This short article is a good summary on the flawed nature of nutritional epidemiological research.

A key thing to remember when you see the latest breathless headline about this is good for you, this is bad, so-and-so is a miracle food, etc.: If the claim is made based on NER, it is almost certainly weak at best, and more often than not, wrong.

Instead, ask: Is the claim based on an experiment? Was there a control arm? What is the mechanism that is causing the gain, or harm.

Here is a simple example: Seventh Day Adventists in So. Cal:
-Eat less meat than is common
-Have a 7 year longer lifespan
Breathless headline: Meat is bad!
The SDA's also weigh less, exercise more, have strong social bonds, are not poor, and critically, have very low levels of alcohol, cigarette, and drug use.

Guess what? There is another group that looks just like the SDA's in almost every way including lifespan, except they eat meat!
Breathless headline: Meat is good!
Or not.
This other group is Mormon's in So Cal.

[crossfit.com]