竞教比赛的多层面Rasch模型分析

运用多层面Rasch模型，从选手、评委、性别和评分量表等几个层面对英语专业青年教师竞教比赛评分质量进行了研究。该研究涉及5名评委和6名选手。研究结果表明，选手的教学能力有显著差别。评委在评分严厉度上存在差异，这种差异可能会影响比赛结果的信度。评分量表不能较好地区分选手的教学能力，评委在评分过程中有较为严重的趋中现象。偏差分析表明，部分评委对某些选手的评分有显著偏差。

A Many Facets Analysis of Lecture Demonstration Contest

This study attempts to validate a lecture demonstration contest performed by junior lecturers in an English faculty of China. In all, five raters and six contestants were involved in the study and five weighted categories of the rating scale were employed to score the performance of these contestants. Results show that some raters were found to have poor internal self-consistency and they exhibited significant difference in severity. Most of the probability curves for the categories were shown to have indistinct peaks and some of the raters were evidenced to have halo effect in rating. Also, bias interaction was found among these facets. Generally, the rating scales of the rubric used in this study worked poorly for their purposes.