ABSTRACT
This article presents a meta-analysis of the impact of minimal exposure to affective stimuli on the emergence of enduring conscious affective feelings. Theories often assume that such affective feelings are linked to automatic appraisals of events (i.e. in the absence of an evaluative processing goal). However, few studies have tested this hypothesis. Moreover, they have provided divergent results. We propose a meta-analysis of these studies to get a clearer picture on this issue. The meta-analysis includes 22 studies (37 effect sizes; combined N = 2159) in which participants were repeatedly exposed to affective stimuli in the absence of an evaluative processing goal before their mood was measured. In this analysis, we focused on the type of stimulus presentation (i.e. visible vs. masked) as well as on the type of stimulus (i.e. faces, pictures, words). The results indicate that the effect of a stimulus is moderated by the visibility of the stimuli. Repeated exposure to visible stimuli leads to congruency effects (i.e. positive stimuli lead to positive feelings), whereas exposure to masked stimuli leads to contrast effects (i.e. positive stimuli lead to negative feelings). Moreover, these effects seem to be restricted to some types of stimuli, with no detectable effects of emotional faces.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 “AND” refers to Boolean search operators.
2 In this meta-analysis, we were interested in the impact of minimal exposure to affective stimuli on the emergence of affective feelings. Therefore, we did not include studies in which participants were instructed to process such stimuli with an evaluative (or feeling) goal in mind, as such situations cannot be characterised as minimal exposure. For the same reason, we did not include within-participants designs because these designs involve repetitions of the measure. If the measures are repeated, participants are likely to process the stimuli after the first measure of affective feelings with such a processing goal in mind.
3 Included articles are preceded by “*” in the references list.
4 The type of measures used can potentially play a role in the emergence of these effects. However, given the diversity of measures used in the studies included in this meta-analysis, we were unable to test this potential moderating role.
5 In addition, funnel plots showing the effects as ds versus their respective standard errors are presented in Supplemental Materials: Figure S1 for the meta-analysis on studies with visible stimuli and in Figure S2 for the meta-analysis on studies with masked stimuli.
6 A common method to detect outliers is to define a study as an outlier if the confidence interval of a study does not overlap with the confidence interval of the pooled effect (Viechtbauer & Cheung, Citation2010). One outlier was identified. When the outlier was excluded, the results remained unchanged.
7 We identified one outlier. The results remained similar when this outlier was excluded.
8 We identified one outlier. The results remained similar when this outlier was excluded.
9 We identified one outlier. The results remained similar when this outlier was excluded.