Re-analysis of data from Dong et al.: [Evidence for a limit to human lifespan] (http://www.nature.com/nature/journal/vaop/ncurrent/full/nature19793.html)
In their paper, Dong et al. claim to provide statistical evidence for an absolute limit to human lifespan. Here we re-analyze the data from their central figure using model comparison. The data from the central figure of the paper provides no support for or against their model of saturation / decline relative to a simple linear increase, but supplementary data recovered from the Gerontological Research Group does provide evidence for their model (see our R report).
It would be ultimately more appropriate to model these data using Extreme Value Theory. We have done so here. While the EVA models clearly fit the data better then the linear models, they did not change the core conclusion above: the GRG data provide evidence for the authors' trend break model over a linear model. The data from Dong et al Fig 2a conversely provide support for the linear model.
In our opinion, such statistical comparisons should have been part of the Dong et al. paper.
