Young for testing the subjects in these experiments and Thomas Nelson, Charles Weaver, Arthur Glenberg, and an anonymous reviewer for their helpful comments on an earlier version of this article.

Subjects' abilities to predict future multiple-choice test performance after reading sections of text were investigated in two experiments. In Experiment 1, subjects who scored above median test performance showed some accuracy in their predictions of that test performance. They gave higher mean ratings to material related to correct than to incorrect test answers. Subjects who scored below median test performance did not show this prediction accuracy. The retention interval between reading and the test was manipulated in Experiment 2. Subjects who were tested after at least a 24-hr delay showed results identical to those of Experiment 1. However, when subjects were tested immediately after reading, subjects above and below median test performance gave accurate predictions for the first immediate test. In contrast, both types of subjects gave inaccurate predictions for the second immediate test. Structural variables, such as length, serial position, and hierarchical level of the sections of text were related to subjects' predictions. These variables, in general, were not related to test performance, although the predictions were related to test performance in the conditions described above.

Two experiments investigated whether increased processing increases the relation between test performance predictions and test performance, i.e., increases calibration of comprehension. The amount of processing of text was manipulated by having subjects read intact text or text with deleted letters. In Experiment 1, intact versus deleted letters were manipulated within subjects, and subjects made either comprehension ease or test prediction ratings. Paragraphs with deleted letters produced higher correlations between predictions and test performance than did intact paragraphs. Better calibration with more processing was not observed for ratings of comprehension ease. In Experiment 2, in a between-subjects design, the prediction results were replicated; calibration was better for text with deleted letters than for intact text. The results show that subjects can predict performance on text material with greater than chance accuracy and that these predictions are better when subjects do more active processing during reading.

We offered introductory psychology on the World-Wide Web (WWW) and evaluated the on-line format relative to the traditional lecture-test format, using a pretest-posttest nonequivalent control group design. Multiple sections of the introductory course were offered each semester; on-line and lecture sections were taught by the same instructor, the same textbook was used, and the same in-class examinations were taken. For on-line sections, mastery quizzes, interactive individual exercises, and weekly laboratory meetings replaced lectures. Increased content knowledge was greater for the students in the Websections, as was in-class examination performance. Use of the WWW and computers for academic purposes increased more in the on-line sections, and the on-line students showed a greater decrease in computer anxiety. The students in the on-line sections expressed appreciation for course components and the convenience of the course, but the lecture sections received higher ratings on course evaluations than did the on-line sections. Learning and course satisfaction were dissociated in the two course formats.Information technology appears to hold great promise for improving learning and increasing access to higher education via distance learning opportunities (see, e.g., Bork, 1997; Dede, 1996). However, despite a long history of computer-assisted instruction and distance education, there is little well-controlled empirical research regarding the effects of such technology. In two recent reviews, one on distance learning (Merisotis & Phipps, 1999) and one on hypermedia (Dillon & Gabbard, 1998), few papers remained after reports ofinterface design, opinion pieces, and reports of research lacking control groups were excluded. Thus, even while the potential of information technology continues to grow, our knowledge about its effects on learning remains scant. Such knowledge, gained from sound designs, is needed in order to deploy instructional technology most effectively.We developed an introductory psychology course, using the World-Wide Web (WWW) as an alternative to a lecture format.' This alternative format addressed several problems associated with large introductory courses (see,

The authors investigated absolute and relative metacomprehension accuracy as a function of verbal ability in college students. Students read hard texts, revised texts, or a mixed set of texts. They then predicted their performance, took a multiple-choice test on the texts, and made posttest judgments about their performance. With hard texts, students with lower verbal abilities were overconfident in predictions of future performance, and students with higher verbal abilities were underconfident in judging past performance. Revised texts produced overconfidence for predictions. Thus, absolute accuracy of predictions and confidence judgments depended on students' abilities and text difficulty. In contrast, relative metacomprehension accuracy as measured by gamma correlations did not depend on verbal ability or on text difficulty. Absolute metacomprehension accuracy was much more dependent on types of materials and verbal skills than was relative accuracy, suggesting that they may tap different aspects of metacomprehension.