[np422]

According to the statistic available the number of drowning accidents among children and sale of ice-cream is highly correlated, we must take measures to lower sale of ice-cream, please think of the children!

Most of the time a journalist will easily believe a statement like above and rewrite into a newspaper article.

And soon thereafter official recommendations and legislation will follow.

I'm looking forward to http://nusi.org/ publishing some results, they appear to be intellectual honest.

[plafl]

I'm always wary of articles that try to support their point by showing some ridiculous correlations. It's very easy to look at thousands of correlations and select the few most ridiculous, ignoring the vast amount of them that maybe make sense. Suddenly it seems like statistics are irrelevant and we cannot trust anything. I suppose that nobody is free from personal bias since even Fisher, yes the great statistician, rejected correlation between smoking and lung cancer as a spurious correlation (look up in Wikipedia)

[TeMPOraL]

> It's very easy to look at thousands of correlations and select the few most ridiculous, ignoring the vast amount of them that maybe make sense.

The reality is in fact reverse. Out of thousands of correlations you could pick, only very few actually make sense. It's trivially easy to find - or accidentally stumble upon - a meaningless correlation.

[NoGravitas]

Drowning accidents and the sale of ice cream probably have a common causal factor -- warm weather. A lot of the correlations in the article are a lot weirder and harder to explain.

[noondip]

NuSI is straight-up fraud. See http://plantpositive.com/warning-signs-nusi-guys-1/