Re: Soda is bad for you, again.
littlek wrote:A new study shows that soda, diet or regular contribute to heart disease and diabetes. Drinking as little as one can per day can increase your risk of these diseases by 48%. 48%!!!!
That's not what the LA Times says. It says it "is associated with" these diseases. This distinction matters because soda-drinking isn't the only difference between people who drink soda and those who don't. To tease out of the data the actual impact of soda on health risks, you have to control for these other things.
The authors of the study knew this of course: That's why "[t]he authors tried to control for all those factors in the diet, but 'even after all that, we still found an increased risk,' he said. 'Maybe it is very difficult to adjust for lifestyle.'" Notice, however, that they don't quantify the risk they found after correcting for these other factors. It seems safe to bet it was underwhelming, or else the LA Times would have reported it.
littlek wrote:Wowsa! Is this study going to be found to be flawed?
I'd have to read the study to answer it. But from reading the article, I'm getting the impression that the flaw is not the study itself, but in the LA Times's coverage of it. The most conspicuous points I'm gleaning in the article are the following three: (1) Impressive-looking but irreleveant correlations between drinking soda and suffering certain diseases. (2) A multiple regression of the data that controls for other features in which soda-drinkers and non-soda-drinkers difference. This regression, which would yield the relevant numbers, seems to show that soda does have a bad effect on your health, but the effect isn't impressive enough for the LA Times to report. (3) Careful wording by the LA Times that avoids saying anything outright false, but systematically misleads the unprepared reader into believing that the impressive numbers describe a causation, not a correlation.
I am not impressed by this article.