The reliability of attentional biases for gambling-related images in free-viewing eye-tracking paradigms

•Free-viewing eye-tracking paradigms can reliably measure attentional bias for gambling images.•Both two- and four-image displays can reliably measure gambling-related attentional bias.•Attentional bias indices capturing the full trial (rather than first fixation) are most reliable.•Attentional bias difference scores for gambling images are reliable. Attentional biases for gambling-related stimuli are a robust correlate of problem gambling. Free-viewing eye-tracking paradigms are considered the gold standard for measuring attentional bias in addiction research, but their reliability in measuring biases for gambling-related stimuli remains unclear. Using secondary data from two different free-viewing eye-tracking paradigms (two-image and four-image displays), this study examined the internal consistency of fixation indices in samples with varying degrees of gambling involvement and problem gambling risk. Cronbach’s alpha, McDonald’s omega, and split-half reliability coefficients were used to assess internal consistency of several fixation indices (total dwell time, total dwell time percent, total fixation count, time of first fixation) for gambling-related images, neutral images, and computed attentional bias difference scores. For both two- and four-image displays, internal consistency estimates for total dwell time, total dwell time percent, and total fixation count were good to excellent for gambling-related images, neutral images, and attentional bias scores. Only time of first fixation exhibited low internal consistency. These findings indicate that both two-image and four-image free-viewing eye-tracking paradigms can reliably measure attentional biases for gambling-related stimuli among participants reporting varying degrees of gambling involvement and problem gambling risk.