Zum Seitenende Übersicht Artikel Home & Impressum
First the link to this week's complete list as HTML and as PDF.
***
As if there weren't already enough, Lelieveld et al. offer another example of epidemiology gone mad. All they have to offer is a correlation with a tiny risk ratio. All possible and probable confounders are treated sloppily or not at all and completely omitted from the declared error ranges. They totally omit any consideration of in-house air quality. Air pollution is very strongly correlated to noise, neighbourhood, and general housing conditions – all known and proven to aggravate cardiovascular disease.
But much worse than all that and ample reason for this study to be thrown out by any conscientious reviewer: The smallest grid element they look at and use to classify air and people is 100 by 100 km, 10 000 km2, or 1 Million hectares. Offhand I can think of no area this size outside of the central Sahara that does not cover widely varying natural and living conditions. It is a long time since I last saw a published study as nonsensical as this, no wonder the yellow press lapped it up.
Please compare it to the methodologically very similar study by Engemann et al., that offers a real, plausible, and credible result.
***
Although what Moore et al. say is obvious and essentially trivial, they do give it a quantitative foundation. Their result is of course far more general than their chosen example and equally applies to e.g. censorship, surveillance, reduced freedom, and banned language.
Zum Anfang Übersicht Artikel Home & Impressum