英译汉任务中的评分员效应研究

A Study of Rater Effects in an English-Chinese Translation Task

本文运用多面Rasch模型考察英译汉任务中由评分员主观因素造成的效应。研究使用FACETS软件分析了3名评分员对300名考生在15个句子译文上的语义、形式评分,结果表明:1)3名评分员的严厉度具有显著差异;2)在评价两种性别的考生时,评分员严厉度相当;在评价不同考生时,评分员严厉度一致;但在评价不同句子时,评分员严厉度差异显著;3)评分员所评分数未出现趋中性或两极分化趋势。这表明,主观性评分差异主要源自评分员对句子难易度的理解不同,对此需要加强对评分员的培训。

This study investigates rater effects arising from subjective judgment in an English-Chinese translation task by implementing Many-facet Rasch Model.Using FACETS,it analyzes three raters' scoring of 300 examinees' translations of 15 sentences in terms of form and meaning.The results indicate that:1) three raters' severity is significantly different;2) rater severity is not biased towards gender or towards examinees,but biased towards sentences;3) ratings do not display central tendency or go to extremes.These results reveal that subjective rating differences mainly result from dissimilar understanding of some sentences' difficulty level,and raters need more training in this aspect.