This study attempts to determine if the stimulus generalization concept is applicable to school-relevant instruction when the material to be learned is verbal science information. The delivery techniques under investigation were audio, printed, and a simultaneous combination of both media. Three hundred and thirteen tenth grade biology students were randomly assigned to nine groups and a variation of the posttest-only control groups experimental design was used. Each group was tested over the presented information using one of the three delivery techniques. Both learning and evaluative materials were presented using a programed text format group-paced delivery technique. Analysis of covariance indicated no advantage in using the same media in the learning and testings sessions. The Verbal Reasoning Test scores of the Differential Aptitude Tests served as the covariate scores. These findings question the extent of the theory's applicability and usefulness to meaningful prose material in science, especially where the instruction more closely approximates school-relevant conditions at the secondary school level. (Author/CP)