Interactive videos are frequently employed in education. Although several reviews and syntheses indicate that interactively engaging with videos might benefit learning, up until now no quantitative synthesis of the effectiveness of enhanced interaction features in educational videos has been published. Enhanced interaction features explicitly aim to encourage learning processes. Very often, they consist of domain-specific questions and tasks to promote retention and understanding. The presented research evaluates the effectiveness of enhanced interaction features in educational videos. A quantitative synthesis of research on learning from interactive videos was conducted. Sixteen articles reporting 17 studies with 22 effect sizes met the inclusion criteria. Based on a random-effects model, both a meta-analysis and a moderator analysis were conducted. Videos that include enhanced interaction features are significantly more effective than videos that comprise merely navigational interaction features or no interaction features. The analysis resulted in an overall effect size of Hedges' g = 0.522 in favor of videos that include enhanced interaction features. However, if a video is not paused while an enhanced interaction feature is offered to the learners, a split-attention effect occurs, and learning is impeded.