Hacker News Reader

Hacker News

Comment by 1_over_n | Hacker News Reader

Parent Full thread

1_over_n

"Funding

This study has been supported by Tchibo GmbH, Hamburg, Germany."

Tchibo is a German chain of coffee retailers and cafés known for its range of non-coffee products that change weekly.

https://en.wikipedia.org/wiki/Tchibo

Number of locations 700 shops (Germany) 300 (rest of world)

They might want to make some in-store claims about Dark Roast...

View on HN

I'm generally sceptical of all studies about coffee for this very reason. There's a lot of bullshit studies around food and supplements funded with people with strongly vested interests in selling more product. It's a real problem. For any given supplement, you can often find multiple studies supporting its use and multiple studies saying it's harmful.

When it comes to coffee specifically, if you listen to the studies about coffee, it's some panacea full of anti-oxidants. However, it's a stimulant, it increases cortisol (stress hormone). Stimulants generally makes the body prioritize immediate energy spending over regeneration/healing. There's been a study claiming that caffeine decreases neurogenesis, which should negatively impact both long-term memory and mood.

> you can often find multiple studies supporting its use and multiple studies saying it's harmful.

While true and I get your point, there isn't anything inherently wrong with that. We should expect and welcome research with the awareness of vested interests.

Afterall it's not surprising for a substance to be beneficial in certain ways yet harmful in other ways. Research from opposing groups will help identify all interesting attributes of a substance. Until we learn all there is to learn, you should approach these papers more as additional clues pointing to a certain direction and not anything conclusive about the substance in question.

I work for a biotech company that publishes peer-reviewed science. We use scientific publishing for marketing purposes, to influence the regulatory environment in our favor, and to look better than our competitors. Some of it is done to a very high standard. But in general, doing science with such predefined strategic objectives results in a corpus of scientific output that is hard to interpret. Essentially this sort of stuff pollutes the scientific literature and we should be ashamed that as a society we can't even let the technical content of the scientific literature be free from corporate interests.

I think in the current research climate, you can replace "biotech company" with "university department" and most of your statements still hold true.

Many purely academic studies are done with predefined strategic objectives of advancing one's own prestige, only publishing positive results that fit "your narrative", and indeed polluting the literature with results that are at best incremental improvements, at worst regressions made up to look like improvements.

Are you directly involved in this? If so - how do you justify it to yourself?

(if not then ignore the question - I'm just curious about the mindset of people who do ethically-questionable work)

> free from corporate interests

That won't help. All research is subject to the bias of the researchers. For example, many researchers have been caught fabricating results so they'll get famous. And just because someone is being funded by the government hardly means the government doesn't have a biased interest in the result. Everyone running the government has an agenda that isn't exactly altruistic.

This is funny.

Just a while back there was an article posted about how studies SHOULD start out with defined objectives. The argument was that academics are so desperate to publish positive results that they'll draw any positive conclusion, even if it wasn't what the study started out trying to accomplish.

Indeed. What you are describing is p-hacking

https://en.wikipedia.org/wiki/Data_dredging

Thank you! I didn't know there was a term for it.

Second that

If p hacking wasn’t so prevalent I would agree with you. Otherwise the researchers funded by MEGA COFFEE will just keep running trials until they get lucky and publish the positive coffee result without mentioning all of the other failed attempts.

>We should expect and welcome research with the awareness of vested interests.

No, we should ban vested interests from funding research studies and reduce the SnR.

Who could possibly not be a vested interest? It's hard to think of more than a few cases where I've personally wanted to learn about something and didn't, at some level, hope for one answer or outcome over another. "Ban vested interests" is an impractical policy. stevenhuang's suggestion to be aware of them is much more practical and has the benefit that "vested interests" who truly do wish to fund an objective study aren't stopped from doing so.

You mean push them into second order vested interest funding?

No, I mean harsh criminals penalties if they can be connected to higher-order vested interest funding too...

