The learning advantages associated with performer-generated performance assessment is well established in the motor learning literature. In such contexts, performers actively interpret their response-produced feedback to provide an explicit hypothesis about the success of their motor action. Such processing could be considered to place higher demands on the cognitive processes of the learner, compared to those not required to estimate their motor performance. The purpose of the present experiment was to determine whether the degree of invested cognitive effort differentially modulates the accuracy of performance appraisal, and subsequent learning. Participants were randomized into 1 of 3 experimental conditions: 1) the Recall group was asked to estimate their overall movement time after each acquisition trial, similar to past research and considered to require high cognitive effort; 2) the Recognition group was presented with 3 movement times (one of them the actual movement time) and were required to select their perceived movement time (considered moderate cognitive effort), and 3) the Control condition did not require explicit performance assessment (low cognitive effort). Following each acquisition trial, KR was presented outlining goal movement time, actual movement time, and estimated time (excluding the control condition). All participants were required to learn a serial 5-keypressing task in a goal time of 2550 ms over 90 acquisition trials. A no-KR retention test was conducted approximately 24-hours later. The retention test showed the performance estimating conditions (recall and recognition) demonstrated superior learning (indexed by |CE|) and performance appraisal (indexed by an Absolute difference measure) compared to the control condition; with no difference between the performance appraisal conditions. Our results are consistent with the ‘hypothesis testing’ notion proposed by Guadagnoli & Kohl (2001), and further suggest an alternative method of facilitating performance appraisal for the learner.