Rationale. The functional analysis of the behavioral effects of drugs has lagged behind more biological research approaches in elucidating how drugs influence behavior. Part of the problem is the scarcity of mathematical models of responding under standardized behavioral procedures. A two-parameter model derived from the matching law to account for response rate as a function of reinforcement rate in single response alternative procedures provides a good description of responding under multi-component multiple variable interval (mult VI) schedules of reinforcement, both in the absence of drugs and in the presence of a variety of drugs. Objective. This paper explores how well the two-parameter model and its exponentiated variant (three-parameter model) describe drug effects on responding under mult VI schedules. Methods. Data from seven published papers concerning the effects of drugs on mult VI schedules are reanalyzed. Results. The three-parameter model describes the effects of drugs on responding more accurately than does the two-parameter model. The three-parameter model also describes well the type of relation between response rates under drug and no-drug conditions that the rate dependency hypothesis is usually invoked to explain. Conclusions. The three-parameter model can be improved by adding a parameter to account for base response rate, that is, response rate when reinforcement rate is zero. The parameters of the best descriptive model could be used to devise a classification system for the behavioral effects of drugs under mult VI schedules. The parameters should be conceptualized by relating the drug-induced changes in parameter value to converging changes in measurements of brain function. This would connect the behavioral and the biological analyses of drug action to the benefit of both research fields.