基于试题内部结构和评分过程的口试构念效度研究——以PETS四级为例

Construct Validity of an Oral Test: A Study Based on the Internal Structure and Assessment Process of the Test— Taking PETS4 as an Example

通过分析试题内部因子结构和评分员的评分心理反应过程证据来验证PETS口试的构念效度。分析结果表明：1）PETS口试试卷内部结构主要呈现一个因子；2）PETS口试较好地体现了所要测试的构念；3）评分受到了与构念无关因素的影响，如句子结构、自我修正、故事完整性和配对；4）评分员背景对测试的构念效度有影响。本文据此提出了影响PETS口试构念效度的因素模型。

This study, by analyzing the internal factors of an oral test and the raters＇ psychological verbal reports while rating, aims to investigate the construct validity of the oral test. The results indicate： 1 ） PETS oral test chiefly obtains one factor; 2） PETS oral test possesses good construct validity; 3） the raters are influenced by some variances,such as structures of sentences, self-correction, integrality of story and the partner, which are irrelevant to construct validity; 4） the raters＇ background affects the test＇s construct validity. Then, one tentative model is proposed to account for and improve the construct validity of PETS oral test.