By Marine Rougier

This research received an IPA from the Journal of Experimental Social Psychology.

Approach and avoidance are among the most fundamental and adaptive actions toward environmental aspects. Research on approach/avoidance training (AAT) effects shows that approaching stimuli (i.e., reducing the distance between the self and a stimulus) leads to more positive evaluations or representations of stimuli than avoiding stimuli (i.e., increasing the distance between the self and a stimulus). So far, however, AAT effects have been considered without any further specification of the meaning of approach and avoidance actions. Yet, approach behaviors can sometimes convey a radically different meaning, one being positive (when approach means affiliation) and one being negative (when approach means aggression).

The present experiments rely on a grounded cognition approach to extend AAT effects by investigating their framing-dependency, that is, whether AAT effects can vary as a function of how approach is framed. We tested this effect on facial representations of target stimuli with the reverse correlation paradigm. In a Preliminary Experiment, using antagonistic types of approach (affiliative vs. aggressive), we found that approach led to more positive facial representations than avoidance when approach was portrayed as affiliative, but this effect decreased and tended to reverse (i.e., yielding more negative facial representations) when approach was portrayed as aggressive. Two registered experiments extended these results while also addressing important limitations of the Preliminary Experiment.

First, to prevent any contrast emerging from the joint use of approach and avoidance, Experiment 1 isolated the unique effects of affiliative approach, aggressive approach, and avoidance when compared to a control action. We also explored whether aggressive approach and avoidance (two negative yet distinct actions) produced negative effects characterized by divergent outcomes on facial features. In line with the framing-dependency hypothesis, we observed that facial representations in the affiliative approach condition were evaluated more positively than the ones in the aggressive approach condition. Moreover, facial representations in the affiliative and aggressive approach considered jointly received more positive ratings than classification images in the avoidance condition. The results regarding a differential effect of aggressive approach and avoidance on traits pertaining to aggressive approach vs. avoidance did not emerge.

Second, Experiment 2 tested the importance of the experiential component of approach/avoidance actions by comparing the AAT with a mere instructions condition. We replicated the difference between affiliative vs. aggressive approach for participants in the AAT condition. Crucially, this effect was larger in the AAT condition as compared to the mere instructions condition. These results are in line with the idea that the AAT effect is framing-dependent and that this effect hinges on the experiential aspect of the training.

Results of Experiments 1 and 2 are consistent with a framing-dependency of AAT effects. Unveiling the framing-dependency of AAT effects challenges some of the current theoretical views on AAT effects. In empirically demonstrating the importance of the framing in this evaluative learning context, this research project perfectly fits in the research activities of our research group.