Effects of Total and Partial Feedback in Multiple-Choice Testing Upon Learning

Abstract

Posttest performance resulting from total feedback, partial feedback, and no feedback in a multiple-choice test were compared. About 1,400 fifth and sixth graders first took a completion-format pretest which was used to match triads. Next, they all took a multiple-choice test, one-third under each of three experimental feedback conditions-total immediate feedback, partial immediate feedback, and no feedback. Finally, all examinees took a completion-format posttest. As predicted, those taking the multiple-choice test without feedback had significantly lower posttest performance than either feedback group. An hypothesized aptitude-treatment interaction was found between achievement level and feedback treatment. An unexpected interaction between sex and feedback also emerged.