Conquer Club

https://www.buzzfeed.com/stephaniemlee/brian-wansink-cornell-p-hacking?utm_term=.jc8YLQzgzm#.ivwByZ9v9w

These type of wide-scale retractions have become a matter of course in the scientific world, where peer-reviewed has meant "Not reviewed at all" and statistically significant means nothing whatsoever. It's nice to see a mainstream media take on this story finally, although it's probably politically motivated in some way.

To recap - not a single nutritional study has believable results, unless the nutritionists start from a firm molecular biology/biochemical background. It doesn't matter how many articles you can find, garlic and green tea don't cure cancer. Rather, they cause a particular type of cancer known as nutritional studies.

That's why I drink coffee and not green tea.

Coffee the true cancer fighter.....after beer of course.

Look at you little fella. Half right is pretty good for you. Coffee and beer are healthy but so are garlic and green tea.

Chemo is the only real cancer killer, it's no picnic though. Just watching your loved ones do it can be rough.

DoomYoshi wrote:https://www.buzzfeed.com/stephaniemlee/brian-wansink-cornell-p-hacking?utm_term=.jc8YLQzgzm#.ivwByZ9v9w

These type of wide-scale retractions have become a matter of course in the scientific world, where peer-reviewed has meant "Not reviewed at all" and statistically significant means nothing whatsoever. It's nice to see a mainstream media take on this story finally, although it's probably politically motivated in some way.

To recap - not a single nutritional study has believable results, unless the nutritionists start from a firm molecular biology/biochemical background. It doesn't matter how many articles you can find, garlic and green tea don't cure cancer. Rather, they cause a particular type of cancer known as nutritional studies.

Your leap from "some nutritional studies are bad" to "all studies nutritional studies are bad" is just as bad as the worst of those studies. It's lazy logic, and quite honestly you should be ashamed of it. In fact, it isn't logic at all; just a lazy variation of confirmation bias.

Why would they be good? They have no good methodologies and are based on the worst assumptions? Some might be correct, but that is merely accidental. Giving people some type of food item and then looking at the results is not a legitimate form of study.

Allow me to switch this on you. Can you find a good nutritional study?

DoomYoshi wrote:Why would they be good? They have no good methodologies and are based on the worst assumptions? Some might be correct, but that is merely accidental. Giving people some type of food item and then looking at the results is not a legitimate form of study.

Of course it's a legitimate form of study. Why would it be not? Empirical evidence is the evidence that matters: does something work, or not? Natives chewed pine bark for thousands of years to relieve their pain before a scientist showed up to tell them that pine bark contains aspirin. We want to fix people's lives RIGHT NOW, not wait a thousand years for the theoretical background for why a particular fix works.

This all feeds into the discussion of whether the broader field of sociology, which is the parent field of the scientists in question (not nutrition), is a real science. The answer to this is still no.

Dukasaur wrote:

DoomYoshi wrote:Why would they be good? They have no good methodologies and are based on the worst assumptions? Some might be correct, but that is merely accidental. Giving people some type of food item and then looking at the results is not a legitimate form of study.

Of course it's a legitimate form of study. Why would it be not? Empirical evidence is the evidence that matters: does something work, or not? Natives chewed pine bark for thousands of years to relieve their pain before a scientist showed up to tell them that pine bark contains aspirin. We want to fix people's lives RIGHT NOW, not wait a thousand years for the theoretical background for why a particular fix works.

It's fine to pick one type of cure that worked but there are thousands of folk cures that didn't work.

You are splitting the argument into where it is not either. I am not talking about non-statistical approaches to medicine. I am talking about statistical manipulation. With a large enough nutritional study, the evidence can say anything.

Thank you neoteny for giving me a new perspective on it. The reason I don't like nutritional studies is because they are too much like sociological studies. They may work in some cases (random chance), but they aren't science.

Neoteny wrote:This all feeds into the discussion of whether the broader field of sociology, which is the parent field of the scientists in question (not nutrition), is a real science. The answer to this is still no.

