- Letter to the editor
- Open Access
Response to Farrokhi et al.’s statistical comments on ‘no seasonal variation in physical activity of Han Chinese living in Beijing’
International Journal of Behavioral Nutrition and Physical Activity volume 14, Article number: 152 (2017)
We have previously shown that physical activity patterns of Han Chinese adults living in Beijing were relatively invariant over the course of a single year , despite enormous differences in ambient temperature over the annual cycle. In addition we also indicated that BMI and body composition of the same subjects was also not significantly changed over the course of the year long measurements. We are grateful to Farrokhi et al.  for pointing out an error in our analysis in this latter section of the paper, and welcome the opportunity to amend this analysis in the light of their constructive comments. As noted by Farrokhi et al.  in our analysis of the effects of time on body composition parameters we used a simple one-way ANOVA that did not account for the fact the measurements were repeated in the same individuals and hence not independent. When we repeat the analysis as a General Linear Model including individual ID as a random factor, then it turns out that there are indeed significant temporal effects on the body composition measurements (Body weight: FMonth (5195) = 0.66, p = 0.024, FID (33,195) = 247.19, p < 0.001; BF%: FMonth (5195) = 11.49, p < 0.001, FID (33,195) = 229.50, p < 0.001; FM: FMonth (5195) = 9.15, p < 0.001, FID (33,195) = 159.63, p < 0.001; FFM: FMonth (5195) = 6.44, p < 0.001, FID (33,195) = 524.33, p < 0.001). These effects are illustrated in Fig. 1. The body weight in January was significantly lower than in November, but the other months did not differ significantly from each other. For body fat percentage and fat mass, the value in November was greater than in the other months. Fat-free mass was higher in the summer months compared to the winter (Fig. 1d).
We note from these revised analyses that for the body composition measurements we made the main effect was that the individuals were significantly fatter and heavier later in the study (November) compared with earlier (January). Because our measurements spanned only a single year we cannot distinguish whether this was a seasonal effect, or because the individuals were all getting older as the study progressed. In contrast fat-free mass did not progressively increase but was higher both at the start and end, but lower in the summer period in the middle. This may be a true seasonal effect, but additional years of study would be required to confirm that. Previous studies have shown adult individuals in the USA gain around 0.5 kg each year as they age  and the magnitude of this effect is consistent with the change we observed in weight and fatness. We emphasise that the other statistical analyses in the original paper concerning physical activity levels did account for repeated measurements, and these analyses, and the conclusions drawn from them, are unaffected by this error.
Wang G, Li BG, Zhang XY, Niu CQ, Li JB, Li L, Speakman JR. No seasonal variation in physical activity of Han Chinese living in Beijing. Int J Behav Nutr Phys Act. 2017;14:48.
Farrokhi M, Shiriyan N, Arjaki D. Letter to the editor: statistical comment on “no seasonal variation in physical activity of Han Chinese living in Beijing”. Int J Behav Nutr Phys Act. 2017;
Van Wye G, Dublin JA, Blair SN, DiPietro L. Adult obesity does not predict 6-year weight gain in men: the aerobics center longitudinal study. Obesity. 2007;15:1571–7.
The authors declare that they have no competing interests.
About this article
Cite this article
Wang, G., Speakman, J.R. Response to Farrokhi et al.’s statistical comments on ‘no seasonal variation in physical activity of Han Chinese living in Beijing’. Int J Behav Nutr Phys Act 14, 152 (2017) doi:10.1186/s12966-017-0603-y