규척빈칸메우기 시혐의 문항수에 관한 연구

Abstract

This paper examines the ideal number of blanks in cloze procedure. 25 subjects who were English major juniors volunteered to participate in the experiment and took six cloze tests each of which consisted of 25 blanks. Both exact word scoring method and acceptable word scoring method were employed and the results of both scoring methods seemed to favor a test of more than 75 blanks. Especially, the K-R reliability excluded the employment of a cloze test of less than 50 blanks. Spearman ra퍼ζ order correlation between 75 blanks vs 100 blanks" marked .92 by exact word scoring method and .93 by acceptable word scoring method. In the light of all these results and their interpretations, it may be concluded that an ideal number of blanks of a cloze test should be at least 75 or more.