THE PURPOSES OF THE STUDY WERE TO COMPARE (1) CONVENTIONAL LECTURE, PROGRAM-DISCUSSION, AND PROGRAM-PROBLEM METHODS OF INSTRUCTION, (2) STUDENT PERFORMANCE WITH AND WITHOUT CONFIDENCE IN PROGRAMED INSTRUCTION, AND (3) ACHIEVEMENT ON THE BASIS OF TEST SCORES ON THE FIRST ONE-THIRD OF THE CONVENTIONAL COURSE. TWO EXPERIMENTS WERE CONDUCTED IN INTRODUCTORY COLLEGE BACTERIOLOGY WITH A TOTAL OF 235 STUDENTS DURING 2 TERMS. IN EACH TERM, STUDENTS WERE DIVIDED INTO THREE TREATMENT GROUPS ACCORDING TO METHOD OF INSTRUCTION--(1) CONVENTIONAL, (2) PROGRAM ONLY (FIRST TERM), AND PROGRAM-PROBLEMS (SECOND TERM), AND (3) PROGRAM-DISCUSSION. FINDINGS WERE BASED ON RESULTS OF THREE TESTS GIVEN TO EACH GROUP. ANALYSIS OF COVARIANCE REVEALED NO SIGNIFICANT DIFFERENCES IN PERFORMANCE AMONG TREATMENT GROUPS. THERE WAS GREATER VARIABILITY OF SCORES WITHIN PROGRAM GROUPS THAN WITHIN CONVENTIONAL GROUPS. HIGH ACHIEVEMENT STUDENTS LEARNED BY PROGRAMED INSTRUCTION AS WELL AS STUDENTS OF LOW ACHIEVEMENT. IN THE SECOND TERM, STUDENTS WHO EXPRESSED NO CONFIDENCE IN THE PROGRAM METHOD DID SIGNIFICANTLY BETTER ON TESTS THAN THOSE WHO HAD CONFIDENCE. NO SIGNIFICANT DIFFERENCES EXISTED DURING THE FIRST TERM BETWEEN THE CONFIDENCE AND NO CONFIDENCE GROUPS. IT WAS CONCLUDED THAT PROGRAMED MATERIALS WERE AS EFFECTIVE AS CONVENTIONAL LECTURES IN TEACHING AN INTRODUCTORY COURSE IN BIOLOGICAL SCIENCES TO COLLEGE STUDENTS. (JM)