The benefits of observation in simulation-based education in healthcare are increasingly recognized. However, how it compares with active participation remains unclear. We aimed to compare effectiveness of observation versus active participation through a systematic review and meta-analysis. Effectiveness was defined using Kirkpatrick's 4-level model, namely, participants' reactions, learning outcomes, behavior changes, and patient outcomes. The peer-reviewed search strategy included 8 major databases and gray literature. Only randomized controlled trials were included. A total of 13 trials were included (426 active participants and 374 observers). There was no significant difference in reactions (Kirkpatrick level 1) to training between groups, but active participants learned (Kirkpatrick level 2) significantly better than observers (standardized mean difference = -0.2, 95% confidence interval = -0.37 to -0.02, P = 0.03). Only one study reported behavior change (Kirkpatrick level 3) and found no significant difference. No studies reported effects on patient outcomes (Kirkpatrick level 4). Further research is needed to understand how to effectively integrate and leverage the benefits of observation in simulation-based education in healthcare.