Cognitive Diagnosis Modelling Incorporating Response Times and Fixation Counts: Providing Comprehensive Feedback and Accurate Diagnosis

Abstract

Problem-solving behavior consists of activities carried out by respondents, which reflect complex cognitive processes that are often systematically related to one another. To measure this variety in problem-solving behavior, biometric data including visual fixation counts (FCs), an essential eye-tracking indicator, can jointly be modeled with other types of variables reflecting different facets of problem-solving. To improve the comprehensiveness and accuracy in using such multimodal data for diagnosis of problem-solving, the present study proposes a multimodal joint cognitive diagnosis model, accounting for latent attributes, processing speed, and visual engagement by including three types of data simultaneously: response accuracy, response times, and FCs. The feasibility of the proposed model is examined by using a set of simulated data. Results indicate that utilizing FCs could improve diagnostic inferences regarding concentration, along with improving the accuracy of diagnostic classification. An empirical example is conducted to illustrate an application of the proposed model.