With increasing medical knowledge, procedural, and diagnostic skills to learn, it is vital for educators to make the limited amount of teaching time available to students effective and efficient. Generative retrieval is an effective and efficient learning tool, improving long-term retention through the practice of retrieval from memory. Forty medical students were randomized to learn normal cardiovascular anatomy using transthoracic echocardiography video clips in a generative retrieval (GR) or standard practice (SP) group. GR participants were required to verbally identify each unlabeled cardiovascular structure after viewing the video. After answering, participants viewed the correctly labeled video. SP participants viewed the same video clips labeled with the correct cardiovascular structure for the same amount of total time without verbally generating an answer. All participants were tested for intermediate (1-wk), late (1-mo), and long-term (6- to 9-mo) retention of cardiovascular anatomy. Additionally, a three-question survey was incorporated to assess perceptions of the learning method. There was no difference in pretest scores. The GR group demonstrated a trend toward improvement in recall at 1 wk [GR 74.3 (SD 12.3); SP 65.4 (SD 16.7); P 0.10] and 1 mo [GR 69.9 (SD15.6); SP 64.3 (SD 15.4);P 0.33]. At the 6- to 9-mo time point, there was a statistically significant difference in scores [GR 74.3 (SD 9.9); SP 65.0 (SD 14.1); P 0.042]. At nearly every time point, learners had a statistically significantly higher perception of effectiveness, enjoyment, and satisfaction with GR. In addition to improved recall, GR is associated with increased perceptions of effectiveness, enjoyment, and satisfaction, which may lead to increased engagement, time spent studying, and improved retention.