Spaced retrieval practice consists of repetitions of the same retrieval event distributed through time. This learning strategy combines two "desirable difficulties": retrieval practice and spacing effects. We carried out meta-analyses on 29 studies investigating the benefit of spacing out retrieval practice episodes on final retention. The total dataset was divided into two subsets to investigate two main questions: (1) Does spaced retrieval practice induce better memory retention than massed retrieval practice? (subset 1); and (2) Is the expanding spacing schedule superior to the uniform spacing schedule when learning with retrieval practice? (subset 2). Using meta-regression with robust variance estimation, 39 effect sizes were aggregated in subset 1 and 54 in subset 2. Results from subset 1 indicated a strong benefit of spaced retrieval practice in comparison with massed retrieval practice (g = 0.74). Results from subset 2 indicated no significant difference between expanding and uniform spacing schedules of retrieval practice (g = 0.034). Moderator analyses on this subset showed that the number of exposures of an item during retrieval practice explains inconsistencies between studies: the more learners are tested, the more beneficial the expanding schedule is compared with the uniform one. Overall, these results support the advantage of spacing out the retrieval practice episodes on the same content, but do not support the widely held belief that inter-retrieval intervals should be progressively increased until a retention test.