A study of reading comprehension using computer-assisted instruction

This study investigated comprehension training through sentence combining, through main idea reading, and through the cloze procedure (usinq computer-assisted instruction). Fifty-one general psychology students served as the control group, and ninety-three developmental reading students served as the experimental subject pool. The experimental group was divided randomly into three different training groups. Effects of instruction were measured by pre- and post-test gains on the Reading Comprehension subtest (Descriptive Tests of Language Skills) and the General Reading subtest (Diagnostic Reading Test). Overall gains were significant. Subscore comparisons showed significant group differences: the sentence-combining group scored higher on the direct statement section of the Reading Comprehension subtest than did either the main idea group or the cloze group; the main idea group scored higher on the main idea portion of the Reading Comprehension subtest than did either the sentence-combining group or the cloze group; and the cloze group scored higher on the inference portion of the Reading Comprehension subtest than did the main idea subgroup or the sentence-combining subgroup (although this last difference failed to meet the criterion for significance). These results have important test-selection implications for treatment- effect evaluation. The results further suggest there are kinds of comprehension tasks which can be taught effectively in solation.