No abstract

8 groups of about 20 undergraduates each were presented with 60 factual multiple-choice items, answered each question, and either received feedback immediately or 24 hr. later. For 4 groups feedback included the stem and 4 alternatives to each question. 4 groups received only the 4 alternatives. Just prior to feedback 4 groups were given a retention set and 4 groups were not. The design of the experiment was 2 (feedback) X 2 (stem, no stem) X 2 (set, no set). A 60-item immediate and 5-day delayed retention test was administered. On immediate retention only the effect of stem, no stem was significant. On the delayed retention test the groups receiving delayed feedback, the stem of the question, and the retention set performed reliably higher than their counterparts.

About 120 students in each of 4 groups received 1 of the following feedback treatments on 3 midsemester examinations: (a) looking up wrong answers in the textbook, (b) having questions discussed by the instructor, (c) checking over answers from correct ones on the board, and (d) no feedback on the questions. On a pretest early in the semester the 4 treatment groups did not differ in initial ability level. A 4 X 4 (treatments X ability levels) analysis of variance on the retention and the transfer sections of the final examination indicated that the differences among the treatments and ability levels were significant at beyond the .001 level. Further analyses by Duncan's multiple range test indicated that the discussion method was best.

Sixty factual multiple-choice items were presented by slides to 311 undergraduates who answered each question and either received immediate or 10-second delayed-information feedback. Two types of feedback cues were employed: with or without the stem of the question, and with right or with right plus wrong alternatives. The experiment was a 2 (time of feedback) X 2 (stem, no stem) X 2 (rights, rights plus wrongs) design. A 60-item mimeographed retention test was administered both immediately after and 5 days after feedback to all eight groups. On immediate retention there were no significant differences. On delayed retention Ss receiving delayed feedback performed slightly but reliably higher than 5s receiving immediate feedback, and Ss receiving no stem performed slightly but reliably higher than Sa receiving the stem of the question.