The classrooms of 40 teachers in the Kamehameha (Honolulu) Elementary Education Program (KEEP) in grades 1 through 3 were used to investigate the influence on achievement of 12 important instructional variables. The KEEP is an elementary-level language arts program designed for Hawaiian learners. The independent variables concerned were, as follows: quality (program strategies, teaching strategies, program experience, teaching experience, and a composite of eight variables); task-ability match (work completion); motivation (participation); and time (daily instruction, attention, on task when not receiving direct instruction, management, and homework). The achievement of the 40 classrooms, with the median number of students in each classroom being 22, 21, and 25 for the 1st, 2nd, and 3rd grades, respectively, was measured using the Metropolitan Achievement Tests (MAT). The expected MAT mean for each classroom was estimated from that class's mean ability as measured by the Otis-Lennon School Ability Test. The measure of achievement used in the analysis was the standardized residual of the MAT score from the estimated mean. Only the number of years as a program teacher was related to achievement. This result was found in 2 consecutive years (1985 and 1986), but was puzzling because none of the 12 variables expected to be critical for instructional effectiveness was associated with achievement. Possible explanations are considered. Because the amount of teaching experience with this instructional program was associated with achievement, the results provide encouraging evidence for the effectiveness of KEEP, although they do not support the effectiveness of current models of instruction. (SLD)