Five studies investigated the conjunction effect (or conjunction fallacy), in which participants report that the conjunction of two events is more rather than less likely than one of the events alone. There was no evidence that feedback or monetary reinforcement for correct answers affected students' performance on conjunction problems. Under some circumstances the context in which the conjunction problem was presented (after questions emphasizing logic or questions emphasizing opinions) affected occurrence of the effect. Location of the conjunction among the statements being rated had a significant effect. The effect occurred with or without a framing description and whether the conjunction consisted of two or three simple statements. However, statements representing the conjunction of three simple statements were (appropriately) judged less likely than those representing the conjunction of two simple statements. The substantial incidence of the effect, even without the descriptive frame and even when incentive and feedback were provided for correct answers, argues for its robustness.