Prior research has investigated whether learners spontaneously adapt their encoding strategies in anticipation of particular test formats (i.e., the "encoding-strategy adaptation hypothesis"; Finley & Benjamin, 2012). However, the strongest evidence supporting this hypothesis is confounded with test experience (as argued by Cho & Neely, 2017). When learners gain equal experience with each test format, do they adapt their encoding strategy use? Across 3 experiments, participants studied lists of cue-target word pairs and after each list completed either a cued-recall test (recall targets given cues) or a free-recall test (recall targets only). Participants received equal experience with each test format. On a final critical test, participants either received a test in a format they expected or one that violated their expectations. On this critical test, participants who received a test they expected outperformed those who did not, and this was true for both cued and free recall. Also, a manipulation of cue-target associative strength had a greater effect on cued-recall tests than free-recall tests (Experiment 1), whereas a manipulation of target-target associative strength had a greater influence on free-recall tests than cued-recall tests (Experiments 2 and 3). These findings, along with divergent patterns of self-reported strategy use for the 2 anticipated test formats, support the encoding-strategy adaptation hypothesis. In particular, learners tend to use more cue-target associative strategies when expecting a cued-recall test, and more target-focused strategies when expecting a free-recall test.