英語能力の測定技法としてのクローズ法に関する実験的研究
スポンサーリンク
概要
- 論文の詳細を見る
Background: In Japan little research has been done in the field of foreign language proficiency testing, especially, concerning English proficiency testing. The cloze procedure or the cloze test has been said to be appropriate as an English proficiency test. The cloze procedure is a generic term given to a testing technique which deletes the words of English passages according to some mechanical rules (usually every n-th). The subjects are required to fill in blanks. Since the cloze procedure was first discovered by W. Taylor in 1953, it has been applied in various areas, for example, as a testing device in Teaching English as a Second Language. Many experiments performed on the cloze showed signs of its validity and reliability. Especially, the cloze test was said to have a high practicality, such as the ease of , construction, scoring and administration. Although it seems necessary to prove the validity and the reliability of the test before it is widely used, experiments of this type have not been performed in Japan. Since both the cloze test and the listening comprehension test probably measure the internalized grammer of expectancy in English of the subjects who are non-native speakers of English, the correlation between them was found to be high. (r=.7) Unfortunately, in Japan, the subjects of the experiments are linked to 40-50 college students. However, in order to reinforce the validity of the experiments, the subjects must be more inclusive in number as well as in grade level. (i.e. the junior high and highschool) Purpose: The primary purpose of this study was to evaluate the cloze test in terms of three criteria, that is, validity, reliability and practicality. The secondary purpose of this study was to investigate the relationship between the cloze test and the listening comprehension test in case the subjects were Japanese. When the correlation between them was significantly high, the cloze test was able to measure the listening comprehension ability of the subjects. Method: As pre-tests cloze tests and listening comprehension tests were administered in March 1979. The subjects were 116 junior highschool students and 87 highschool student. The Exact-Word scoring method was used in both cases. The item analysis was conducted and based on these results, test formats were improved for the actual experiment in April 1979. The subjects were 133 junior highschool students in Experiment I and 86 highool students in Experiment II. Both in Experiment I and II, cloze tests and listening comprehension tests were administered. In Experiment I, teacher's grades, IQ score, the score of the standard test and the score of the English achievement test were gained as the external criteria. In Experiment II teacher's grades and the scores of the standard test were obtained as the external criteria. The Exact-Word scoring method was used in both Experiments. In order to evaluate the validity of the tests, 1) difficulty 2) item discriminating index were computed and 3) χ^2 test was employed on each item. Fcr the purpose of examining the internal consistency of all the tests administered in this study, the reliability coefficients were obtained by the Kuder-Richardson formula 20. In order to evaluate the criterion-related validity and the construct validity, the Pearson Product-Moment coefficient correlations were computed. For the purpose of investigating the relationship between the cloze test and the listening comprehension test, the Pearson Product-Moment coefficient correlations between them were also obtained. The factor analysis was employed to examine the relationships of all the variables in this study. Results and Discussion: The results of χ^2 test on the cloze tests showed that 20 out of 22 items were significant in Experiment I, while 38 out of 46 items were significant in Experiment II. The reliability coefficients were obtained by the Kuder-Richardson formura 20:r=.868(Experiment I); r=.916 (Experiment II). These results proved thet the cloze had a high discrimination power and internal consistency. The Pearson Product-Moment coefficient correlaticns among 6 variables gained in Experiment I were all significant at the 1% level. (r=.6-r=.8) The Pearson Product-Moment coefficient correlations among 8 variables obtained in Experiment II were also significant at the 1% level. Therefore, the cloze test proved to have high criterion-validity and construct validity. Although the coefficient correlations between the cloze test and the comprehension test were significant at the 1% level, they were not as high as those gained in foreign experiments. However, as far as this study is concerned, the cloze test may be appropriate for measuring the listening comprehension ability of the Japanese students. The internal consistency of the listening comprehension was not as high as that of the cloze test. If the listening comprehension test formats are improved, the higher correlation between the cloze test and the listening comprehension test. In order to examine the relationships of all the variables in this study, the factor analysis was emplyed. In both Experiment I and II, only one factor was extracted, which could not be named due to insufficient data, but which showed relevancy to both the listening comprehension test and the cloze test. Cloze tests attract our attention as excellent in quality with a high degree of reliability, validity, and practicality. It seems that further research should be done: (1) To increace both the reliability and validity of the listening comprehension tests. (2) To see the correlation between the Acceptable-Word scoring method and Exact-Word scoring method, both of which are also classified as scoring methods of cloze tests, and at the same time, to compute the correlation between the two methods and those measures which serve as the external criteria. (3) To increase the number of measures and investigate what items cloze tests and listening comprehension tests have in common. (4) To explain more clearly the relationship between cloze tests and speaking ability, and then point out the determinants of English proficiency.
- 国際基督教大学の論文