One crucial field of the research on retrieval-induced forgetting (RIF) is to explore the mechanism which underlying this phenomenon from the perspective of individual-differences. However, there remain some inconsistent findings so far, because of failing to consider the correlated costs and benefits problem. Such ignoring make it difficult to separate costs created by an inhibition process during retrieval practice from the benefits produced by an inhibition process in the final test. The goal of this study is to examine the effect of the correlated costs and benefits problem on RIF for undergraduate students, using two variables of item competitive intensity and response inhibition ability, and furtherly clarify the issue of the mechanism that underlies RIF. In the present experiment, the stop-signal task was first administered for 137 undergraduate students, and intended to distinguish the group with high levels of response inhibition ability from groups with low levels of response inhibition ability, consisting of 32 participants respectively. And then each participant from both groups was instructed to perform retrieval practice paradigm, which consisted of three stages including study, retrieval practice, and test, in an individual cubicles. Specifically, participants were asked to study 60 categories-exemplar pairs presented successively at 5-s rate on a computer screen, each exemplar together with its category name. The order of presentation was random with the constraint that no two pairs from the same category were presented consecutively. Immediately following the study phase, participants were cued to retrieve half of exemplars from half of the categories using a category-cued word-stem completion test (30% of ending strokes were removed from the first character of each exemplar). Each pairs were practice three times. Finally, after a limited 2-min distractor task, category-cued word-stem completion test that 50% of ending strokes were removed from the first character of each exemplar were administered for all studied exemplars. A 2 (response inhibition ability: high, low) × 2 (item competitive intensity: strong, weak) × 2 (item type: Rp+, Rp-, Nrp) mixed design were conducted. The findings found a significant difference in performance of stop-signal reaction time (SSRT) between two groups of participants assessed by stop-signal task, with high levels of response inhibition individuals, as compared with low levels of response inhibition individuals, presenting less SSRT. More importantly, in retrieval practice task following the stop-signal task, the results showed that participants with high levels of response inhibition ability presented significant RIF in the condition that unpracticed items interfered the recalling of practiced items strongly, but not in the condition in which the retrieval of practiced items were disrupted slightly. In contrast, participants with low levels of response inhibition ability didn’t show significant RIF for unpracticed items regardless of item competitive intensity. Taken together, the current results reveal that, for participants with a normal level of response inhibition ability, RIF can be attributed to inhibitory control that is generally considered as an intentional, resources-demanding processes. In particular, the finding supports that item competitive intensity is an effective operation that can separate the effect of the correlated costs and benefits problem from RIF.