Children calibrate their willingness to wait for rewards on the basis of time-interval experience

• In Event: Poster Session 01
  In Poster Session: PS 01 Topic - Cognitive Processes

Abstract

Delay of gratification requires children to forgo an immediate reward in order to obtain a more desirable reward in the future (Mischel et al.,1989). Failing to wait has classically been attributed to a lack of self-regulatory ability in children (Shoda et al. 1990). However, recent research suggests that children may engage in rational decision making about the relative utility of waiting, such as by assessing the trustworthiness or reliability of an experimenter to deliver promised rewards (Michaelson & Munakata, 2016; Kidd et al., 2013). Research on adults' intertemporal decision making amid uncertainty has found that adults are more likely to wait for a reward when the distribution of possible delays has a well-defined time scale, compared to when delays are more unpredictable (McGuire & Kable, 2012). We therefore investigated whether children would also calibrate their waiting decisions according to the statistical distribution of delays they had previously encountered in the same environment.

We adapted a task from McGuire and Kable (2012) in order to investigate waiting decisions in 5-to-7-year-old children. Children (n=46, M=6.54 years, SD=0.87, 21 girls) participated online via videoconferencing. Children were introduced to two characters and told that they would be collecting rewards for the characters by waiting at designated “reward spots”. Children were told that the amount of time they had to wait at each reward spot varied, and that their goal was to collect as many rewards as possible before the character arrived at the finish line (Figure 1). Children had to decide whether to continue to wait for the reward or end the trial early in hopes of the next reward spot having a shorter delay. Children completed two five-minute test blocks: a high persistence (HP) block, in which delays were uniformly distributed and lasted 2 to 16 s per trial, and a limited persistence (LP) block, in which delays had a heavy-tailed distribution and lasted 0.3 to 32 s per trial. Block order was counterbalanced across participants, and all participants received identical instructions.

A Kaplan-Meier survival curve was used to assess the time each child waited before quitting, accounting for the fact that observed waiting times were censored when rewards were delivered. Analyses were restricted to an interval of 0-16 s to account for the different ranges of delays in the HP and LP conditions. Results (Figure 2a) showed that children were willing to wait longer in the HP condition than in the LP condition, with the strongest effects appearing in the first block (between subjects p< 0.01, Wilcoxon rank sum test). In the HP condition, children become more willing to wait as the test block went on, whereas the opposite trend was observed in the LP condition (Figure 2b). Children’s age was not related to their willingness to wait (all p>.2). These results are consistent with previous findings in adults (McGuire & Kable, 2012), and suggest that children are able to use environmental information to predict remaining delay time, and then employ this information to calibrate their waiting behaviors.

Authors

• Shiba Mina Esfand, Boston University
  Presenting Author

• Joseph T McGuire, Boston University
  Non-Presenting Author

• Melissa M. Kibbe, Boston University
  Non-Presenting Author