Wrong. Again, just lazy over-generalization. There's an awful lot of bad sociology out there; that does not mean that all of it is bad.

Many phenomena described by sociologists are unquestionably real. You can take a phenomenon such as civil inattention, and confirm it for yourself next time you step on an elevator.

Others might be not so valid.

Dukasaur wrote:

Neoteny wrote:This all feeds into the discussion of whether the broader field of sociology, which is the parent field of the scientists in question (not nutrition), is a real science. The answer to this is still no.

Wrong. Again, just lazy over-generalization. There's an awful lot of bad sociology out there; that does not mean that all of it is bad.

Many phenomena described by sociologists are unquestionably real. You can take a phenomenon such as civil inattention, and confirm it for yourself next time you step on an elevator.

Others might be not so valid.

Was Civil Inattention the result of an experiment or of a keen mind describing things that others saw but didn't describe?

Dukasaur wrote:

Neoteny wrote:This all feeds into the discussion of whether the broader field of sociology, which is the parent field of the scientists in question (not nutrition), is a real science. The answer to this is still no.

Wrong. Again, just lazy over-generalization. There's an awful lot of bad sociology out there; that does not mean that all of it is bad.

Many phenomena described by sociologists are unquestionably real. You can take a phenomenon such as civil inattention, and confirm it for yourself next time you step on an elevator.

Others might be not so valid.

No, it's not lazy or a generalization. You can probably give plenty of legitimate examples of decent cultural anthropology that hold water, but one has to admit after more than a cursory glance at the field that sociology, like economics, is a psuedoscience.

I think there is a miscommunication here. Dukasaur seems to be equating the terms scientific and truthful. Therefore, when we say something is unscientific, he takes it as an attack on the truth of the thing in question.

Any study that announces 'correlation between x and y discovered' is meaningless anyway, unless it can also produce compelling evidence (not just speculation or common sense) demonstrating exactly why the correlation exists. I had thought studies like this only ever gained traction in clickbait newspapers and websites, but apparently conducting studies like this is enough to make someone a senior member of staff at an elite American university too.

Neoteny wrote:This all feeds into the discussion of whether the broader field of sociology, which is the parent field of the scientists in question (not nutrition), is a real science. The answer to this is still no.

This isn't sociology. The article quite clearly states that it is (shit) psychology.

DoomYoshi wrote:

Dukasaur wrote:

DoomYoshi wrote:Why would they be good? They have no good methodologies and are based on the worst assumptions? Some might be correct, but that is merely accidental. Giving people some type of food item and then looking at the results is not a legitimate form of study.

Of course it's a legitimate form of study. Why would it be not? Empirical evidence is the evidence that matters: does something work, or not? Natives chewed pine bark for thousands of years to relieve their pain before a scientist showed up to tell them that pine bark contains aspirin. We want to fix people's lives RIGHT NOW, not wait a thousand years for the theoretical background for why a particular fix works.

It's fine to pick one type of cure that worked but there are thousands of folk cures that didn't work.

Of course there are thousands that are bogus. That's why we need rigorous studies to see which ones work and which ones don't. But if you're going to condemn all such studies a priori, then you're throwing out the baby with the bathwater. Hell, you're throwing out the baby without even drawing the bath.

DoomYoshi wrote:You are splitting the argument into where it is not either. I am not talking about non-statistical approaches to medicine. I am talking about statistical manipulation. With a large enough nutritional study, the evidence can say anything.

Yes, I know you're talking about statistical manipulation. And that's what I'm addressing. You're saying that because some nutritional researchers have engaged in fraudulent stats, that means that all of them have. Which doesn't follow. Because some cops are crooked doesn't mean all cops are crooked. Because some whores have false teeth doesn't mean all whores have false teeth.

Even in the article you posted, it said possibly half of all studies might be tainted. If half are tainted, that implies the other half are not.

Neoteny wrote:

Dukasaur wrote:

Neoteny wrote:This all feeds into the discussion of whether the broader field of sociology, which is the parent field of the scientists in question (not nutrition), is a real science. The answer to this is still no.

