Reacción a "Dark chocolate (but not milk chocolate) is associated with a lower risk of diabetes"

I do not like the study because the authors recognise that it does not allow causal relationships to be established, but they repeat the message of association as if it were causal, inducing a certain consumption that would only make sense if the relationship were causal. The press release merely repeats the same thing, only shorter. Moreover:

1. They repeat the obsolete term ‘significant’, associated with the much discussed P<0.05.
2. If they are interested in non-causal association, they should quantify its predictive value with measures of the ability to anticipate the future, such as R2 [coefficient of determination] or ROC [ Receiver Operating Characteristic].
3. They insist on a measure, relative risk, which only makes sense for causal associations. But they do not facilitate its practical interpretation. For example, how long will individuals remain free of the complication [diabetes]?
4. The analysis of their weaknesses is poor, not useful for learning.
5. Can we believe that they have no conflict of interest simply because of public funding? We have no guarantee that the analysis was designed impartially, in the so-called ‘original position’, before the results were known.