The current trend of Korean language teaching for foreign students weigh on communication purposes more than in the past. Specially in writing, the direct writing assessment has been on the rise rather than indirect writing assessment. At the same tim... The current trend of Korean language teaching for foreign students weigh on communication purposes more than in the past. Specially in writing, the direct writing assessment has been on the rise rather than indirect writing assessment. At the same time, the importance of the evaluator reliability has been increased largely because of the possibility of biased evaluation among evaluators. In line with this trend, this study investigated the result of Korean writing assessment of foreign students applying Multi-Faceted Rasch Model to analyze various aspects. Two research results were considered. First was the distribution charts of evaluator, test takers, evaluation area and evaluation questions. Second is the inter-relationship between evaluator and evaluation area, evaluator and evaluation questions. For the research, three Korean language teachers who had more 6 years experience in teaching were selected and they analytically assessed the writing assignment of the 41 foreign undergraduate students from Gyeongsang National University. The difficulty of writing assignment was about the same level as the writing test of the Korean language classes they had attended earlier in GNU. The assignment was selected by professors of the department of Korean Language and Literature. There were 4 evaluation categories and 12 items on the assignment which were provided to teachers. The results are as following; First, Korean language teacher’s severity turned out to be similar to each other by the scale ranging from .52 logit to .63 logit. The consistency of Korean language teachers engaged in writing assessment was between .99 logit to 1.10 logit. It was shown that Korean language teachers evaluated consistently as we have seen the result is placed in the scale between .6 logit to 1.4 logit. Second, the distribution chart for the capacity of test takers shows logits ranging from –3.45 to 2.22 which indicates that there were the large difference in level among test takers. Third, the difficulty levels of each evaluation categories were statable as we see its scale( .6 logit to 1.4 logit). The logit for Item difficulty also ranged –1 to 1, which is proper for students to solve. Fourth, We confirmed that there were no interactions between both evaluator and evaluation area and also with the evaluator and evaluation items from the result. In this study, we raised questions about reliability of the direct writing assessment in aspect of Korean language teacher’s severity and their consistency on evaluating. From the result, we can draw somewhat positive conclusion about evaluators' fairness and objectivity on assessing direct writing. Our discussion now are following; First, Multi-Faceted Rasch Modeling program could be used in various fields of the studies and not only for evaluator's tendency in writing test but also in other aspects too, since it can analyze the numerous educational data. More researches using FACETS are expected in near future. Second, as an evaluation tool, Multi-Faceted Rasch Modeling program can make students plan the most efficient personalized study plan. of their progress in studying. Through the FACETS, students can have not only the result of a test, but also analyzed aspects of it which will let them understand their weaknesses Third, teachers can realize their tendency on rating tests. They could put theirs severity and characteristics in mind when it comes to evaluation, they can see whether he or she works consistently and also consider other evaluator's tendency on evaluation, too. Teacher's severity, consistency, and tendency should be known to the evaluator himself. Forth, in this study, we confirmed the direct writing assessment is reliable for Korean writing proficiency. At first, we raised the question about reliability on evaluating direct writing, but the result showed that severity, consistency, and tendency of the evaluator's were not intervene as we are concerned. From the aspects of the fairness on scoring the direct writing assessment, the test reliability would be enhanced if we know the tendency of the evaluator forehead. In this study, we selected the evaluators who were under same condition so that they could not be analyzed in various way. The study could be held in comparisons of evaluators based on their working years, rating training experiences, gender and age in detail depending on evaluators' condition. We expect the findings from this study would contribute to further studies of the test score tendency and assessment area. ，韩语论文，韩语论文 |
