การเปรียบเทียบแบบทดสอบโคลซและแบบทดสอบเลือกตอบ ในการวัดความเข้าใจการอ่านภาษาอังกฤษ / นิรมล สถิตย์ทอง = A comparison of cloze and multiple-choice tests in English reading comprehension / Niramon Satidthong
Purpose The purpose of this study was to examine the correlation between a cloze test and a multiple-choice reading comprehension test in order to determine if a cloze test can be used to measure upper secondary school Thai students’ English reading comprehension and to substitute a multiple-choice test. Procedure Two English reading comprehension tests were used in this study. A cloze test was constructed by deleting every 8th word from each of the three passages which was about 300-350 words in length. A multiple-choice comprehension test was also constructed from the same passages and each of them contained 20 questions. The cloze test was first administered to 70 Matayom Suksa 5 arts and science students in Singharatpidhayakom School, and 60 Matayom Suksa 5 arts and science students in Indharasuksa School. The subjects were allowed an hour and a half to complete the test. One week after the cloze test was [administered;] the same subjects took the multiple-choice test in the same amount of time. A Pearson correlation was then computed between the cloze test results and the corresponding multiple-choice comprehension test results, and, by using a regression equation, the cloze percentage scores which were predictive of specified multiple-choice percentage scores were calculated. A t-test was used afterwards to determine the significant differences in the ability to complete a cloze test and a multiple-choice test between boys and girls. Findings and Conclusions 1. The average coefficient correlation between the cloze test results and the multiple-choice test results was .57 at the .01 level of significance. It indicated that a cloze test could be used, as efficiently as a multiple-choice test, to measure upper secondary school Thai students’ English reading comprehension. 2. There were no statistically significant differences in the ability to complete a cloze test and a multiple-choice test between boys and girls. 3. The equivalent sets of scores obtained in this study revealed that a person who got a score of 27 per cent on a cloze test most probably got a score of 50 per cent on a multiple-choice test; a cloze score of 41 per cent was most probably comparable to a multiple-choice score of 70 percent; and, similarly, a cloze score of 55 per cent was most probably comparable to a multiple-choice score of 90 per cent. Recommendations In order to obtain an effective measurement and the most suitable ways of cloze test construction and scoring, a cloze test of varied deletion and scoring should be tried on Thai students of various levels. Moreover, the cloze procedure should be employed by teachers of English to construct the reading comprehension exercises for upper secondary school students.