[Mase]

Again, it's hard to discern their conclusions from the linked article, but I have read enough of this kind of research to believe that even if they don't directly make the link, their conclusions will make any link highly suggestive.

They have to do this if they are going to attract more money to continue their "research." Mere correlation isn't worth much, unless you think there really is something legit between the amount of rainfall in Fiji and the price of peanuts in Georgia. These people are going to have to make some sort of connection here....and they will. They are already laying the groundwork. The media are willing accomplices in the charade that passes for legitimate research today.

You appear willing to give them the benefit of the doubt. Not me, I've seen far too much of this in my career to believe otherwise.

[stormer]

Clearly without reading the paper one cannot with certainty describe their conclusions. That said, in science there is nothing wrong with saying that a correlation is highly suggestive of any given hypothesis - that's how it's supposed to work. As far as your example of peanuts and rain in Fiji, that sounds like something from Huff's "How to Lie with Statistics" - I briefly thumbed through my copy and couldn't find it and it doesn't have an index, so, who knows? Anyway, examples like that are meant to illustrate the importance of recognizing a post hoc argument and just how ludicrous some findings may be. In the case of the pop drinking kids - indeed, with most health studies - the demonstrability of correlations can be a very valuable tool in identifying trends and developing hypotheses. I couldn't hazard a guess as to the value of the soft drink industry, nor would I want to speculate on the cost to society of aggression in children and their trajectory through time, but given the importance of those, I think that research like this may be quite relevant.