpability to estimate alcohol and drug impairment effects when exposure data are unavailable to calculate relative crash risks. Previous studies using culpability judgments provided some evidence that culpability is related to BAC. Other studies, some inferring culpability, indicated a relation between culpability and relative crash risk. Data are presented showing that culpability judgments with a rating scale have high interrater reliability, and systematic relations with BAC suggest some validity in the ratings. A method is demonstrated for estimating relative crash risk from culpability judgments with accident data, and the limitations of culpability analysis are discussed. While alcohol and drug impairment effects are best determined with relative crash risks determined from accident and exposure data, culpability analysis may provide useful indications in the absence of exposure data.