A self-instructional program for teaching elementary descriptive statistics was developed. The program consisted of 35 mm. slide-and-tape presentations of twenty half-hour lessons. To test the effectiveness of this approach, students majoring in elementary education were divided into four treatment groups. One group used the self-instructional program on their own in a study carrel. A second group used the program in a study carrel supplemented by mandatory attendance at a 30-minute laboratory session scheduled once a week. A third treatment group used the program in the classroom and were lectured by an instructor adhering closely to the taped materials. The fourth group consisted of a traditional lecture-discussion approach to teaching. Those students in the study-carrel-only group scored significantly lower than any other group. No other significant differences were discovered between any of the other treatments. The major conclusion based on the results of this study was that self-instructional programs are not as effective as are the traditional approaches to teaching unless the self-instructional programs are supplemented by a minimum of student-instructor contact and interaction. (JY)