“The game would have been better for me if…”: children’s counterfactual thinking about their own performance in a game

Abstract

In two studies, we investigated for the first time the content of children’s counterfactual thoughts about their own experiences. Results showed that the majority of children aged 8-13 were able to produce valid counterfactuals regarding an event that happened to them, despite not achieving an adult-level ability. Comparing counterfactual and prefactual thinking, in Study 1 we found that children showed the same temporal asymmetry previously found in adults: They focused on the controllable features of their experience more in prefactual than counterfactual thinking. However, in Study 2, comparing counterfactuals produced by children and adults after a task in which making errors became salient, children produced more controllable counterfactuals (modifying their own errors) than adults, who still focused on uncontrollable features (as in Study 1). These results suggest that the ability to reason counterfactually in complex and real-life situations is not yet fully developed at age 8-13 years, affecting counterfactual content.

Keywords:

• Counterfactual
• prefactual
• children
• controllability

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Here we focus on counterfactuals in which the mutation of the antecedent undoes the outcome, as most of the counterfactual research does, even if the term also includes the cases in which a change in the antecedent does not change the factual outcome (i.e. semifactuals; Goodman, 1983).

2 Data were collected during collective experimental sessions lasting about two hours, in which the individually administrated tasks were alternated with long breaks in a common area. The tasks, individually administrated, comprised: cognitive tasks (lasting about 20 minutes), self-report questionnaires (lasting about 20 minutes) and motor tasks (lasting about 15 minutes).

3 The thought validity score of the second/third thought (when generated) was consistent with the score of the first one generated in all adults (n = 4) and in 92% of the children (one child received a better score and one child received a worse score with respect to the first thought generated). As regards the controllable vs. uncontrollable coding, all adults and 71% of children who generated more than one thought were consistent in their responses (generating all controllable or all uncontrollable thoughts).

4 We followed a common practice in hierarchical regression analyses, in which demographic variables are usually entered first, and then substantive predictors are entered to assess their effects, above and beyond the demographics.

5 The thought validity score of the second/third thought (when generated) was consistent with the score of the first one in all adults (n = 21) and in 80% of the children (only three children received a better score for the second thought, whereas the other eight children received a worse score). As regards the controllable vs. uncontrollable coding, 62% of adults and 84% of children who generated more than one thought were consistent in their responses.

6 Note that we cannot consider the score obtained on our adapted trail making test as a measure of cognitive flexibility given that the path, the instruction and the procedure were modified for the aim of the study.