Effects of personal pronouns on children's numerical problem solving

Personal pronouns such as 'you' can be used to elicit self-referencing, whereby information is processed with reference to self rather than other people or contexts. Self-referencing is associated with robust attentional biases like attention capture and increased attentional resource availability, which may have implications for numerical processing. Previous research suggests that performance on numerical problems (e.g., "Tom has 3 balls. Bob has 2 more (less) balls than Tom. How many balls does Bob have?") can be improved when one character in the problem is replaced with the pronoun 'you' (D'Ailly et al., 1997). However, evidence for this effect is limited, and is complicated by confounds between problem difficulty and position of the self-pronoun in the extant research. Further, there has not been a test of the most parsimonious explanation: that self-pronouns reduce the working memory load of the problem simply by reducing the list of new characters to be tracked. To address these issues, we conducted two experiments in which school children solved numerical word problems, half of which included the self-pronoun 'you'. In Exp 1, children aged 7-11 years (N = 96) were asked to complete addition or subtraction problems in which the self-pronoun was positioned either first or not first in the problem. Children were required to track either multiple characters (e.g., "you have 3 balls. Bob has 2 more...") or one character but multiple objects (e.g., "you have 3 balls and 2 bats...'') to solve the problem. Accuracy data showed a reliable advantage for problems that included the self-pronoun, and children were also faster to solve these problems when they involved addition, although were slower when the question involved the more difficult subtraction operation. The position of the self-pronoun had no impact on performance, and the effect of self was consistent regardless of whether the problem included multiple characters or only one. This suggests that self-pronouns have a facilitative effect on problem solving regardless of whether or not there is a list of characters to hold in working memory. To explore the effects of difficulty further, in Exp 2, 9-11-year-old children (N = 144) were asked to solve addition and subtraction problems that were either low or high in difficulty, half of which included a self-pronoun. As in Exp. 1, accuracy was significantly higher when a self-pronoun was included in the problem, although this effect was greater in the most difficult questions. Response time data showed that self-referent questions were answered more quickly than problems without a self-pronoun regardless of difficulty. Together, these findings suggest that self-pronouns have a reliable enhancing effect on children's numerical problem-solving. The effect of self-pronouns on accuracy is greatest when the difficulty of the problems is highest, although the effect on response time is more complicated. These data suggest that incorporating self-pronouns in children's numerical problems would be good practice in schools, reducing some of the processing or load barriers children may experience when learning to solve these problems.