Skip to main content
Fig. 3 | International Journal of Behavioral Nutrition and Physical Activity

Fig. 3

From: Strength exercises during physical education classes in secondary schools improve body composition: a cluster randomized controlled trial

Fig. 3

Effect of the intervention on physical activity. Physical activity measured by accelerometers: Observed data, possibly biased by missingness (left panel), and predicted means based on the mixed regression (right panel). Note that observed means and SDs can be biased by ignoring both the clustering and dropout/missingness. The predicted means based on mixed regression are the best estimates of the time courses of all outcomes

Back to article page