["TESOL Quarterly, EarlyView. ", "\nAbstract\nCumulative testing is known to improve vocabulary learning by integrating both new and previously introduced words in weekly quizzes. While evidence for its benefits is promising, prior research has primarily focused on the written mode of vocabulary, with target words studied, practiced, and tested in the visual mode only. This article reports on a classroom study investigating whether cumulative testing also facilitates the learning of spoken word forms. Two intact university English classes were assigned to either a cumulative n = 30) or noncumulative (n = 38) condition and learned 100 words over 10 weekly quizzes. The cumulative group was tested on both new and previously studied words, whereas the noncumulative group was tested only on the most recent words. Students practiced using online flashcards by retrieving word meanings upon hearing the spoken form or seeing the written form, with practice counts tracked for both types. Posttest results on spoken meaning recall indicated that (a) the cumulative group outperformed the noncumulative group despite slower initial progress on weekly quizzes; (b) the cumulative group engaged in practice both before and after quizzes, whereas the noncumulative group rarely reviewed; and (c) students practiced more in the written mode despite weekly quizzes delivered in the spoken mode, but it was practice in the spoken modality that showed a stronger association with learning outcomes (twice in size). Our findings underscore the importance of using tests in appropriate modalities to induce transfer‐appropriate practice for vocabulary learning.\n"]