Person Fit Analysis For Cognitive Diagnostic Assessment

In recent years,Cognitive Diagnostic Assessment(CDA)has been widely used in educational assessment.It can provide guidance for further study and teaching by analyzing whether the test-takers have acquired knowledge points or skills.The principally task of cognitive diagnostic assessment is to provide the diagnostic information indicating the strengths and weaknesses of each test-takers,based on the results of their responses to the test.The validity of the diagnostic information generated by cognitive diagnostic assessment depends on the accuracy of the estimated attribute mastery pattern.It is necessary to use the measurement model to accurately estimate the test-takers’attribute patterns.In order to avoid wrong remedy measures or invalid selection decisions that lead to serious consequences,the misfitting test-taking were selected by person-fit statistics to ensure the accurate measurement of the test-takers’performance.By analyzing the test data,the CDA can provide information about test-takers’mastery of knowledge,and provide reference for teachers to carry out accurate“Differential teaching”and students to carry out targeted“Remedial learning”,as a result,more and more attention has been paid.However,this diagnostic information is only useful if it is valid.The study of person-fit in cognitive diagnosis is to assess the fit of test-takers’item responses to a postulated psychometric model.The analysis based on the model and the conclusions drawn from it are reliable and valid only when the cognitive diagnostic model is fitted to the examinee’s response pattern.In order to optimize the use of diagnostic results,it is very important to develop new methods or improve existing techniques for person-fit.Based on this,this study puts forward a new person-fit index of weighted residuals and applies it to the person-fit analysis in cognitive diagnosis evaluation,discusses the validity of the R index,and carries out three sub-studies.Study 1 puts forward a new person-fit index which based on the residual to determine the critical value and distribution characteristics of R.The critical value results of different experimental conditions were applied to the second study.Study 2 tests a class of type I error and detection rates of R index under different experimental conditions in the cognitive diagnostic model,and it is compared with l_z、RCI.The effects of item discrimination,item number and attribute number on the fitting index are discussed.Study 3 is the empirical data application research.The practicability of the index is demonstrated.The findings suggest:(1)The R index distribution tends to normal distribution.(2)The type I error of R index is relatively stable.With the increase of the discrimination of items and the increase of the number of items,the power of R index has improved,and the performance of R index is the most robust when the discrimination of item is low.The R index has a high power for the types of abnormal behavior such as creative responding behavior,random guessing responses and cheating with randomness.(3)In the empirical data research,the abnormal detection rate is 4.29%,and the response pattern of typical students was analyzed.