For health science students to learn huge amounts of information they must optimize their ability to retain information. Quiz‐based learning has been shown to be more effective than passive restudying, a phenomenon known as the “testing effect.” In medicine, appropriate self‐assessment of ones’ own knowledge is an important quality to limit medical error. We hypothesized partner study groups (PG) would outperform individual groups (IG) on knowledge tests based on increased adherence to quiz‐style learning. We thought PG would appropriately indicate their confidence on answer choices since they quizzed with and received feedback from their partner on incorrect answers while studying.Graduate level students recruited from Pacific Northwest University of Health Sciences were given information on a novel anatomical topic in an initial study session. They then were split into groups: IG restudied the material by themselves and PG as student pairs. Following the restudy session, students’ knowledge of the topic was assessed by a 15‐question test. Follow‐up tests occurred 1 week and 3 weeks later. Each test item required the student to indicate if they felt confident or not in their answer. Answers marked as confident that were incorrect were labeled as “misinformed” answers. Students filled out surveys to report prior knowledge, study habits, and reflections on their learning experience on day 1 of the study. Data were analyzed using SPSS software with p < .05 via two‐sided testing considered statistically significant using repeated measures ANOVA and Mann‐Whitney U tests.Day 1 surveys completed (n = 17) showed PG and IG were not significantly different for age, gender, memory scores, or prior knowledge. PG reported feeling the material was easier to learn compared to IG (p = .024). Knowledge tests at day 1, week 1 and week 3 were completed by 15 students. PG had lower average test scores than IG on each of the 3 tests; the difference between groups was not significant (p = .302), but there was a moderate effect size (ηp2= .082). PG reported more confidence on answers (p = .272, ηp2 = .092). The number of misinformed answers was more for the PG compared to IG; while not significant, there was a moderate to large effect size measured (p = .205, ηp2= .121). For all students, confidence on answers dropped significantly between day 1 and week 1 (p < .0001), and on average had less misinformed answers (p = .0854).This study found, contrary to our initial hypothesis, students who studied alone were more likely to do better on knowledge tests. However, the findings also suggest that students who studied in partner groups found it easier to learn and felt more confident while taking the tests despite performing worse. There was an increase in misinformed answers for partner groups compared to individuals, likely because they were more confident overall. This corresponded with total student data; participants were more accurate in their assessment of their knowledge after day 1. When total confidence decreased, there were fewer misinformed answers. A follow up study is warranted to evaluate why there was higher confidence seen in partner groups, even when they did not know the correct answer. The limitations of this study include a low sample size and students were from a single institution.