Negative Binomial Models for Visual Fixation Counts on Test Items

Abstract

With the development of technology-enhanced learning platforms, eye-tracking biometric indicators can be recorded simultaneously with students item responses. In the current study, visual fixation, an essential eye-tracking indicator, is modeled to reflect the degree of test engagement when a test taker solves a set of test questions. Three negative binomial regression models are proposed for modeling visual fixation counts of test takers solving a set of items. These models follow a similar structure to the lognormal response time model and the two-parameter logistic item response model. The proposed modeling structures include individualized latent person parameters reflecting the level of engagement of each test taker and two item parameters indicating the visual attention intensity and discriminating power of each test item. A Markov chain Monte Carlo estimation method is implemented for parameter estimation. Real data are fitted to the three proposed models, and the results are discussed.

Assessing Preknowledge Cheating via Innovative Measures: A Multiple-Group Analysis of Jointly Modeling Item Responses, Response Times, and Visual Fixation Counts

“…VFCs are fitted using a negative binomial fixation (NBF) model proposed by Man and Harring (2019), which describes the relation between observed VFCs and latent test visual engagement. 1 The NBF model is specified as:…”

“…The rapid shift toward a digitally reliant society necessitates the continual updating of the environments in which students learn. To improve the effectiveness and efficiency of learning, technology-enhanced learning infrastructures and approaches, such as artificial intelligence–enhanced and virtual reality–based learning, are called with greater frequency to bridge this gap (see, e.g., Ercikan & Pellegrino, 2017; Hao et al, 2016; Jiao & Lissitz, 2018; Man & Harring, 2019; Mislevy, 2011). In the technology-enhanced learning system (TELS), students’ learning status regarding biological and psychological reactions are continuously recorded in a formative manner via integrated detectors (e.g., eye-tracker, motion detectors, and virtual reality goggles) within the environment.…”

How Do Test Takers Interact With Simulation-Based Tasks? A Response-Time Perspective

“…It is likely that simulation-based tasks are more interesting and engaging to test takers, but more research with different datasets—in terms of different tasks, different test-taker populations, different test conditions, and so on—is needed to further investigate the benefit of using simulation-based tasks in various settings. It may also be useful to retrieve more fine-grained data at the action level, including timing, processes, and others, to look further into person-task interactions (see, e.g., Ercikan and Pellegrino, 2017; Man and Harring, 2019). In any case, it will be valuable to conduct a systematic RT study similar to the one presented herein to assess different timing aspects of a test-taker sample’s behavioral features in the simulation-based task of interest.…”

Longitudinal Learning Diagnosis: Minireview and Future Research Directions

“…Only outcome data (or item response accuracy) were considered in most current longitudinal LDMs. Although a few studies have incorporated item response time into current models (e.g., Wang et al, 2018b ), future studies may attempt to introduce other types of process data (e.g., number of trial and error and operation process), or even biometric data (e.g., eye-tracking data; Man and Harring, 2019 ). Utilizing multimodal data can evaluate the growth of students in multiple aspects, which is conducive to a more comprehensive understanding of the development of students (Zhan, 2019 ).…”