期刊
JOURNAL OF EDUCATIONAL AND BEHAVIORAL STATISTICS
卷 45, 期 3, 页码 297-315出版社
SAGE PUBLICATIONS INC
DOI: 10.3102/1076998619885636
关键词
item characteristic curves; multinomial log-odds transformation; log-odds slope weighting; sum scoring; optimal scoring; two-stage test data analysis
资金
- Natural Sciences and Engineering Research Council of Canada [320-2012-RGPIN, 2014-578]
Ramsay and Wiberg used a new version of item response theory that represents test performance over nonnegative closed intervals such as [0, 100] or [0, n] and demonstrated that optimal scoring of binary test data yielded substantial improvements in point-wise root-mean-squared error and bias over number right or sum scoring. We extend these results by showing that optimal scoring of the full information in option choices produces about as much further improvement in these measures of score performance as was achieved by going from sum scoring to optimal binary scoring.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据