Automated Test Assembly For Multistage Testing With Cognitive Diagnosis And Application Research

The computer multistage adaptive test(MST)combines the advantages of paper and pencil based test(P&P)and computer-adaptive test(CAT).As CAT,MST is adaptive based on item module;As P&P,MST can meet the needs of tests developers to management test forms and to keep tests forms parallel.Cognitive diagnosis(CD)can accurately measure the student’s knowledge states and provide diagnostic information,which is conducive to student’s self-learning and teacher’s targeting and effectiveness teaching.This article firstly tries to employ automated test assembly(ATA)to achieve the MST in the application of CD(called CD-MST)via heuristic algorithms.At the same time,this paper presents a new item difficulty index of cognitive diagnosis,and the attribute reliability in CD is defined as test quantitative target.These two indicators with heuristic methods applied for CD-MST.This study adopts the Monte Carlo simulation and empirical application research to verify the scientific and effect of the proposed CD-MST based on automatic test assembly algorithm.Study 1 developed an automated test assembly algorithm for CD-MST and use Monte Carlo simulation to verify its effectiveness.Study 2 applied the automated test assembly algorithm based on CD-MST to depression assessment.The results are as follows:(1)This study proposed automatic test assembly algorithm based on CD-MST Successfully achieved automatic assembly,and fully satisfied with both statistical and non-statistical constraints;(2)The research results show that the assembled panels of CD-MST are completely parallel;(3)Applied the automatic test assembly algorithm based on CD-MST to depression assessment can ensure that the depression questionnaire has good reliability and content validity,and ensure parallelism between panels.