Abstract
This paper reports the results of two studies which focus on the comprehension monitoring abilities of middle-grade poor comprehenders. In Experiment 1, the ability of 54 middle-school students to detect and correct text-based and reader-based inconsistencies was compared. The results indicated that the students could detect and correct reader-based inconsistencies to a significantly greater degree than text-based inconsistencies. Twenty-four of the same students participated in Experiment 2, which was designed to investigate the effects of direct training of comprehension monitoring. One-half of the students received direct instruction in two sessions; the other half served as a control group. Following training, the experimental group scored significantly higher on reader-based inconsistencies than the control group, but no significant group differences were found for text-based inconsistencies. Implications for future research are discussed.