Abstract
A glance at a face, a few syllables uttered by a speaker, or the verbal description of a behaviour
are sufficient for forming a first impression. Impressions are often determined by the combination of
several modalities. Moreover, impressions are updated when additional, diagnostic information is
presented. Only a few studies have tested the integration of multi-modal cues presented sequentially.
This research aimed to extend and complement previous work on multi-modal impression updating.
In seven pretests (N = 317) and eight experiments (N = 1613) we examined whether, how,
and why multi-modal cues update first impressions based on faces, voices, and verbal descriptions of
others' behaviour. In all experiments, participants were presented with a first information concerning
the target and formed a first global (valence-based) impression. Then, they were presented with
valence-inconsistent information from a different source before providing a second impression. In
Experiments 1-3, we analysed face-to-voice and voice-to-face impression updating. We showed that
faces and voices update impressions based on cross-modal cues. In addition, voices had a greater
impact on updating impressions. In Experiments 4-7, we analysed behaviour-to-face and behaviour-to-voice updating.
We demonstrated that faces and voices update first impressions based on behaviours.
We also found that voices were on average more effective than faces, but that this effect
depended on the type of faces used (i.e., artificial or real faces) and on perceived cue attractiveness.
In Experiment 8, we provided an explanation for the higher impact of voice (vs. face) in impression
updating. Comparing face-to-voice and voice-to-face updating, we found that voices were perceived
as more diagnostic of others’ self and intentions than faces, and that diagnosticity mediated their
difference in updating.
Our results provide evidence for the involvement of multi-modal cues in impression updating
and open to further investigations on multi-modal integration.