A meta-analysis was conducted of simple confirmation feedback, which informs the learner whether a response was correct or incorrect, and feedback which provides the correct answer. Following initial selection of 72 studies of computerized and printed instruction, 15 studies were identified which met the selection criteria of involving: (1) adults (either high school or university students) learning meaningful verbal material; (2) questioning about the material and immediate feedback following responses; (3) measurement of responses to all criterion test items; (4) an immediate post-test; and (5) a no-feedback control group. Results indicate that correct response or confirmation feedback is more effective than no feedback. One possible explanation for the inconsistency in feedback findings of previous research is that various characteristics of the instructional content itself (particularly prose variations such as concreteness, vividness, and vocabulary range), independent of feedback, may influence recall to an extent that overshadows any effects of feedback. An appendix presents tables showing studies used in computing effect size of both feedback types. (Author/LMM)