Wrong. Again, just lazy over-generalization. There's an awful lot of bad sociology out there; that does not mean that all of it is bad.

Many phenomena described by sociologists are unquestionably real. You can take a phenomenon such as civil inattention, and confirm it for yourself next time you step on an elevator.

Others might be not so valid.

No, it's not lazy or a generalization. You can probably give plenty of legitimate examples of decent cultural anthropology that hold water, but one has to admit after more than a cursory glance at the field that sociology, like economics, is a psuedoscience.

All you're doing is adding more slander to the list. Economics is unquestionably a real science. Economic laws describe how things work no less than the laws of physics. Gresham's Law is no less observably true than Boyle's Law.

mrswdk wrote:

Neoteny wrote:This all feeds into the discussion of whether the broader field of sociology, which is the parent field of the scientists in question (not nutrition), is a real science. The answer to this is still no.

This isn't sociology. The article quite clearly states that it is (shit) psychology.

I mean the article calls him a social scientist and psychologist and bro has his phd in fucking "consumer behavior" so who even knows. I stand by my statement.

I remember thinking the Laffer Curve was really interesting, until I read up on it and found out the Curve was just a doodle some economist had drawn on a napkin and that actual studies using actual data have totally discredited it.

Neoteny wrote:I mean the article calls him a social scientist and psychologist and bro has his phd in fucking "consumer behavior" so who even knows. I stand by my statement.

Yes.

• The social sciences are a group of subjects, including economics, sociology and psychology.
• Psychologists study psychology.
• Consumer behaviour is the application of psychology specifically to the way in which consumers behave.

The nice thing about sociology definitions is that you can use whichever one best suits your needs:

Consumer behaviour is the study of individuals, groups, or organizations and all the activities associated with the purchase, use and disposal of goods and services... Consumer behaviour is an inter-disciplinary social science that blends elements from psychology, sociology, social anthropology, ethnography, marketing and economics, especially behavioural economics.

Dukasaur wrote:
Even in the article you posted, it said possibly half of all studies might be tainted. If half are tainted, that implies the other half are not.

Even the accepted statistical methods aren't great though. It comes down to correlation does not equal causation.



Those are acceptable statistics. They are done using statistical best practices. The conclusion doesn't make sense.

You can accuse me of rejecting all fad diets a priori. I will accept that accusation and claim that it is the lesser of two evils.

DoomYoshi wrote:
You can accuse me of rejecting all fad diets a priori. I will accept that accusation and claim that it is the lesser of two evils.

Don't ridicualize it. You said that no nutritional studies are valid without biochemical theory. You're painting with an awful broad brush, not just fad diets.

You want to throw out a VAST body of knowledge about the health benefits of berries for instance, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3068482/ because we don't fully understand all the biochemical pathways involved.

You want to throw out dietary knowledge that's not biochemically-based, like this tidbit.

You even want to throw out modern medicines, a vast number of which were found by simple trial and error rather than through any overarching biochemical theory of how they work.

The problem isn't the studies. It's the rumors.

"I heard xxxxxx is an anti oxidant. That cures cancer!" Next thing everyone is overdosing on xxxxxx.

No, that won't cure cancer, it probably helps prevent it if you use it in moderation with other healthy habits.

2dimes wrote:The problem isn't the studies. It's the rumors.

"I heard xxxxxx is an anti oxidant. That cures cancer!" Next thing everyone is overdosing on xxxxxx.

No, that won't cure cancer, it probably helps prevent it if you use it in moderation with other healthy habits.

A classic problem in biochemistry and medical research. Every week there's an article where such-and-such drug discovery will lead to cure for cancer or Alzheimer's, but when you read the article it's a minor correlation of some inhibitor protein or something that gets blown out of proportion.

Which i guess goes to why those amped-up article get views in the first place.

Dukasaur wrote:

DoomYoshi wrote:
You can accuse me of rejecting all fad diets a priori. I will accept that accusation and claim that it is the lesser of two evils.

Don't ridicualize it. You said that no nutritional studies are valid without biochemical theory. You're painting with an awful broad brush, not just fad diets.

