When my actions shape your looks: Experience-based properties of approach/avoidance bias the visual representation of others.

The literature on the approach/avoidance training (AAT) effect has focused on its evaluative consequences (with approached stimuli evaluated as more positive than avoided ones). Building on a grounded cognition framework, we investigated AAT effects on the visual representation of stimuli (here, neutral faces). We formulated specific predictions regarding the facial features that should be the most biased and the conditions under which the effect should be the strongest. We tested these predictions in five preregistered experiments using a reverse correlation paradigm. In Experiments 1-2, the facial representations resulting from an AAT looked more "approachable" and "avoidable," respectively. Specifically, we observed more bias on facial traits related to approach/avoidance (e.g., trustworthiness) than on traits less relevant for these actions (e.g., cleverness). Experiment 2 additionally tested the unique contribution of both approach and avoidance as compared to control actions. Experiments 3A-4 showed that this effect depends on experiencing the AAT (compared to the mere instructions of approach/avoidance) and on the sensory aspects of approach/avoidance that are mimicked in the AAT. Finally, Experiment 5 supported the idea that the AAT effect still emerges in the absence of explicit instructions of approach/avoidance (i.e., by only leaving the sensory aspects of approach/avoidance). The present research enriches the literature by revealing AAT effects that extend beyond mere evaluative consequences to the visual representation of the target stimuli. Our results also inform existing theoretical views on AAT effects. (PsycInfo Database Record (c) 2021 APA, all rights reserved).