Integrating eye-tracking into psychometric research: computational approaches to explore response burden in questionnaire completion.

INTRODUCTION

Long questionnaires can increase cognitive and motivational demands, causing response burden. This study investigates whether eye-movement behavior varies with instrument length.

METHODS

Twenty-five participants completed the BFI-44 and the longer BFQ-2 during gaze recording. Analysis included a 2 x 3 repeated-measures ANOVA (Test x Time) and the evaluation of seven machine-learning classifiers (including Random Forest and AdaBoost) trained on the final questionnaire segment.

RESULTS

Fixation, visit, glance, and pupil metrics decreased over time across both instruments. Significant Test x Time interactions revealed that the longer BFQ-2 produced more pronounced temporal variability in gaze behavior. The AdaBoost classifier achieved 73% accuracy in distinguishing between the two instruments.

DISCUSSION: Findings suggest that longer questionnaires elicit stronger temporal fluctuations in gaze, likely reflecting fatigue or strategic adaptation. Eye-tracking provides a valuable objective tool for assessing the stability of respondent engagement over time.