Human abuse potential studies are randomized crossover studies, with numerous measures, endpoints and comparisons. These studies use enriched populations and multiple doses of investigational drug, comparator(s) and placebo. These study design factors, along with the often non-linear dose-responses and non-normal distribution of subjective responses can create challenges in analysis and interpretation of study results. This presentation will introduce some of the key concepts in design of abuse potential studies, and discuss some of the inherent issues associated with analysis and interpretation of these data from a non-statistician's perspective, including potential underlying causes of these issues from a design perspective. The presentation will also offer some practical approaches to addressing some of these issues.