decision-making in low-ses children_2021_data.xlsx

Abstract<br>Traditional perspectives on self-regulation have emphasized the importance of future-oriented decision-making in the expression of adaptive behaviors. In the present study, we examined whether decision-making varies as a function of socioeconomic status (SES) using the Children’s Gambling Task (CGT). We administered the CGT to 227 children (49% female, 48% low SES) between the ages of 5 and 7 years. After completing the CGT, children were given an awareness test to assess their knowledge of the reward/loss contingencies. Data analysis was conducted through multilevel modelling. Fluid intelligence, as measured by the Test of Nonverbal Intelligence, was included as a covariate in the analysis. Results indicated that low-SES children performed lower in the CGT than their middle/high-SES peers. Children from middle/high-SES backgrounds made more advantageous choices than the low-SES children in the last blocks of the task. No differences were found in the level of explicit understanding of the task reached by the two SES groups. Further analysis indicated that, while middle/high-SES children with higher knowledge of the game performed better on the last blocks of the task in comparison with their SES peers with no understanding, low-SES children with higher explicit knowledge did not exhibit an improvement in their decision-making strategy in comparison with their SES low-awareness counterparts. Fluid intelligence did not predict CGT performance, suggesting that SES differences were not mediated by reasoning capabilities. The finding that children from low-SES families continued exhibiting an immediate reward-oriented strategy despite being aware of deck contingencies fits with (although speculatively) the evolutionary-developmental framework.<br>Ethical requirementsWritten informed consent was obtained from parents/caregivers of the children. The study was approved by the Research Ethics Committee of School of Psychology ([blinded]) (Protocol Number: 191175, Project Title: Desarrollo de la autorregulación en la primera infancia: efectos interactivos entre la sensibilidad biológica al contexto y el nivel socioeconómico).<br><b>Statistical analyses</b>To test the study hypotheses, we conducted two-level multilevel analysis using the nlme package<b> </b>(Pinheiro et al., 2013) on the R programming platform (Team, 2013). Multilevel modeling is advantageous as it allows for the estimation of fixed effects while simultaneously accounting for the clustered or hierarchical structure of data, namely, within-cluster relations (the random effect).<br>In this study, independent variables included Block (a categorical variable with 5 levels by dividing the 60 card selections into five instances of 12 cards each (Kerr &amp; Zelazo, 2004)) as the within-subject variable (Level 1), and Sex, Age in months (centered by calculating the difference in relation to the youngest child in the sample), SES (low-SES vs. middle/high-SES), and Awareness (Level 1 vs. Level 2) as the between-subject variables (Level 2). As in Kerr &amp; Zelazo (2004), the primary dependent measure was whether children made an advantageous or disadvantageous choice on each trial. Proportion scores were used to analyze performance across the 5 blocks of 12 trials each. The proportion score of each block was calculated by subtracting the proportion of disadvantageous choices from the proportion of advantageous choices, which yielded difference scores ranging from -1 to 1. A net score above zero implied that children were selecting cards advantageously, and a net score below zero implied a disadvantageous selection.