This paper examines the relationship between physical activity and individual productivity among agricultural workers paid on a piece-rate basis. In the context studied, physical activity has a clear correspondence with worker effort. Agricultural workers' physical activity is directly observed from accelerometer data and is robustly associated with their daily productivity. In addition the impact of a health intervention, which provides malaria testing and treatment, on physical activity and productivity, indicates that the increased daily productivity of workers who are offered this program is explained by worker effort reallocation from low-intensity to high-intensity work within a fixed time period. This demonstrates, in settings when individual productivity is observed, that physical activity measures can help disentangle productivity effects due to effort. When productivity is unobserved, physical activity measures may proxy for individual productivity in physically demanding tasks. The challenges and limitations of physical activity measurement using accelerometers is discussed including their potential use for alternative contexts and the importance of field and data analysis protocols.