Methods for determining the likelihood that a subject will be a placebo responder in a clinical study are provided. Also provided are methods for eliminating likely placebo responders from a clinical study a priori, thereby simplifying data analysis and minimizing or eliminating any confound that arises in the analysis as a result of placebo response. Databases and computer systems using the methods are also disclosed herein. Methods for assessing likelihood of a subject experiencing a response shift are also provided.
