Moving well-being well: Using machine learning to explore the relationship between physical literacy and well-being in children

Abstract

Physical literacy provides a foundation for lifelong engagement in physical activity, resulting in positive health outcomes. Direct pathways between physical literacy and health have not yet been investigated thoroughly. Associations between physical literacy and well-being in children (n = 1073, mean age 10.86 ± 1.20 years) were analysed using machine learning. Motor competence (TGMD-3 and BOT-2) and health-related fitness (PACER and plank) were assessed in the physical competence domain. Motivation (adapted-Behavioural Regulation in Exercise Questionnaire) and confidence (modified-Physical Activity Self-Efficacy Scale) were assessed in the affective domain. Well-being was measured using the KIDSCREEN-27. Accuracy of predicting well-being from physical literacy was investigated using five machine learning classifiers (decision tree, random forest, XGBoost, AdaBoost, k-nearest neighbour) in the full sample and across subgroups (sex, socioeconomic status [SES], age). XGBoost predicted well-being from physical literacy with an accuracy of 87% in the full sample. Predictive accuracy was lowest in low SES participants. Contribution of physical literacy features differed substantially across subgroups. Physical literacy predicts well-being in children but the relative contribution of physical literacy features to well-being differs substantially between subgroups.

Keywords: children; health; machine learning; physical literacy; prediction; well-being.