THE COMPARISON OF ITEMS‟ AND TESTEES‟ABILITY PARAMETERudESTIMATIONudIN DICHOTOMOUS AND POLITOMUS SCORINGud(STUDIES IN THE READING ABILITY OF TEST OF ENGLISH PROFICIENCY)

摘要

This study aimed to compare the testees‘ ability estimation in the politomus anduddichotomous scoring model. The data used in this study are the responses of testees to theudTest of English Proficiency (TOEP) set 1 in reading subtest, which are usually scoring inuddichotomous model then they are scoring in politomus model. In the reading subtest ofudTOEP, in one text presented several items related to the text. In the dichotomous scoring,udeach item is scored one by one item. As alternative, every item item is scored usinguddichotomous model separately, but for every text, the acquisition of these items are added toudthe score attained politomous model. The estimation of items‘ and abilities‘ parameter inuddichotomous scoring were done using the Rasch models and in the politomous scoring wereuddone with partial credit models using QUEST software. Comparative analysis of the twoudmodels are seen based on the average results of the estimated difficulty level, graphicaludanalysis, calculating the correlation, and the results of the value of information function. Theudresults of the analysis showed that the average item difficulty dichotomous scoring model isud0.486 with a standard deviation of 0.895 and the mean level of difficulty politomous scoringudmodel is -0.105 with a standard deviation of 0.695. The correlations between abilities ofudparticipants using the dichotomous and the politomous scoring model is 0.94. The value ofudinformation function in the dichotomous scoring model is higher than in the politomousudscoring models. These results indicate that the Reading of TOEP set 1,the dichotomousudscoring model is better than the politomous scoring model.