This paper discussed a multiple linear regression approach to the evaluation of instructional strategies in science. A treatment by levels experimental design and the methods required for forming and solving research problems associated with it were described.
Two methods of college biology instruction, the Audio-Tutorial and conventional techniques, were compared in terms of promoting achievement. Methods for making comparisons between the two treatments in the form of linear models were discussed. Interpreting regression solutions to linear models was also presented.
The principles expressed in this paper can be applied to other research problems as an effective alternative to the two-way analysis of variance.