| 英文摘要 |
This research was objectively to evaluate the effectiveness of college students’elective selection of primary sign language textbook for one semester. Based on the content of primary sign language textbook, referring to relevant literature , and considering the distribution and balance of the content of primary sign language textbook, it was based on the content of vocabulary learning, including literacy, understanding of word meaning, using vocabulary to make sentences, comprehension of sentences, and comprehension of applying vocabulary to articles, etc., to compile the ''Primary Sign Language Test'' . This research compiled two sets of test for A and B copies, to explore the effect of college students after one semester of study, this research also established a norm of two sets of test A and test B at the same time, as an evaluation tool for college students to learn more sign language courses in the future. According to the content of the primary sign language textbook, this test had four sub-tests, which were:1. Vocabulary selection, 2. Vocabulary filling in the blanks, 3. Sentence comprehension, 4. Text comprehension. After the test structure and questions were determined, two deaf teachers were invited as sign language demonstration, and the test questions were recorded on a digitized computer disc for easy playback of the test to ensure the consistency of the test administration process. After the formal tests were determined, the students who chose this sign language textbook in the previous and next semesters were the subjects of the tests. There were 744 subjects in test A and 735 subjects in test B. The internal consistency coefficient of each sub-test of test A was between .66 and .82, the internal consistency of the whole test A was.91, and the internal consistency reliability coefficient of formula B was between .71 and . 80, the internal consistency of the whole test was .91, test A and test B all had good internal consistency reliability. The sub-tests of the A and B tests were all measured the same construct, and the correlations between the sub-tests were all above moderate. Both test A and test B all had good content validity. |