On Sunday, April 23, 2017 at 12:16:39 PM UTC-4, Ignoramus8879 wrote:
On 2017-04-23, wrote:
Here is the study's abstract, with links to the full study:

http://stroke.ahajournals.org/conten...AHA.116.016027

Nobody "loaded the dice." They had plenty of anecdotal data to
suggest the risk-factor correlation. Then they applied
straightforward statistical methods to see what the data tell us.

Don't guess about this. If you haven't read at least the abstract,
and if you don't know how these associations are researched and
measured in medicine, find out before jumping to conclusions.


That abstract speaks for itself. Thanks for posting a link to it.

It's not that impressive, though. The wide confidence intervals, which come close to including 1.0, mean that the statistical significance is borderline. Add a bit of Bonferroni correction, and it'll disappear. And Bonferroni correction is appropriate here, since they investigated at least two hypotheses (sugar-sweetened and artificially sweetened). Did they use such a correction? Well, searching the article for "Bonferroni" comes up empty, as does "correction", but possibly they used it under another name. It wouldn't be unusual if they omitted it, though.

https://en.wikipedia.org/wiki/Bonferroni_correction

And I'm sure they explain what they really mean somewhere, but it is rather odd to find any dietary factor significant after adjusting for "diet quality".

My beef with artificially sweetened soft drinks is that they fool the body: when taste buds register sweetness, the body reacts by increasing blood sugar in anticipation of its carbohydrate storage being replenished by the incoming food. When no carbohydrate is really incoming, this leaves the body's storage more drawn-down than it should be. So I wouldn't have a problem with there being a real effect here, even if this study can't pick it out of the statistical noise.