RATIONALE/OBJECTIVES: Research has revealed associative learning deficits among users of ecstasy; the present study explored the component processes underlying these deficits. METHODS: Thirty-five ecstasy users and 62 non-ecstasy users completed a computer-based, verbal paired-associates learning task. Participants attempted to learn eight sequentially presented word pairs. After all eight had been presented, the first member of each pair was displayed and participants attempted to recall the second. Eight trials were administered. Correct responses on each trial, forgetting at various levels of learning, perseveration errors and the rate at which the associations were learned (trials to completion) were all recorded. RESULTS: MANOVA revealed that ecstasy users performed worse overall and subsequent ANOVAs showed that users performed significantly worse on virtually all measures. Regression analysis revealed that over half of the ecstasy-group related variance in trials to completion was attributable to group differences in initial learning and forgetting. In relation to forgetting, it appears that cannabis use may be an important determinant. In relation to rate of learning (trials to completion) and initial learning, both ecstasy and cannabis may be implicated. CONCLUSIONS: There appears to be abundant evidence of associative learning deficits among ecstasy users. However, it appears that a range of illicit drugs including cannabis and ecstasy may contribute to these deficits. (Author' s abstract)