Attentional biases in low-risk and high-risk gamblers and the moderating effect of daily psychosocial stress

Abstract

Background and aims

Psychosocial stress is associated with gambling urges and behavior. A large body of evidence also indicates that individuals who are at risk for disordered gambling are susceptible to attentional biases for gambling-related stimuli. Yet the extent to which psychosocial stress moderates these biases has not been fully investigated. The present study examined the effect of psychosocial stress on attentional biases for gambling stimuli in regular gamblers.

Methods

The sample was 87 undergraduate students who were categorized according to gambling severity scores (no-risk, low-risk, moderate/high-risk). Participants viewed pairs of gambling-related and neutral images for 8 s while their eye movements were tracked and recorded to measure their attention to the images.

Results

Each group exhibited a unique pattern of attending to the images over the 8-s presentations. As predicted, the moderate/high-risk group exhibited an attentional bias for gambling-related images that sustained throughout the 8-s presentations. Psychosocial stress was found to moderate attentional biases only for the low-risk group.

Conclusions

Psychosocial stress appears to influence attention toward gambling images only for low-risk gamblers. Our findings suggest that addressing underlying causes of stress in gamblers may be an effective method for reducing attentional biases to gambling stimuli and gambling problems and behaviors.

Keywords:

• Gambling
• stress
• attentional biases
• cue-reactivity

Disclosure statement

DM has been supported in part by the Alberta Gambling Research Institute (AGRI). DM also discloses a speaker’s honorarium from Spectrum Therapeutics. AF received a scholarship from AGRI. The other authors report no conflicts of interest.

Notes

1 The full set of images is available from the corresponding author.

2 Three participants were excluded from all the analyses because they had no eye-tracking data (due to calibration problems) or poor eye-tracking data (less than 60% usable data due to tracking issues). Two participants in the no-risk group were excluded from all analyses because their scores on the Beck Depression Inventory were very high (>29), which suggested they were in a depressive episode.

3 The number of fixations to the gambling-related and neutral images were used to create a complimentary measure of AB. The sum of the fixations to the gambling-related images and to the neutral images were averaged over the 25 experimental trials, and AB was defined as the difference between these two means, calculated separately for each participant. An analysis of this bias measure yielded the same pattern of group differences. A between-groups ANOVA revealed a significant difference between the groups, F(2, 84) = 12.775, p < 0.001, partial η² = 0.23, and t-tests showed that the bias of the moderate/high-risk participants (M = 3.04) was significantly different than the bias of the no-risk participants (M = −1.65), t(61) = 4.89, p < 0.001, d = 1.26, and the bias of the low-risk participants (M = 0.57), t(47) = 2.31, p = 0.025, d = 0.66. The no-risk and low-risk participants also differed significantly, t(60) = 2.49, p = 0.015, d = 0.65.

4 We also analyzed overall AB using multiple regression, treating the PGSI scores as a continuous variable and using BDI scores as a covariate (as the BDI scores of the moderate/high-risk group were higher than those of the no-risk and low-risk groups). In this analysis the PGSI scores accounted for a significant percentage of the variance in AB (R = 0.521, R² = 0.272, adjusted R² = 0.254), F(2, 84) = 15.67, p < 0.001. The partial correlation between PGSI scores and AB, controlling for BDI scores, was .43 (p < 0.001).