PRETCO口试评分标准效度验证  

A validation study of rating criteria of PRETCO Speaking Test

作　　者：杨志强[1] 全冬[1] 

机构地区：[1]重庆科技学院外国语学院,重庆401331

出　　处：《外语测试与教学》2016年第1期13-21,31,共10页Foreign Language Testing and Teaching

摘　　要：本研究基于多层面Rasch模型,通过分析评分结果对PRETCO口试评分标准的效度进行了验证,结果发现:PRETCO口试评分标准能够有效区分考生的能力;评分员评分的内部一致性总体较好;虽然评分员之间的严厉度有明显差别,但是这些差异并不影响总体的评分质量;评分员使用了评分标准中所有分数段;各个分数段的评分总体分布较为合理;因此,PRETCO口试评分标准的整体效度较高。但是,研究同样发现一些问题:"朗读"任务评分标准中两个分数段与考生能力出现了负增长;"陈述"任务评分标准一个分数段的使用频次过低,个别分数段阈值差距较小,评分员对这些分数段难以把握,容易产生混淆。基于此,本研究尝试为完善PRETCO口试评分标准以及评分员的培训提供一些启示与建议。This paper explores the validity of rating criteria of PRETCO speaking test by analyzing the rating in multi- faceted Rasch model. The results indicate that the rating criteria can help differentiate examinees of different English proficiency and maintain rating consistency of raters ; despite different leniency/severity between raters, the rating re- sult is overall desirable; all rating categories are covered; the rating distribution is reasonable; therefore, the rating criteria is valid on the whole. However, the results also show some problems： one rating category of Reading advances reversely with examinees＇ proficiency ; one rating category of Presentation is rarely used ; the advance of thresholds is less than expected and raters are likely to confuse some of these categories. Based on the results, this paper attempts to put forward some suggestions for the refinement of rating criteria and raters＇ training of PRETCO.

关 键 词：PRETCO 口试 评分标准 效度验证 

分 类 号：H319[语言文字—英语]