A Comparison of Several Feedback Methods for Correcting Errors by Computer-Assisted Instruction.

Studies which have utilized low-error-rate linear type programs have not been able to compare the effectiveness of various modes of feedback in correcting error in programed learning. In the present study using 75 university students, it was possible to correct errors without teaching erroneous material by using materials designed to teach 30 commonly misunderstood concepts in general science by means of computer-assisted instruction (CAI). The five treatment groups differed only with respect to feedback modes which were no feedback; feedback of "correct" or "wrong"; feedback of the correct response; feedback appropriate to the student's response; and a combination of modes two, three, and four. Items missed were presented repeatedly until a criterion of correct response to each item had been attained. There are indications that the subjects who received feedback guiding them to the correct response were learning more effectively and performed better than those who were forced to "discover" the correct response. Data indicate that for correcting error, providing a student with a statement of which response was correct or why the correct response was correct may be the most valuable. The analysis of variance on posttest scores indicates that a combination of modes is slightly superior to some of the individual feedback modes in affecting immediate retention. (MT)