Hacker News Reader

Comment by mrjangles | Hacker News Reader

I can't extract anything meaningful from this comment. It seems you are repeating a lot of "go to" sayings about how science should be done in a hand wavy way without understanding the why, then using that to dismiss the results.

For example you say

>There is no causality here; just correlation and many potential cofounders...

what does this specifically mean? When A and B are correlated to statistical significance, either A causes B, B causes A, or something else is causing both A and B. So, how could the future act of living longer be causing people to eat more protein? Or, what possible things could be causing people to eat more animal protein and live longer at the same time? I can't think of anything plausible.

It is fair to say that this study is very good evidence, as good as anything can be, that animal protein is important for longevity.

View on HN

Another comment in this thread points out that a huge confounder could be income. Indeed, as @robwwilliams points out, there could be many potential confounders, including genetics and environment. All that means is that, if drilled into, researchers might find that X (where X is whatever confounder you can think of) might be a better predictor of mortality than protein intake.

It seems to me that the parent comment doesn't aim to "hand wave away" this paper, but instead to suggest that it is a singular datum whose findings should be tested by larger, more in-depth, research (which could test for a whole set of confounding variables that might come to mind). The authors of the original study state the same thing right in their abstract!

I'm not sure why you think that point is not meaningful. IMO, that is exactly how science should be done ---interesting correlations should be stress-tested. If the findings hold up, then we can be especially confident in whatever choices we make. If the findings don't hold up, we then have new avenues to research.

> Or, what possible things could be causing people to eat more animal protein and live longer at the same time? I can't think of anything plausible.

That's a failure of imagination. A plausible example might be socioeconomic status (rich people live longer and also eat more expensive foods such as meat).

An environmental factor that reduces longevity and alters your sense of smell or taste in a way that affects your diet is semi-plausible (heavy metal poisoning such as cadmium can cause anosmia, which will affect food preferences), smaller but more widespread effects are possible.

Arguably we're already the beneficiaries/victims of mutations that make cooked food (that is more easily ingested and digested) taste better, but also make food subjected to combustion processes like smoking, searing, roasting, etc. (which can produce various carcinogenic polycyclic aromatic hydrocarbons as a side effect) smell and taste good.

A fanciful hypothesis would therefore be a mutation that directly increases longevity and also happens to make meat taste better (or vegetables taste worse, like they do to supertasters) and so affect dietary preferences.

Thanks to McDonald's, KFC et al at least the first part can safely be dismissed? Meat is among the cheapest ways to feed yourself if you disregard quality. But I don't think the study mentions they specifically only considered people eating high quality meat.

The study was done in the EU on older adults.. They are very unlikely to eat McDonalds and meat is not as subsidized as in the US. This study might return opposite results in the US, or even elsewhere within the EU like Germany, where obesity is closer to the US.

Also their reference to animal protein might mean they are on average moving higher fat meat/lower quality meat eaters toward the lower animal protein intake than average if this demographic eats a lot of meat but of different quality.

Reply on news.ycombinator.com