GP just means that effectively that's what will happen. just like wow churches push money through think tanks and orgs that donate to political campaigns. The money will never go away it will just get darker and darker and darker.

> There's been a study claiming that caffeine decreases neurogenesis, which should negatively impact both long-term memory and mood.

How much caffeine was required for this effect and how long did it last?

Roasting aka burning the beans seems like it would be a cause of cancer, similar to research about cancer causing agents in the burnt aspects of meat, etc.

I think the cancerous agents in burnt meat are caused by extremely high temps, higher than would be reached in coffee roasting. High temps in coffee roasting could cause the undesirable effect of caffeine break down.

I often see charring on beans. Starbuck's beans are notoriously over-roasted.

Completely - you could also say that reducing DNA damage might actually also decrease immune system surveillance that might be happening causing worse than expected outcomes. Imagine if a lot of cells that might have killed themselves stay alive because of drinking coffee might actually turn out to be a bad thing. We just don't know enough to say for sure and these sort of studies are wildly difficult to get mortality figures from.

We still aren't even sure what general diet is fully optimal long term let alone if one constituent can give actual life expectancy benefits.

So far mortality and morbidity figures say "meh". It is relatively easy to find good quality observational studies with decent power on this exact subject, they're just not sexy so they do not get quoted in pop science magazines.

To some extent isn't that expected?

Let's face it - who would spend the money to fund a study on a substance that's been around for hundreds (thousands?) of years?

Probably if you're going to spend the money to investigate it, you expect some kind of return. So sadly, all these studies end up being funded by people with an interest in a positive outcome.

The bigger question to me is, does that immediately invalidate their results?

The bigger question to me is, does that immediately invalidate their results?

It warrants skepticism of the strongest variety unless the study can be reliably replicated under better circumstances. Don’t be shocked when it doesn’t happen though.

Like alcohol studies paid for by brewers, this is hard to accept at face value given the knowns. Even when money isn’t pulling the strings, nutrition studies are notoriously unreliable by dint of human variability and the nature of the studies themselves.

It depends on the ratio of studies funded vs studies published.

I wouldn't be surprised if around 90-95% of special interest funded studies that aim for a p value of <5% aren't published.

> I wouldn't be surprised if around 90-95% of special interest funded studies that aim for a p value of <5% aren't published.

This is true of the majority of publicly-funded research, too.

> The bigger question to me is, does that immediately invalidate their results?

How many times did they fund this study with different research groups, quietly round filing the results each time until they found a research group willing to p-hack a flattering result?

Yes, I agree, and expect people, companies or whatever to act in their own self-interest, so no implication it is rigged because it may be self-serving. However, I am more suspicious of the study, and studies in general after the recent wave that uncovered an epidemic of unsound statistical methods being used in studies across multiple disciplines. Unfortunately, I am not proficient enough in statistics to vet these things out for myself most of the time, so I remain skeptical, and similarly cliche with the Edgar Allen Poe quote:

“You are young yet, my friend,” replied my host, “but the time will arrive when you will learn to judge for yourself of what is going on in the world, without trusting to the gossip of others. Believe nothing you hear, and only one half that you see."

>The bigger question to me is, does that immediately invalidate their results?

Considering that most such research results are already non-reproducible and p-hacking, the added vested interests land this squarely in the crap category.

People like us nerds thinking it through strict logic ("still this does not necessarily invalidate") might reduce some false negatives, but it mostly just wastes our time and has us taken for suckers from industry/marketing con men. With this kind of "research" taking a shortcut in its evaluation is more productive.

If a YouTuber pulled off this study and vlogged it from start to end, they would make a shit ton of money.

> However, it's a stimulant, it increases cortisol (stress hormone). Stimulants generally makes the body prioritize immediate energy spending over regeneration/healing. There's been a study claiming that caffeine decreases neurogenesis, which should negatively impact both long-term memory and mood.

All true.

Now, by the same token, I am wondering about the effects of decaf. You know, all the alleged "good stuff", none of the sympathetic-nervous-system-stimulation.

