A multivariate multilevel approach to unravel the associations between individual and school factors on children's motor performance in the REACT project

Objective The aim was to (1) estimate the relationship between physical fitness (PF) and object control fundamental movement skills (FMS), (2) identify child characteristics that relate with PF and FMS, and (3) examine associations between the school environment, PF, and FMS. Methods The sample included 1014 Portuguese children aged 6-10 years from the REACT project. PF was assessed via running speed, shuttle run, standing long jump, handgrip, and the PACER test. Object control FMS were assessed with stationary dribble, kick, catch, overhand throw, and underhand roll. Test performances were transformed into z-scores, and their sum was expressed as overall PF and FMS. Child-level variables included body mass index (BMI) z-scores, accelerometer-measured sedentary time and moderate-to-vigorous physical activity, and socioeconomic status (SES). School size, physical education classes, practice areas, and equipment were also assessed. Results Approximately, 90% of the variance in object control PF and FMS was at the child level, and 10% at the school level. The correlation between PF and object control FMS was .62, which declined to .43 with the inclusion of covariates. Older, more active, and higher SES children had higher object control PF and FMS, and boys outperformed girls. BMI was negatively associated with PF but not with object control FMS. Sedentary time and number of physical education classes were not significant predictors. Most school predictors did not jointly associate with PF and object control FMS. Conclusion PF and object control FMS z-scores were moderately related. Not all child characteristics were associated with both PF and object control FMS, and their effect sizes were different. School characteristics only explained 10% of the total variation in PF and object control FMS.