Poster #72 - Preschoolers Treat Confidence as Situational Rather Than Indicating Enduring Knowledge

• In Event: Poster Session 19

Abstract

Research has widely demonstrated that children prefer to learn from confident rather than hesitant informants (Brosseau-Liard & Poulin-Dubois, 2014; Hübscher et al., 2016). However, it remains unclear how children interpret confidence cues. Some knowledge cues, such as visual access to information, are situational (i.e., informative of an individual’s current knowledge within a specific situation); others, such as one’s past semantic accuracy, are person-specific (i.e., informative of an individual’s knowledge across situations). Confidence, however, could conceivably be treated either as situational or person-specific. Most past studies have focused on the event in which confidence cues were being displayed, thus not allowing to test children’s interpretation. In two experiments, we examined 4- and 5-year-olds’ use of informants’ changing levels of confidence when learning semantic information. In both experiments, children witnessed a history phase where one informant displayed confidence whereas another one expressed hesitancy. In Experiment 1, at test, the informants either continued showing the same confidence cues (i.e., consistent condition) or switched from confident to hesitant and vice versa (i.e., inconsistent condition). In Experiment 2, during the test phases, the informants demonstrated neither confidence nor hesitancy, such that only past confidence cues were available. On one hand, if children interpret confidence cues as indicative of a durable person-specific characteristic, children’s learning during the test phases would be impacted by the informants’ past confidence cues. On the other hand, if children perceive confidence as a strictly situational indicator, they should only use the informant’s present level of confidence to decide from which informant to learn as it reflects their current certainty in that specific situation. In Experiment 1, participants (N=51, M=61 months, SD=6, 26 girls) preferred to learn from the currently confident informant over the currently hesitant one (M=3.65 out of 5 trials, SD=1.40; t(50)=5.86, p<.001, d=.82). Performance in the consistent and inconsistent conditions did not significantly differ (consistent: M=3.75, SD=1.15; inconsistent: M=3.55, SD=1.60; t(49)=.49, p=.625, ns, d=.14). This indicates that participants relied primarily on the informants’ current level of confidence (i.e., confidence cues displayed in the test phases) regardless of their prior confidence displayed in the history phase. In Experiment 2, when no confidence cues were available at test, participants (N=41, M=59 months, SD=6, 21 girls) did not prefer to learn from the previously confident informant over the previously hesitant one (M=2.51 out of 5 trials, SD=1.23; t(40)=.06, p=.475, ns, d=0.02) although a memory check confirmed that most participants remembered the past confidence cues (Nsuccess=28 (68%), p=.014). Overall, participants thus decided whom to learn from based on the informants’ current level of confidence regardless of their prior confidence (Experiment 1) and did not distinguish informants based on past confidence cues when no current confidence cues were displayed (Experiment 2). Taken together, these results suggest that 4- and 5-year-olds perceive confidence cues primarily as situational indicators of a person’s knowledge.

Authors

• Aimie-Lee Juteau, University of Ottawa
  Presenting Author

• Patricia E Brosseau-Liard, University of Ottawa