IN THIS EXPERIMENT, 186 FRESHMAN ENGINEERING STUDENTS STUDIED ELEMENTARY PROBABILITY BY THESE INSTRUCTIONAL METHODS--(1) MULTIPLE CHOICE TEACHING MACHINES, (2) FREE-RESPONSE TEACHING MACHINES IN INDIVIDUAL BOOTHS AND IN CLASSROOMS, (3) PROGRAMED TEXTS REQUIRING OVERT RESPONSES AND GIVING CORRECT ANSWERS, (4) PROGRAMED TEXTS REQUIRING NO OVERT RESPONSES, (5) "PROGRAMED" LECTURERS AND (6) STANDARD LECTURERS. THE STUDENTS FIRST TOOK AN ENGINEERING APTITUDE TEST, ON WHOSE BASIS THEY WERE RANKED AND THEN RANDOMLY ASSIGNED TO GROUPS. THEY WERE TESTED FOR LEARNING IMMEDIATELY AFTER INSTRUCTION. RESULTS INDICATED NO SIGNIFICANT DIFFERENCE IN STUDENT LEARNING BY ANY OF THE PROGRAMED METHODS, BUT ALL OF THE PROGRAMED METHODS WERE SIGNIFICANTLY MORE EFFECTIVE THAN THE STANDARD LECTURE. LEARNING TIME WAS SIGNIFICANTLY DIFFERENT FOR THE VARIOUS METHODS, MACHINE METHODS REQUIRING MOST TIME AND LECTURE METHODS LEAST TIME. (MS)