Please wait a minute...
   2006, Vol. 38 Issue (03) : 453-460     DOI:
Minimum Chi-square/EM Estimation Under IRT
Zhu Wei,Ding-Shuliang,Chen Xiaopan

Shenzhen Archive Bureau, Shenzhen 518000, China

Computer Information Engineering College of Jiangxi Normal University, Nanchang 330027, China

Download: PDF(552 KB)  
Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks    
Abstract  A new parameter-estimation method, the minimum c2/EM algorithm for unknown parameters of the 2PLM, was proposed. The new estimation paradigm was based on careful considerations of the differences between item response theory (IRT) and classical test theory (CTT). Specifically, it is derived from a modified version of the minimum c2 algorithm originally proposed by Berkson (1955).
The starting point of the minimum c2 algorithm is the Pearson c2. Given ability score level, examinees can be classified into categories; the congruence of the sample and the expected distribution can be measured by c2 statistic. The subsequent estimation procedure is to seek appropriate item parameters to minimize c2. Because true ability scores are unobservable, most of the time, examinees are classified according to observed scores. We believe this practice is based on the point of view of CTT, which assumes that the examinees with the same observed scores have the same ability scores.
As we all know, the posterior distribution of ability parameter is affected by item parameters. Thus, the new method takes the posterior distribution of ability parameter into account and introduces artificial data in the EM algorithm for estimating the unknown parameters in IRT models. The new method redefines , (the observed proportion of correct responses and incorrect responses) of Berkson’s minimum c2 algorithm, and replaces it with artificial datum and respectively. The statistical reasoning and operations behind this method can be intuitively explained as the following:
In the minimum c2 algorithm, the observed proportion of responses is fixed and the theoretical distribution is changed with the new estimated value of the unknown parameters. In other words, the algorithm draws the theoretical distribution closer to the observed distribution and, as a consequence, the estimating speed slows down. In order to accelerate estimation, the new method connects artificial data to the item parameters through the EM algorithm so that the theoretical and the observed distribution represented by the artificial data can change simultaneously. Because item parameters are the so called “structural” parameters, whereas ability parameters of examinees are “incidental” parameters. In order to remove the effect of item parameters’ estimation to ability parameters, the new method also factored out ability parameters during estimation.
With the new estimation procedure, examinees can be classified just according to the posterior distribution of the ability parameter. After arriving at the final item parameters, examinee ability scores can be estimated using Bayesian EAP method. Through these procedures, the new method overcomes the restriction that the ability parameters must be known before estimation and expands the application range.
The results of a Monte Carlo simulation test demonstrated that the new method was not restricted by either the number of items or the number of examinees. It is also more effective and more robust than BILOG in terms of ability parameters recovery. When the number of examinees exceeded 2000, the new method was also much more effective than BILOG for item parameter recovery. The best advantage of the new method is that the ABS (the absolute value of the difference between the true and the estimated parameters) of item parameters were smaller than 0.08 when the number of examinee was 2000; the value decreased further with an increase in the number of examinees
Keywords IRT      parameters estimate      EM algorithm     
:  B841  
Corresponding Authors: Ding Shuliang   
Issue Date: 30 May 2006
E-mail this article
E-mail Alert
Articles by authors
Cite this article:   
Zhu Wei,Ding-Shuliang,Chen Xiaopan. Minimum Chi-square/EM Estimation Under IRT[J]. ,2006, 38(03): 453-460.
URL:     OR
[1] LI Meijuan,LIU Yue,LIU Hongyun. Analysis of the Problem-solving strategies in computer-based dynamic assessment: The extension and application of multilevel mixture IRT model[J]. Acta Psychologica Sinica, 2020, 52(4): 528-540.
[2] GUO Jichengsi,HUANG Jianping,WAN Xiaoang. The influence of target knowledge on path integration[J]. Acta Psychologica Sinica, 2019, 51(2): 188-195.
[3] ZHANG Weiwei,HUANG Jianping,WAN Xiaoang. Influence of expectations on human path integration[J]. Acta Psychologica Sinica, 2019, 51(11): 1219-1228.
[4] Xiaojun YUAN, Xiaoxia CUI, Zhengcao CAO, Hong KAN, Xiao WANG, Yamin WANG. Attentional bias towards threatening visual stimuli in a virtual reality-based visual search task[J]. Acta Psychologica Sinica, 2018, 50(6): 622-636.
[5] HUANG Feng, DING Qian, WEI Hua, HONG Jianzhong.  Effects of post thematic characteristics on knowledge sharing in the virtual community: The bystander effect perspective[J]. Acta Psychologica Sinica, 2018, 50(2): 226-234.
[6] ZHOU Xi; WAN Xiaoang; DU Dikang; XIONG Yilei; HUANG Weixin. Reorientation in uncontinuous virtual reality space[J]. Acta Psychologica Sinica, 2016, 48(8): 924-932.
[7] ZHANG Jing; CHEN Wei. Is body image plastic? The impact of synchrony and distance reference frame on body ownership[J]. Acta Psychologica Sinica, 2016, 48(8): 933-945.
[8] YU Liutao; BAO Jianzhang; CHEN Qinghua; WANG Dahui. The effect of individual confidence on dyadic decision making[J]. Acta Psychologica Sinica, 2016, 48(8): 1013-1025.
[9] JIAN Xiaozhu; DAI Buyun; DAI Haiqi. The weighted-score logistic model and Monte Carlo simulation study[J]. Acta Psychologica Sinica, 2016, 48(12): 1625-1630.
[10] GUO Jichengsi; WAN Xiaoang. The Effect of Learning in Virtual Path Integration[J]. Acta Psychologica Sinica, 2015, 47(6): 711-720.
[11] BIAN Yulong; HAN Lei; ZHOU Chao; CHEN Yingmin; GAO Fengqiang. The Proteus Effect in Virtual Reality Social Environments: Influence of Situation and Shyness[J]. Acta Psychologica Sinica, 2015, 47(3): 363-374.
[12] HU Bin;FENG Chengzhi. The Different Effects of PM Types, Motivation, and Task Sequence on Prospective Memory[J]. Acta Psychologica Sinica, 2013, 45(9): 944-960.
[13] DU Wen-Jiu;XIAO Han-Min. Multidimensional Grade Response Model[J]. Acta Psychologica Sinica, 2012, 44(10): 1402-1407.
[14] ZENG Li,XIN Tao,ZHANG Shu-Mei. Comparison of Two MCMC Approaches to Missing Response Data in 2PL Model[J]. , 2009, 41(03): 276-282.
[15] TIAN Jian-Quan,MIAO Dan-Min,YANG Ye-Bing,HE Ning,XIAO Wei. The Development of Computerized Adaptive Picture Assembling Test for Recruits in China[J]. , 2009, 41(02): 167-174.
Full text



Copyright © Acta Psychologica Sinica
Support by Beijing Magtech