Retrieval enhances subsequent memory more than restudy (i.e., the testing effect), demonstrating the encoding (or reencoding) effects of retrieval. It is important to delineate the nature of the encoding effects of retrieval especially in comparison to traditional encoding processes. The current study examined if the level of retrieval, analogous to the level of processing during encoding, has an effect on subsequent memory. In 4 experiments, participants studied short lists of words, each followed by a retrieval or restudy trial. A final free recall test was given at the end of the experiment. The level of retrieval was manipulated by asking participants to retrieve words with a semantic or phonemic cue in the retrieval trial. In order to isolate the effects of retrieval per se, the semantic or phonemic cue was also presented in the restudy trial. Experiment 1 manipulated levels of retrieval (and restudy) between subjects while Experiment 2 manipulated levels within subjects. Experiment 3 sought to enhance the levels effect by adding an overt levels judgment, and Experiment 4 sought to rule out an alternative account of the equality of the testing effects across levels by increasing the list length. In all 4 experiments, a robust testing effect was obtained but it was not moderated by level of retrieval, a result supported by a small-scale meta-analysis, which demonstrated an overall effect of levels and testing condition, but no interaction.