OBJECTIVE The aim of this article is to examine the validity of a cue-reactivity paradigm for evaluating medications to treat cocaine dependence and to critically review cocaine pharmacotherapy studies that use this method. METHODS A Medline computerized search was performed to identify randomized, controlled medication studies for cocaine dependence that employed a cue-reactivity paradigm. Relevant bibliographies of these articles were also reviewed. Eleven placebo-controlled studies were identified in the English language literature. Four of these studies used agents that block dopaminergic neurotransmission, two studies used agents that modify the serotonergic system, and two studies used nicotinergic agents. The other three studies employed a mood stabilizer, an opioid antagonist or a psychostimulant. RESULTS There has been little research examining the theoretical basis of the cue-reactivity model, as applied to the screening of medications to treat cocaine dependence. From a methodological viewpoint, most studies have shown that exposure to cocaine-related stimuli increases subjective and physiological reactivity in cocaine-dependent patients, but methods used to present the cues and to measure cue reactivity have not been consistent across studies. Similarly, the observed increase in subjective and physiological reactivity to cocaine cues has varied within and across studies. CONCLUSIONS If a cocaine cue-reactivity paradigm is to be used to evaluate medications for treatment of cocaine dependence, the validity of the model must first be demonstrated and a consistent methodology for cue presentation and measurement of responses must be developed.