How Children Weight Social and Nonsocial Information When Searching for Rewards

• In Event: Active Learning Across Domains and Development: How Goals and Experience Shape Children’s Pursuit of Information

Abstract

Introduction. In making choices, children must consider multiple sources of information in order to determine an action to take. How children integrate information may be influenced by the type of information available. Social information is privileged early in life, particularly when there is high uncertainty. For example, infants were sensitive to the reliability of both social and nonsocial cues to a rewarded location; however, when the cue pointed to a never-rewarded location, infants only searched in a social condition (Tummeltshammer, et al., 2014). Therefore, social cues may preferentially direct attention in order to make predictions, whereas nonsocial cues do not achieve the same aim (Wu, et al., 2014). Less is known about how use of social and nonsocial information changes across development as children become better at tracking the reliability of others (Ronfard, et al., 2018). We examined the effect of domain (social vs. nonsocial) and cue reliability on 4- to 9-year-old children’s reward learning.

Hypotheses. We predicted that younger children would overall rely more heavily on social information when making choices than older children, but that children across the age range in our sample would privilege social information when there was a higher degree of uncertainty.

Method. Participants (N=191) searched for a reward among multiple, differentially rewarded locations on a computer. Prior to making a selection on each trial, participants saw a cue regarding the location of the reward (Figure 1). Participants were either told that the cue was a suggestion from another person or that the cue was merely a computer glitch (between-subjects). We also varied the reliability of the cue within each condition (reliable: cue indicated the reward on 90% of trials; unreliable: cue indicated reward on 10% of trials; uncertain: cue indicated reward on 50% of trials). We measured the extent to which participants relied on the cues versus the underlying reward distribution.

Results. Comparing the social and nonsocial cues, children attributed more weight to a confederate’s suggestions as compared to the nonsocial cue (F(1,179)=14.191, p<.001). Participants also distinguished between unreliable, uncertain, and reliable suggestions and increased their use of the cues consistent with their reliability (F(2,179)=232.446, p<.001). Use of both the social and nonsocial cues decreased with age (b=-0.046, F(1,179)=8.774, p=.003), and this effect was driven by the uncertain confederate (F(2,179)=4.489, p= 013; simple slopes: uncertain b=-0.046 p=.003; reliable and unreliable ps>.2; Figure 2). Finally, there was a marginal condition-by-reliability interaction (F(2,179)=2.474, p=.087) suggesting greater differences in weight attributed to social versus nonsocial suggestions when suggestions were uncertain. Consistent with this finding, a greater proportion of participants used a strategy of simply selecting the cued location when an uncertain cue was social as compared to (X2(1)=4.468, p=.034; no difference in the use of this strategy when the cue was reliable or unreliable, ps>.05).

Conclusion. Children were sensitive to the reliability of the cues, yet social information was privileged under uncertainty. This research makes progress in exploring the interplay between the multiple sources of information children must weight when making choices.

Authors

• Rista Plate, University of Pennsylvania
  Presenting Author

• Kristin Shutts, University of Wisconsin - Madison
  Non-Presenting Author

• Aaron Cochrane, University of Wisconsin-Madison
  Non-Presenting Author

• C. Shawn Green, University of Wisconsin-Madison
  Non-Presenting Author

• Seth David Pollak, University of Wisconsin - Madison
  Non-Presenting Author