Despite the robust memory-enhancing benefits of retrieval practice, an initial test can sometimes exacerbate eyewitness’ susceptibility to subsequent misinformation—a phenomenon known as retrieval-enhanced suggestibility. One explanation for this finding is that after taking a memory test, participant witnesses are more likely to treat the subsequently presented misinformation narrative as being credible (the misinformation acceptance account; e.g., Chan, Manley, & Lang, 2017). Another explanation suggests that prior testing enhances participants’ ability to learn the post-event misinformation (the test-potentiated learning account; Chan et al., 2017). In two experiments, we provided a direct test of these alternative hypotheses and showed that testing did not make participants believe the content of the narrative to be more accurate, and the perceived accuracy of the narrative was not associated with the likelihood of misinformation recall. These results are inconsistent with the misinformation acceptance account.