Hacker News Reader

Comment by Aardwolf | Hacker News Reader

The state of food research science is so ridiculously bad. We got paper after paper, and badly written news article after badly written news article, contradicting each other on everything about fat types, coffee, milk, meat, wine, anything else that's edible or drinkable you name it.

It's as if we know exactly zero information about the relationship between food and health, because for every fact, there'll be another contradictory fact.

I have no trust in food related scientific research. Give us properly conducted experiments with large enough (millions?) sample sizes and all the necessary blinds, publish it, and ensure we can STICK with it, don't leave openings for papers with the opposite opinion to be published, by making your experiment good enough to be indisputable. Thank you.

View on HN

One of my Grandpa's go-to joke from the 90s:

"Before <Grandma> starts making breakfast, can someone check the news and see if we're allowed to eat eggs this week?"

One contributing factor is that many studies are funded by the people selling you the food being studied

Not just the foods, but also the books, pills, programmes, etc.

> Give us properly conducted experiments with large enough (millions?) sample sizes, publish it, and ensure we can STICK with it, don't leave openings for papers with the opposite opinion to be published, by making your experiment good enough to be indisputable.

With all due respect, this comment sounds like you've not really ever considered how this research is actually conducted.

I indeed don't know how research is conducted in this field, and I'm sorry if I made the wrong point due to that. But we're getting papers with opposite opinions now, so something is very wrong with how it's conducted now. Either do it properly, or be honest about the correctness/p-values/... to avoid articles on "red meat is bad", then articles with "red meat is not bad", or "coffee causes cancer", then "coffee has antioxidants and is good against cancer", etc...

That's a fair point, and reasonable as someone who's looking for actionable insight. But papers are supposed to contradict each other. This is how scientific knowledge evolves.

I will not claim that nutrition science has achieved methodological purity, because obviously that is far from the truth. But the system doesn't seem to be as broken as folks think.

I think the field just has a long way to go before it settles on methods that are reproducible. Until then, opposing papers are literally the only path we have to get there.

No, it’s totally broken. Typical nutritional studies take a large group of people, ask them what they normally eat, and look for correlation with whatever health problems that have. This can’t possibly work because most people’s diets have been influenced by whatever crap they have read about nutrition over the years.

Or they use animal models, animals whose natural diet looks nothing like a human diet.

The only way to really know how diet affects health is to take a large group of people, assign them different diets, see that they actually follow them, then measure the outcomes. But practically nobody does this because it is really expensive to do.

> I will not claim that nutrition science has achieved methodological purity

I'm not in a position to critique the methodology. But the reports that reach the public (and what use are reports that don't reach the public?) keep contradicting one another. If they can't get consistent results, then these reports shouldn't be appearing on the front pages of newspapers; they should stay in the pages of obscure journals.

I pay no attention at all to "nutrition science". I eat a mixed diet, no fruit at all, few green vegetables, a small steak once a week or so, quite a lot of cheese and whole grains. I let my stomach guide me. If I followed dietary guidelines, I'd probably starve. I simply can't swallow food that doesn't appeal to my appetite.

Like if research was by and large conducted out of any kind of business considerations and agenda. We're getting these contradictory signals for the same reason that it was supposedly debated in the 90s that humans had an impact on the climate or in the 70s that smoking was cancerous.

> "red meat is bad", "red meat is not bad"

Part of your frustration is that you're asking the wrong question. "Is it good or bad?" That is the too simple of a question.

Let me use an example we can all understand on Hacker New: Is React good or bad? Should I start using React? We all know, it depends. Asking whether something is good or bad is not helpful.

The human body is more complicated than even the worst software project. And we're dealing with millions of bodies, all different, not just one. What is healthy for one person will kill another person (allergies, etc). Going back to React, we seem to want the answer to be either "everyone should use React", or "nobody should use React"; we want a clear answer that applies to everyone, but that's impossible.

I found this YouTube video almost exactly answers your question, context and all: https://youtu.be/mQ56uOkjccg

Reply on news.ycombinator.com