One of the main behavioral patterns of the episodic memory decline that accompanies aging is an “associative deficit.” This deficit involves a reduced ability to remember new associations. Retrieval-based learning has been explored in the literature as a potential strategy to reduce the associative deficit when learning non-verbal material. In the present study, the main question of interest was whether retrieval practice could serve as a strategy to reduce the associative deficit when learning verbal material. Other questions of interest involved whether an advantage of retrieval practice is observed when learning semantically unrelated material, and whether individual differences determine who benefits from retrieval practice, who benefits from re-study practice, and who benefits to the same extent from either type of learning practice. Participants from the community (N = 106) completed measures of episodic memory and processing speed, and an experimental paired-associates task. The experimental task involved learning 40 pairs of words (20 semantically related and 20 semantically unrelated). The learning strategy was manipulated so that participants practiced learning half of the related (10 pairs) and half of the unrelated pairs (10 pairs) through re-study and the other half the pairs through retrieval. After 24 hours, participants tried to recall all the pairs. Because a pattern of associative deficit was absent in the data, it was not possible to conduct the tests for hypotheses exploring whether retrieval practice reduces the associative deficit. As for the other questions of interest, the results replicated the findings in the retrieval-based learning literature: retrieval practice produced more memory benefits than re-study practice. However, when considering the semantic relatedness condition, the advantage of retrieval practice was observed only in the semantically related condition. When taking into consideration that not everyone benefits from retrieval practice, the results show that the benefits of retrieval and re-study practice are of similar magnitude, regardless of semantic relatedness. However, the proportion of participants that benefit from retrieval practice is larger than the proportion that benefits from re-study practice. Older participants with lower episodic memory scores do not show memory benefits from the baseline effects of either re-study or retrieval practice.