A Meta-Analysis on the Effect of Visual Attention on Choice

Decision makers attend more to preferred choice options and to the ultimately chosen option, but does visual attention influence preferences and choice? Several theories suggest that attention has a causal effect on preferences and choice and a growing number of studies have examined the question with experimental methods. However, the evidence for an effect of attention on choice is mixed and highly contended. To advance the debate on the role of attention in decision making, we meta-analyze studies that manipulate attention-to-choice options and measure the effect on 2-alternative preferential choices. We identify 3 different methods for manipulating attention and find that studies manipulating total exposure time enhance choices the most, P = .541, 95% CI [.523, .560], p < .001, followed by studies controlling the location of the last fixation, P = .532, 95% CI [.518, .547], p < .001. Studies manipulating the location of the first fixation do not differ from chance level choice proportions, P = .507, 95% CI [0.497, 0.516], p = .18. The PET-PEESE analysis suggests a small degree of publication bias which results in a slight reduction of effect sizes. A meta-regression with absolute attention difference as predictor confirms the robustness of the findings. Our findings show the relevance of assuming an effect of attention on choice, but also indicate a need for further model development to account for the complete pattern of attention effects.