Use Iterative Latent Class Analysis Method For Cognitive Diagnosis Evaluation: Simulation And Empirical Research

CDMs provide more detailed and multi-dimensional diagnostic feedback information on a set of attributes of subjects,reflecting their understanding and mastery of knowledge and skills,and can also reflect the different subjects' ability level.Attribute estimation is one of the main tasks of CDMs.In recent years,CDMs have developed many models and methods in attribute estimation and have been put into use successively,such as GDINA model,but the work focused on improving the accuracy of its attribute estimation needs to be further refined.Therefore,this study explores the performance of a cognitive diagnostic evaluation based on ILCA approach in attribute estimation.In this article,a combination of simulation and empirical research is used to further discuss the performance of ILCA based cognitive diagnostic evaluation in the accuracy of attribute estimation.ILCA is used for attribute estimation of empirical data.The first simulation study explored the accuracy of the attribute estimation of the ILCA method by considering the five factors,including the data generation model,the fitted model,the number of attributes,the sample size,and the attribute correlation level.Attribute estimation performance was compared.The results of the simulation study showed that: the changes in the number of attributes,the sample size,and the level of attribute correlation will affect the accuracy of the ILCA method's attribute estimation;in most cases,the accuracy of the ILCA method's attribute estimation depends on the mastery of marginal attributes.Estimation is better than GDINA model attribute estimation accuracy.The main purpose of the second simulation study was to investigate the performance of the ILCA under the Q matrix misspecification conditions.Five factors were considered: the data generation model,the fitted model,the number of attributes,the sample size,and the level of attribute correlation.And compare it with the performance of attribute estimation produced by the GDINA model.The simulation results show that the incorrect setting of the Q matrix affected the performance of the ILCA method.The results of the simulation study showed that: the changes in the number of attributes,the sample size,and the level of attribute correlation will affect the accuracy of the ILCA method's attribute estimation;in most cases,the accuracy of the ILCA method's attribute estimation depends on the mastery of marginal attributes.Estimation is better than GDINA model attribute estimation accuracy.Research three is empirical research.First of all,through self-made questionnaires of cognitive diagnosis of physical and magnetic phenomena,testing freshman freshmen in two high schools in Shandong Province to collect empirical data needed for research;Secondly,the cognitive diagnosis evaluation of the empirical data is performed using the ILCA method and the GDINA model respectively.Finally,the accuracy of the attribute estimation between the two is compared.Empirical research shows that the cognitive diagnosis and attribute estimation of empirical data prove that the ILCA method has practical application value,so it can be used to generate score reports.