Hacker News Reader

Comment by dsrguru | Hacker News Reader

Right, so people are often really bad at interpreting studies.

But in the case of this article, there is an obvious causal mechanism between eating beef and increasing your risk of heart disease. In fact, we don't need a study to tell us that. Beef, especially the kind of corn-fed beef that's very popular in the U.S. which comes from cattle who are confined to tiny living quarters throughout their lives, has a much higher concentration of saturated fat than, say, chicken. And we've consistently observed that consuming large amounts of saturated fat increases your risk of cardiovascular disease. So eating a lot of beef increases your saturated fat intake, which increases your risk of cardiovascular disease. How is that hocus pocus?

View on HN

The point of the article is that observational studies are terrible and non-scientific, but people seize on them anyhow if they like the conclusions. You reply by saying, in essence, "this article must be wrong, because it disagrees with some observational studies which must be right because I like their conclusions".

Ironic.

(As the article points out, the very limited hard data we have - ie, not from observational studies - indicates that meat-rich diets are, if anything, good for the heart. And that observational studies such as the one under discussion ALWAYS show that the people who comply with advice (even if it's to take a placebo) are healthier. So your blithe "this study must be right, because we already know eating meat is bad, and what else could explain these people being healthy?" is just magical thinking. And thoroughly debunked by the very article you're commenting on.)

The causal mechanism is that there is a correlation between saturated fat intake and cardiovascular disease? That makes no sense.

There is also a correlation between the frequency a male shaves their face and cardiovascular disease.

There is no obvious causal mechanism, we always need a study. It's very easy to be biased (Just read a few of these articles for examples http://wiki.lesswrong.com/wiki/Sequences). Books like "wheat belly" and "Why we get fat?" make an alternative hypothesis to the commonly accepted "Saturated fat is bad"

Isn't Taubes claiming just the opposite was reported?

> And when these [randomized-controlled] experiments have been done, the meat-rich, bacon-rich Atkins diet almost invariably comes out ahead, not just in weight loss but also in heart disease and diabetes risk factors.

I must not have gotten that far in the article. But I seriously find that hard to believe. I recognize that eating a diet rich in saturated fat and very low in other calories would facilitate weight loss, but that saturated fat increasing the risk of heart disease and the like has not been demonstrated outside of studies that incorrectly assumed causation, and that such a diet is likely healthful goes so far against my intuition and observation of almost everyone I've ever interacted with (including many people whose lifestyle decisions were otherwise relatively health-promoting) that I'll need to see evidence to the contrary in order to replace this hypothesis as my dominant theory. And my intuition has almost never proven wrong. XD

Heh. My own intuition is that we're good for roughly forty years of eating just about anything we can find. Some specific diet which consistently gets us more than that certainly hasn't been easy to identify, and I don't see why we should expect one to exist; the http://en.wikipedia.org/wiki/Grandmother_hypothesis seems inconclusive so far.

I say that having already reached 41…

correlation != causation

From the article: as we move from the bottom quintile of meat-eaters (those who are effectively vegetarians) to the top quintile of meat-eaters we see an increase in virtually every accepted unhealthy behavior — smoking goes up, drinking goes up, sedentary behavior (or lack of physical activity) goes up — and we also see an increase in markers for unhealthy behaviors — BMI goes up, blood pressure, etc. A whole host of effects confound the interpretation of the association.

Reply on news.ycombinator.com