A possible correction of the face inversion effect: A methodological commentary

The face inversion effect is based on 2 phenomena: Recognition of faces presented upright (hair above, chin below) is better than that of faces presented inverted (chin above, hair below), and inversion impairs recognition of faces more than of nonfacial objects (buildings, cars). Based on theoretical and empirical considerations, this article suggests that in many cases the size of the second phenomenon is overestimated, because it is influenced by recognition of upright faces being higher than recognition of upright objects (i.e., a baseline-level effect). A mathematically based correction is suggested for the baseline-level effect. It rests on the assumption that recognition of inverted faces depends on recognition of upright faces. Theoretical and empirical support is offered for this assumption. Also, an empirical base for the proposed correction is provided.