To determine the effect of oral responses on cloze test, particularly at the primary grade level and with students of below average reading ability, was the purpose of this study. A total of 135 students from second, fourth, and sixth grades were included in the study. At each grade level the students were divided into above-average, average, and below-average reading ability groups with 15 students in each group. Two cloze forms were prepared for each grade level using a 10 percent lexical deletion pattern. Each student completed both forms at his grade level. The forms were read silently, and one was completed with written responses and the second with oral responses which were recorded by the examiner. The t-tests for correlated means showed no significant differences between oral and written response scores at any of the three grade levels or for any of the reading ability groups. Significant positive correlations between the response modes were found at all three grade levels. The results indicated that oral responses did not affect cloze scores. Both response modes generally ranked students the same way. (Author/WR)