Poster #195 - Preschoolers’ Perception of Confidence and Reliability as Indicators of Knowledge

• In Event: 3-053 - Poster Session 10
  In Poster Session: PS 10 Section - Social Cognition

Integrative Statement

Selective social learning is a useful skill that enables people to endorse one informant’s testimony over another based on the informant’s observed characteristics. This skill has been shown to be present at a very young age and to improve as children get older (Mills, 2013). Some of the informant’s characteristics or indicators can be defined as epistemic (they pertain to knowledge). Past studies have shown that children distinguish speakers based on epistemic cues and prefer to learn from people that are confident (e.g., Birch, Akmal & Frampton, 2010) and accurate (e.g., Koenig, Clément & Harris, 2004). Although children are capable of using this skill, it is still unknown how children understand these cues and what cognitive mechanisms underlie their decision when choosing from whom to learn. More precisely, even though researchers consider these cues to be epistemic, we do not know whether children actually make inferences about an informant’s knowledge based on these cues.

In the present study we aim to investigate the inferences children make about informants that differ in terms of confidence and accuracy. Twenty-three children of ages 4 to 5 were randomly assigned to one of two conditions: they were either shown videos of two individuals who differed in confidence only or who differed in confidence and reliability. In the second condition, confidence and reliability cues were intended to conflict: if an informant was showing confidence, they were also shown to be unreliable and vice versa. Children were then presented with two tasks. In the first task, children were shown labels of familiar objects and had to identify which of the two informants wrote each label. Assuming that children interpret confidence as indicative of knowledge, it is expected that they would identify the confident informant as the one writing the correct labels and the hesitant informant as the one writing incorrect labels. The results indeed showed that children in the Confident-Only condition believed the confident informant wrote the correct labels (M= 2.70, SD= 0.95, p= .045) and that the hesitant informant wrote the incorrect labels (M= 2.70, SD= 0.95, p= .045). Interestingly, children responded at chance when confidence and accuracy cues were conflicting (all ps≥ .31).

In the second task, children were asked explicit questions about the knowledge or other attributes (e.g., talents, moral behaviors) of the informants. Children in the Confident-Only condition did not attribute more knowledge or other traits to the confident individual. However, children in the Confident and Accuracy condition attributed more knowledge (M= 1.69, SD= 2.69, p= .043), but not other attributes, to the accurate but hesitant informant. These results indicate that children make some inferences about an individual’s knowledge based on both confidence and accuracy. More research will be needed to understand why children make these inferences and what impact they have on children’s decisions when choosing an informant to rely on.

Authors

• Isabelle Cossette, University of Ottawa
  Presenting Author

• Gladys Ayson, University of Ottawa
  Non-Presenting Author

• Patricia Brosseau-Liard, University of Ottawa
  Non-Presenting Author