In an attempt to selectively improve student performance, one-half of a set of difficult test items from a FORTRAN programming class had handouts explaining the concepts underlying the items distributed to the students. Each handout contained a written learning objective, a short prose passage explaining the objective, and one or more practice test questions. Change scores were computed by subtracting the proportion of students getting a test item correct during the experimental quarter from the same proportion in a baseline quarter. Analysis showed no significant difference in the change scores for handout-treated-difficult items and untreated-difficult items. Since objectives had been shown to improve performance on difficult items in a previous experiment, the hypothesis was advanced that the effectiveness of an instructional improvement effort may depend on the learning situation in which the effort is made. (Author)