In this study, three grade 5 classes were tested during a 3 week period to determine the effect that classroom discussion has on certain quantitative aspects of composition. Each week, 78 subjects viewed a short film with no narration and--after either no discussion, class discussion, or paired-student discussion--were required to write compositions about the film. Data were analyzed in terms of total subjects and high-, middle-, and low-ability subjects. It was found that (1) numbers of T-units, adjectives, and different movables differed significantly from one writing occasion to another, (2) paired-student discussion consistently produced the greatest number of T-units while no discussion produced the smallest quantity, (3) while class discussion consistently resulted in the least number of adjectives used, the most adjectives were produced by total subjects with no discussion and by low ability students after paired-student discussions, and (4) the largest number of movables were used by total subjects after no discussion, while the least number of movables occurred after class discussion. Findings indicated that, although student participation in pre-planning leads to greater productivity, class discussion does not seem to stimulate subjects to use adjectives and movables in their compositions. (Author/JB)