RESUMO
We aimed to provide evidence for an a priori hypothesis and sample size for subjectively assessing physical activity intensity and duration in paediatric population, adopting objective methods as the reference. We searched electronic databases, reference lists and author databases. Correlation coefficients were pooled as an indicator of agreement estimates. We found 183 agreement analyses (94.5% based on correlation coefficients) from 89 studies that met our inclusion criteria. We identified four physical activity parameters addressing intensity and two parameters addressing duration. The physical activity parameters focusing on intensity were measured only by questionnaires, and the best correlation was achieved by moderate-to-vigorous physical activity compared with a heart rate monitor. In addition, total physical activity duration had a stronger correlation with an accelerometer when measured by diaries or logs than when measured by questionnaires. In both cases, the correlation was moderate. Studies with sample sizes between 50 and 99 subjects showed measurements that were stable in both magnitude and interpretation. Our findings suggest that the agreement between subjective and objective methods for assessing physical activity intensity and duration is weak to moderate. Furthermore, sample sizes ranging from 50 to 99 subjects provide stable agreement estimates between methods.