Does it have "all the alleged good stuff"? Most of the alleged good effects are in caffeine itself...

If you just want antioxidants (which are themselves oversold too) just eat some berries or something...

> Most of the alleged good effects are in caffeine itself...

I vaguely doubt that.

I mean, this could be tested. Three groups: cold turkey controls, crappy tasting instant coffee group (or even just give them a caffeine pill or something), and full-aroma real coffee group. Groups 2 and 3 get the same amount of caffeine. Now compare the outcomes.

That would be very interesting.

You know how coffee is decaffinated, right?

Edit to address Florin_Andrei's critique: last I checked, the organic solvent processes predominant. Just a gestalt, but I suspect there are some adducts produced and the separation, though evaporative, is still imperfect.

Just something to keep in mind.

I mean I can read this just like anybody with an internet connection

https://en.wikipedia.org/wiki/Decaffeination#Decaffeination_...

but you'd have to be more specific as to what your concerns are.

It's the same with studies about alcohol. They seem to always fish for some positive effect and ignore all negative effects.

ya they're mostly red herrings. don't forget the impact on our sleep cycles which every day prove to be critically important to our longevity. for a pretty strong and compelling case for sleep (and against anything that gets in its way) -- check out "Why We Sleep" by Matthew Walker.

It's also a typical horseshit vague claim that almost certainly is the result of p-hacking ludicrous meta-studies and other nasty practices. Same as those damn red wine studies that come out all the time touting its health benefits. There's an astounding number of pure trash studies out there.

No, drinking burnt coffee isn't going to defend anyone's DNA. This is patently absurd. Come on y'all, don't vote for this kind of stuff. Shame.

I'm already addicted to your foul dark energy swill, you don't need to sell it to me twice!

I feel this is an instance of "hate the game not the player". Funding of scientific research should always be good, even if the stakeholders are hoping for a favorable outcome. By it's definition science is merely uncovering the truth and can't be swayed.

The reality however where we don't trust the results (especially around nutrition), the issues with replication of studies, and wrong usage of p-values is the real problem. Perhaps we need to mandate replication studies... I'm not sure what the solution is - but I don't feel blaming the industry is the right approach either. Science has the problem here.

I can't access the full text but from the summary it looks as if blood was taken only once. So the difference between one and the other group was not measured before-after and then the change compared, rather a simple comparison was made between the groups at the end. If that's true the whole thing is just bollocks - run it often enough and you'll find a group with this effect.

http://www.ico.org/promotion_e.asp

Hmm. This is why climate change deniers can get a foot in the door....

....the door has been jammed open by skewed commercial interests for decades.

Wow, this sounds like the alcohol industry funding research publications which emphasize that consumption of 2 glasses of wine daily is beneficial to health.

Thanks for pointing it out.

Thank you! What a ridiculous study.

Also the list of authors is a bit interesting. Nothing shocking - but interesting.

Could someone please explain to me how funding could lead to biased results? Is it disbursed conditionally? Is it the desire for future grants?

If it's just bribery, are all of these scientists just selling out constantly?

Too late to edit but I ask this question in earnest.

I have trouble with the notion that we should be immediately skeptical, without reviewing methodology, of any study with industry funding. Some 70% of research funding comes from private entities.[1]

That said, given the current sorry state of "science" (p value abuse, publish or perish, replicability crisis) it wouldn't surprise me if funding related bias really is a common problem.

1. https://en.m.wikipedia.org/wiki/Funding_of_science

Yeaaaaaaah but who else would do the study

Cmon use your head here

Whether its the sugar company or the coffee company, who else would do the study just so some college conspiracy theorists dont make a netflix documentary

NOBODY

Its okay to acknowledge the conflict of interest thats literally what the disclaimer is for, but lets not pretend like a more objective version of this study would have ever happened

> but lets not pretend like a more objective version of this study would have ever happened

Precisely, because a more objective version would have found no effect whatsoever.

Reply on news.ycombinator.com