This study addressed the question of whether teachers' judgments differ according to the frequency of data collection on student performance and the type of trend. The study also investigated whether teachers' judgments, based on different types of graphs (ascending, descending, flat, or variable), vary with the frequency of data collection. A set of 16 graphs of actual student performance data was analyzed by 59 teachers of students with moderate to profound handicaps. Results indicated that, when asked to evaluate student performance, teachers' judgments tended to be consistent and accurate when the graphed data represented continuous and systematic improvement in performance. However, when the data represented a decrease in performance, no change, or highly variable performance, teachers' judgments tended to differ according to the frequency of data collection. When asked to make program recommendations, teachers' judgments tended to differ according to the frequency of data collection for all types of trends. (Author/DB)