We examined two different accounts of why studying distinctive information reduces false memories within the DRM paradigm. The impoverished relational encoding account predicts that less memorial information, such as overall famililarity, is elicited by the critical lure after distinctive encoding than after non-distinctive encoding. By contrast, the distinctiveness heuristic predicts that participants use a deliberate retrieval strategy to withhold responding to the critical lures. This retrieval strategy refers to a decision rule whereby the absence of memory for expected distinctive information is taken as evidence for an event’s nonoccurrence. We show that the typical false recognition suppression effect only occurs when the recognition test is self-paced. This suppression effect is abolished when participants make recognition decisions under time-pressure, such as within 1 second of seeing the test item. These results are consistent with the distinctiveness heuristic account that a time-consuming retrieval strategy is used to reduce false recognition responses.