Developmental Change in 4- and 5-year-olds’ Face-Trait Predictions

• In Event: 3-124 - The Specificity of Face-Trait Inferences in Infants and Children

Integrative Statement

Adults predict other’s future behavior from their facial features (i.e., thin-slice behavior) (Todorov, Pakrashi, & Oosterhof, 2009). Even young children associate particular facial features with specific character traits (Cogsdill, Todorov, Spelke, & Banaji, 2014). However, young children’s ability to use facial features to make predictive decisions is unclear. Here, we explored whether preschoolers use facial features to make predictions about knowledge. We measured children’s predictions about trustworthy and competent faces to explore whether they are more likely to infer knowledge based on broadly positive traits (i.e., trustworthiness) or on more specific traits associated with knowledge (i.e., competence).

Thirty-two preschoolers (M age = 4.75 years; SD = 6 months; 16 girls) participated in eight test trials in either a competence or trustworthiness condition. Participants were introduced to two puppets with computer-generated faces manipulated on competence or trustworthiness, (Figure 1). On each test trial, participants were asked to predict which of the two puppets would know the name of a novel object. After making a choice, participants heard the (novel) label each puppet provided for that object. Then, children were asked to decide which puppet provided the correct name for the object, and to recall the correct name for the object.

Across test trials, children made eight predictions about which puppet would know the correct name for each novel object. Interestingly, there was an interaction of age (4-years versus 5-years) by condition (competence versus trustworthiness) for these eight predictions, F(1, 28) = 5.951, p = .021, d = .891 (Figure 2). Specifically, four-year-olds in the trustworthiness condition selected the trustworthy puppet (M = 4.667, SD = 2.062) more often than those in the competence condition selected the competent puppet (M = 2.286, SD = 1.80), t(14) = -2.418, p = 0.030, d = 1.231. Five-year-olds, on the other hand, were no more likely to choose the trustworthy puppet than they were the competent puppet, t(14) = 0.940, p = 0.363. To determine whether children’s selection of a particular puppet changed over the course of the eight test trials, we examined the test trials as two blocks of four trials each. Interestingly, the age by condition interaction only remained significant in the first block of trials, F(1, 28) = 5.125, p = .032, d = 0.827.

This study suggests that preschoolers make predictions about future behavior based on facial appearance. The age-related change in children’s evaluations of the puppets indicates a developmental trajectory in which children begin to show domain specificity. When first encountering the puppets, 4-year-olds prefer to seek information from the individual with generally positive features, making the trustworthy puppet is more appealing. Five-year-olds, however, seem to take both competence and trustworthiness into account. This is consistent with prior research indicating that preschool-age children show a “halo effect” when making inferences about others (Brosseau-Liard & Birch, 2010). However, in later trials, once children have received additional information about the puppets (i.e., the object labels they provide on each trial), facial features become less important to children’s predictions.

Authors

• Erika DeAngelis, Amherst College
  Presenting Author

• Carolyn Palmquist, Amherst College
  Non-Presenting Author

• Erik W Cheries, University of Massachusets at Amherst
  Non-Presenting Author