Anything that sick people do less is inversely associated with mortality. The "controlling for blah blah blah" stuff can't be believed---are the researchers succeeding at controlling for everything that affects health? Not even close.
If you see a claim about food and health that doesn't spend a huge amount of time talking about the researchers' experimental or quasi-experimental approach, just ignore it.
What is this describing? First-party apps with Facebook integration and/or OS features connecting to Facebook? The leakage of Facebook information onto MS/Apple/Blackberry servers would be concerning, but having Microsoft software connect to Facebook on a user's device sounds harmless (to the extent we trust MS/Apple/Blackberry software to not leak information so accessed). Right now I'm giving Apple similar access to every single communication I make through my computer, to my bank accounts and health records, to all the work I do for my employer.
This distinction wasn't made clear in the story (or I can't read) and it's an important one. Privacy is complicated enough already.
Note also that any distributional assumptions are really only necessary for inference (i.e., tests and confidence intervals) in finite samples (read: small samples); the central limit theorem guarantees the tests work asymptotically, so you're usually going to be fine.
Most of the attention paid to distributional assumptions in regression is wasted, and would be better spent on really thinking through the assumed moment conditions underlying the estimator.
> If the error vector in our regression model follows any distribution in the family of Elliptically Symmetric distributions, then any test statistic that is scale-invariant has the same null and alternative distributions as they have when the errors are normally distributed.