You want to throw out a VAST body of knowledge about the health benefits of berries for instance, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3068482/ because we don't fully understand all the biochemical pathways involved.

Knowing about the health effects of berries has nothing to do with nutritional studies. It's based on people eating berries for years. It's unscientific knowledge. Nobody has ever advocated not eating berries. However, people publish "eating berries is good for you" as if it is news.

Let's take any particular claim in that study you have presented "demonstrated significant improvements in LDL oxidation". That phrase means something along the lines of "compared to the null hypothesis, people who eat berries have improvement in LDL oxidation". Without going into specifically what their null hypothesis was, let's just assume that it was "compared to people who don't eat berries". That still doesn't actually tell us anything. What if cantaloupe-eaters are the best? Does that mean that it's healthier to eat only cantaloupe instead of berries? No it doesn't, because LDL oxidation is a basic metabolic trait, not directly a result of diet (I really don't want to open the debate about diet and cholesterol, as this has been a giant pissing match for 30 years and I'm not an expert in the field). One problem with nutritional studies is that secondary effects have a more pronounced impact than primary effects*. Greater than any individual item you can eat is the overall diet and metabolism. Perhaps with some supercomputer you could determine the correct number of berries to eat exactly, but I'm not sure how that information helps anyone in any way.

So what are those studies good for? Pretty much nothing. Metabolic pathways are so complete that 99% of what you eat ends up the same no matter what. No matter which animal you eat, it ends up being amino acids. No matter which source of fiber you eat, it ends up being roughage. There are certain chemicals that we can't create and need from our diet, but none of that knowledge came from nutritional studies, it all came from biochemical studies.

I'm not sure the second article you posted is a nutritional study or even accurate. It falls into the same trap that many psychology papers fall into. They aren't describing human psychology but rather the psychology of undergraduates. Undergraduate students care about what it looks like they are eating. Ask a starving man in Africa if he cares about the potato chips that you give him and I bet you will get a different answer. THIS JUST IN: FEMALE UNDERGRADUATES CARE ABOUT IMAGE! SHOCKING DETAILS TONIGHT AT 11!

Trial and Error in a medical sense works very different than trial and error for a nutritional sense. Is anyone suggesting that the best cholesterol drug is to eat more berries?

To put this another way: how can you use this information in your life? What if I find some studies that show things that berries are bad at? How do you decide between the two? Simple, confirmation bias - the cornerstone of nutritional studies.

Maybe if you struggled with cholesterol, you might look up a list of foods that helps you get cholesterol under control. What is more likely is that people look for the best solution(s). Underlying all of this is the possibility that berries may not actually help with cholesterol, without looking at the study, it's impossible to trust it, because it comes out of a murky field of science where doubt is a sensible first defense.

I'm not sure why you have turned an attack on a pseudoscience into a moral crusade. It isn't science, end of story, let's move on with our lives.

*By this I mean that the statistics they use stem from controlled experiments, where all else is equal and the only variable is berries or no berries. This doesn't describe actual studies though, where even controlling for berries is almost impossible (some people eat berries without knowing, some people over- or under-report the amount of berries they eat, not all berries are created equal, some people might think that Swedish Berries count as berries, etc.).

TA1LGUNN3R wrote:
A classic problem in biochemistry and medical research. Every week there's an article where such-and-such drug discovery will lead to cure for cancer or Alzheimer's, but when you read the article it's a minor correlation of some inhibitor protein or something that gets blown out of proportion.

Which i guess goes to why those amped-up article get views in the first place.

We are naturally prone to wanting the easy fix. Take your multi vitamins and live forever with a nice boner and a buzz.

Neoteny wrote:The nice thing about sociology definitions is that you can use whichever one best suits your needs:

Consumer behaviour is the study of individuals, groups, or organizations and all the activities associated with the purchase, use and disposal of goods and services... Consumer behaviour is an inter-disciplinary social science that blends elements from psychology, sociology, social anthropology, ethnography, marketing and economics, especially behavioural economics.

Doesn't change the fact that studies like the one in OP are psychology studies, not sociology.