Despite having studied English for some 33 months, the students at the private junior high school in Japan described in this paper had never before been asked to write original compositions in the language. The researchers undertook a quasi-experimental pilot study in which the 156 (n= 156) participants were each asked to write four compositions, two using Microsoft Word and two using pencil and paper. Their writings were analyzed for length, Japanese word usage, grammar, and spelling. Descriptive statistics and graphs show that the students produced higher word counts using pencil-and-paper, but used fewer Japanese words in their computer-based writings.