ESCON 2022

Abstract

Approach/avoidance is a basic behavior of organisms and is an important part of how they respond to stimuli in their environment. Thus, being able to capture—at list a part of—people’s tendency to approach or to avoid stimuli would be an important scientific advancement, both in terms of the underlying mechanisms comprehension and in terms of applied research. To measure people’s approach/avoidance tendencies toward stimuli researchers developed computerized tasks (e.g., Chen & Bargh, 1999). Often, in these tasks participants have either to approach positive stimuli and to avoid negative ones—a condition in which they are fast—or to avoid positive stimuli and to approach negative ones—a condition in which they are slower. We label this kind of task “evaluative tasks”, because participants have first to evaluate whether the stimulus is positive or negative before deciding to perform an approach or avoidance behavior. However, one could ask whether it is necessary that participants first evaluate the valence of stimuli to observe a facilitation or an inhibition of approach/avoidance behaviors toward positive/negative stimuli. On this issue, diverging conclusions were reached by meta-analyses (Laham et al., 2015; Phaf et al., 2014). In this work, we report new data suggesting that approach/avoidance behaviors are reactivated—and can be measured—even in a non-evaluative task.

To do so, we designed a new procedure based on the Visual Approach/Avoidance by the Self Task (Rougier et al., 2018). In this new procedure, participants were primed with a positive or negative stimulus and had to approach or to avoid depending on a neutral target (i.e., a square or a diamond). Accordingly, this procedure did not require participants to evaluate the valence of the stimulus. Yet, through seven experiments we consistently observed faster responses to approach the target after a positive prime and to avoid the target after a negative prime vs. slower responses to avoid the target after a positive prime and to approach the target after a negative prime (i.e., compatibility effect). We first tested this effect in a pilot lab experiment and in two preregistered replications (in the lab and online), yielding a meta-analytic effect size of dz = 0.66, 95% CI [0.46, 0.85]. Then, we tested the robustness of the effect to various experimental manipulations throughout four preregistered experiments.

In two experiments we diminished the evaluative salience by reducing the proportion of valenced primes to respectively 50% and 25% of the primes (see Everaert et al., 2011). We still observed compatibility effects in these two experiments. We also conducted an experiment in which we omitted the instruction to “ignore the word” presented before the geometric shape (see Duscherer et al., 2008). In this experiment, we also observed a significant compatibility effect. In a last experiment, we asked participants to keep a mental tally of prime words presented in grey ink to encourage them categorizing stimuli on their color dimension instead of their valence. As previously, we observed a significant compatibility effect in this experiment. Taken together, these experiments suggest that approach/avoidance behaviors toward stimuli can be reactivated and measured even in a non-evaluative task.