Articles to 2020-07-26

Zum Seitenende      Übersicht Artikel      Home & Impressum

First the link to this week's complete list as HTML and as PDF.


The R factor is not directly deductible from the infection data but only through the independent determination of the time between infections, the generation time or serial interval. As Ali et al. have now demonstrated, this interval is not only imprecisely known but also widely and rapidly variable, making all stated R values an ambiguous nonsense. R is a necessary and meaningful parameter in sophisticated modeling, for the purpose of descriptive statistics, i.e. looking at and describing what is currently going on, it is useless and meaningless. But that should long have been obvious anyway.


Gibson et al. is yet another example for a bold claim in a glaring headline. Yes, there is a tiny difference between the averages, utterly dwarfed by in-group variability but statistically significant all the same. This very small distinction is an order of magnitude below those derived from house age, house value, or year of testing. All of these distribute unequally between the two groups but contrary to many others, Gibson et al. don't hide the fact but list it fully in their supplementary table S5. From that none of the above is sufficient to explain the end result alone, but they do demonstrate it not to stand out from the confounder noise. There may just be a tiny residual effect but nothing to warrant the headline. A better and more correct one would be “No significant effect of water source on lead exposure in children”. But that wouldn't stir up the yellow press, would it?


I love Lubinski. His result confirms all my politically incorrect prejudices and vindicates the English usage, where contrary to the German “Geisteswissenschaft” there is no “science” in the term “humanities”. I especially enjoy the result for law and find some – by now means all – of my old school teachers correctly described here.

Zum Anfang      Übersicht Artikel      Home & Impressum

Creative Commons Attribution-Share Alike 3.0 Unported License Viewable With Any Browser Valid HTML 4.01! Valid CSS!