This meta‐analysis examined the effectiveness of an additional gloss mode in single versus dual and dual versus triple glossing on second language (L2) learners’ word learning. In total, 22 studies, providing 26 independent effect sizes, were coded, and 11 moderator variables including quality of data sample, learner variables, gloss features, text features, and methodological features were examined. The results show that the overall effect of an additional gloss mode was medium (g = 0.46) for immediate posttests and small (g = 0.28) for delayed posttests. However, analyses of moderator variables indicated that the effect of additional gloss modes is influenced by a range of variables related to learner (e.g., proficiency), gloss (e.g., language), text (e.g., narrative vs. expository), and research design (e.g., test format). Importantly, adding an additional mode to single textual gloss enhances vocabulary learning, whereas adding a mode to dual glossing does not result in significantly better vocabulary learning. The findings suggest that using more than two gloss modes is not necessary because it does not always lead to better learning of new words.