Abstract: This is a case study of vocabulary learning in Japanese for five students under two different instructional conditions. Instruction under both conditions was accomplished orally and interactively. Under one condition (the “output” condition), students initiated talking about video scenes using target words in groups of two to three students. Under the other condition (the “input-dominant” condition), vocabulary activities were conducted primarily through the instructor, questions about video scenes. The author examines whether output activities facilitated the retention of words more efficiently than input-dominant activities, which require students to comprehend questions involving target words but do not encourage them to use the words in communication. Data were collected at three points during one semester: one month, two months, and two and one-half months after the initial exposure to the target words. A comprehension measure shows that the students retained more words from the output condition than they did from the input-dominant condition two months after the initial exposures. Furthermore, the results show that two and one-half months after the initial exposures, although there was no difference between the two conditions with respect to contextual appropriateness of word usage, students used more words studied under the output condition than under the input-dominant condition. These results suggest that output-focused activities help students encode surface-level forms into memory better than input-dominant learning activities.