An Investigation of the Reliability of the Cloze Procedure Compared with Teacher Judgment for the Purpose of Determining Instructional Reading Texts

Abstract

Numerous investigations have been conducted to examine the cloze procedure as an assessment of both the readability of materials and reading comprehension. An extension of this research has been in the use of cloze procedure to determine performance levels. Of the research in this area, only a limited amount has been conducted that investigates the cloze procedure as a reliable predictor of instructional level as compared to teacher judgment. This study was designed to investigate the reliability of the cloze procedure as compared to teacher judgment for the purpose of placing students in an instructional reading text. The eight participating teachers received specific information on each of the 24 randomly selected third-, fourth-, and fifth-grade students. On the basis of the given information the teachers each recommended reading text placement within the Houghton, Mifflin Readers (Durr, 1971). The students were given cloze passages constructed from representative passages at each level of the basal. Testing terminated when the student's cloze test score was within the 30-45% instructional level. Inter-rater agreement, as assessed by an analysis of variance, was .29. The mean placement differences between cloze procedure and teacher judgment were assessed by dependent t-tests (for correlated samples). With respect to book level, the difference between the two methods was statistically significant. The difference between cloze and teacher placement with respect to readability level was not significant. The two placement methods were highly correlated. In this study the cloze procedure was determined to be reliable compared to teacher judgment with respect to readability level. However, the cloze procedure did not produce the same placement as teacher judgment when established book level was used as the basis of comparison.