Children rationally adjust cognitive effort based on task difficulty and reward probability

• In Event: Poster Session 04
  In Poster Session: PS 04 Topic - Cognitive Processes

Abstract

Like physical effort, cognitive effort is costly. Young children, however, are famous for being more optimistic in their abilities (Dweck, 2002) and more generous in spending their cognitive effort (Gopnik et al., 2017). Adults, on the other hand, are able to rationally adjust cognitive effort based on its utility (e.g., [how much effort required] * [how much reward it yields]; Westbrook & Braver, 2015). Are children sensitive to their cognitive effort? Recent research showed that 7- to 12-year-old students put more value on tasks that required more cognitive effort (Chevalier, 2018). However, little is known about younger children’s sensitivity to cognitive effort. Can young children rationally adjust their cognitive effort based on its utility?

Here we asked whether children rationally adjusted their cognitive effort based on task difficulty and reward probability. Four- to 5-year-old children played a computer game, where they saw pairs of dot arrays that differed in number and had to count and decide which set had more dots. Children were told that if they answered correctly, they would spin a fortune wheel and have a chance to get a star. The difficulty of the trials varied based on the number of dots. The larger the number, the more difficult the trial. There were four trials in each block. The probability of getting a star from the fortune wheel varied from block to block in a predetermined pseudo-random order (25%, 50%, 75%, or 100%). Before each trial, children were first asked whether they wanted to count and figure out the answer to the current trial, or to skip to see a different trial. The order of the correct side was counterbalanced within subjects. The order of the instructions asking children to count or skip, and the order of trial blocks (varying in reward probability) were counterbalanced between subjects.

Children (N=20; Mean Age = 5.05 years) overall performed better on the easier trials (Figure 1a), p < .001, and the probability of reward did not affect their overall accuracy. Critically to our main hypothesis, children were more likely to skip a problem when it was more difficult, p = .006. And this effect was particularly prevalent when the probability of reward was lower, p = .044 (Figure 1b). These results converge with other findings that revealed sensitivity to cognitive effort in young children. Importantly, we extend this past work by showing that children as young as five years adjust their decisions to engage in such effort using utility estimations (based on the likelihood of reward). We will discuss how these precocious abilities are related to resource rational models of cognition, specifying the integrated factors of cognitive effort utility in reasoning, learning, and decision making.

Authors

• Jinjing Jenny Wang, Rutgers University - New Brunswick
  Presenting Author

• Alejandra Reinoso, Northwestern University
  Non-Presenting Author

• Elizabeth Bonawitz, Rutgers University - Newark
  Non-Presenting Author