國小數學領域基本能力測驗編製與分析研究

• Main Authors: SHEN YI LING, 沈怡伶
• Other Authors: 許天維
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/43307158751756669910

碩士 === 國立臺中教育大學 === 教育測驗統計研究所 === 95 === This research focuses on one of the ability indexes in mathematics studies in elementary schools, using the principle of test writing to design items. In order to examine the adequacy of items, the testing materials of examinees are undergone a series of testing statistic analysis by relative computer programs after testing. The systematic mechanism for analyzing how test items are formed and the tests model will be able to provide a guideline to teachers and related researchers in developing a test, and also in developing test item pool of elementary school core competence test. The data is qualitatively analyzed on the item checklists, table of specification, and quantitatively on the item response theory (IRT). First of all, it examines if the testing data conforms to unidimensionality assumption of IRT and goodness-of-fit analysis. Then a model was chosen therein, and analyzed its estimating peculiarity of item analysis, such as difficult, discrimination, pseudo-chance parameters, and option analysis, and so on. As for test analysis, reliability and content validity are used as the target to decide whether this test quality is good or bad. Finally this research integrates the qualitative and quantitative analyses, checks the effect of item analysis and tools of test writing. From the analysis results, the following points can be concluded. 1. The test data fit IRT, it can be quantitatively analyzed by IRT. 2. In a serious checking by kinds of check lists, most items are good in the qualitative analysis, but some items must be checked more seriously by item setters to avoid some mistakes made in the process of setting items. 3. In quantitative analysis, we find a good result in the estimating peculiarity. On the other hand, the test reliability is over 0.8 and the test validity is good, too. 4. Through the cautious process of test and check by our tools: item checklists, table of specification, and quantitative analysis on the item response theory. The results of qualitative analysis and quantitative analysis are different. But the results are mutual influence. It provides the requirement and direction to examine items. 5. In qualitative analysis, the apprehensions of test experts were not found in the check lists. So the check lists have to be improved.