This study was conducted to determine the effects of instruction on using a formal media selection procedure on the media selection choices made by novice instructional designers. Twenty-nine male and female graduate students enrolled in a media design course at Arizona State University participated in the study. Media design problems were used which required students to develop an instructional presentation for various situations: Personnel Interviewing, Aircraft Instrument Comprehension, C.T. Scan, and Instructional Objectives. The participants were randomly assigned to groups which selected media under two conditions: an informal intuitive condition and a formal media selection condition using a model developed by Reiser and Gagne (1983). Each media selection problem had a designated "correct answer" identified by two experienced instructional designers. Results indicate that 32% of the students working in the intuitive condition and 55% of the students working with the model made the correct media selection; however, 56% of those who made correct decisions with the model did not work through it correctly. In the final, test condition, a total of 28 (96%) made the correct media selection with the model, although only 40% used the model correctly. It was tentatively concluded that use of the Reiser-Gagne media selection model can help students make more correct media selection decisions, and that the redundancy in the model and its restricted choices increase the probability of making a correct choice even if the model is not used correctly. A list of references and two data tables complete the paper. (JB)