Assessment of Local Influence on Item Parameter Estimation
Local influence analysis is an instrument for diagnosing influential observations of a statistical model. Cook’s (1986) initial proposal and Poon’s (1999) development of local influence method are discussed in this paper. Poon’s measurement is derived for assessing local influence of response patterns in item response theory modeling, particularly in the 2-parameter logistic model. The impact on estimating all item parameters and part of parameters is both considered. A real data set is used as illustrative examples to show the power of the proposed diagnostic approaches.
Local Influence IRT Response Pattern Joint Mazimum Likelihood
WU Ruilin YUAN Ke Hai ZU Jiyun
Institute of Psychology and Behavior, Beihang University, P.R.China, 100191 Department of Psychology, University of Notre Dame, United States, 46556
国际会议
2009 International Institute of Applied Statistics Studies(2009 国际应用统计学术研讨会)
青岛
英文
1-5
2009-07-25（万方平台首次上网日期，不代表论文的发表时间）