ISSN 0439-755X
CN 11-1911/B
主办：中国心理学会
　　　中国科学院心理研究所
出版：科学出版社

心理学报 ›› 2002, Vol. 34 ›› Issue (04): 37-40.

• • 上一篇    下一篇

锚题题型与等值估计方法对等值的影响

戴海崎;刘启辉   

  1. 江西师范大学教育科学学院,江西师范大学教育科学学院 南昌330027 ,南昌330027
  • 出版日期:2002-08-25 发布日期:2002-08-25
  • 通讯作者: 戴海崎

EFFECT OF ANCHOR ITEMS AND EQUATING RELATIONSHIP ESTIMATION METHOD ON TEST EQUATING

Dai Haiqi, Liu Qihui (College of Educational Science , Jiangxi Normal University , Nanchang 330027)   

  • Published:2002-08-25 Online:2002-08-25

摘要： 锚测验———非等组设计是一种非常重要的等值设计方法。研究证明 :在此设计之下作为等值媒体的锚测验采用的题型不同对等值结果会有不同影响 ;采用的等值关系估计方法不同对等值结果也有不同影响 ;题型与估计方法之间还有明显的交互作用。研究认为 ,在当前的命题与评分技术水平条件下 ,锚测验以纯客观题组成为最佳 ;在锚测验题量固定的条件下 ,等值关系估计以选用频数估计法为最佳。

关键词: 测验等值, 锚题, 等值关系估计

Abstract: Anchor items nonequivalent group design plays a rather key part in the equation design. Based on our research, the results showed the following facts: If different anchor items as equating media are adopted, the outcomes are obviously not the same, also if different equating relationship estimation methods are adopted. So we believe that best anchor testing is the objective item group under the present technology and the best equating relationship estimation method is the frequency estimation with fixed number of anchor items.

Key words: test equating, anchor items, equating relationship estimation