Hacker News Reader

Comment by currymj | Hacker News Reader

there are many separate problems of scientific fraud. I think these issues sometimes get confused which is unhelpful.

1. apparently-legitimate papers in prestigious journals with fraudulent data. extremely bad.

2. legitimate papers in legitimate journals which, innocently or not, just used bad methods and have wrong conclusions. this is "the replication crisis".

3. totally fake papers in paper mills with no meaningful peer review. it's really easy to spot these, no one is individually getting taken in by the results, but...

3a. sometimes they wind up in a meta-analysis, which is really bad because people might trust the meta analysis.

Problem 1 is morally worst and much more common than one would hope. Outright fabricated data in a Nature or NEJM publication (as has happened) is a disaster.

Problem 2 is amenable to reform for the most part (fields are already doing this).

Problem 3 isn't a problem at all for scientific knowledge per se, although universities and funding bodies might not be pleased their scientists are buying fake papers. You can just ignore the paper mills.

But Problem 3a can actually alter policy, which is pretty serious.

View on HN

> Problem 1 is morally worst and much more common than one would hope. Outright fabricated data in a Nature or NEJM publication (as has happened) is a disaster.

In my experience, fabricated data is rare. But cherry-picking data and misuse of statistics to prove the desired hypotheses is extremely common in top journals from e.g. the Nature Portfolio journal family. I'd say at least 1/5 of the articles in biology and medicine have serious methodological errors.

IMHO, the main reason are perverse incentives and power imbalance. If you are a student or a postdoc and you get pressure from your supervisor to obtain certain result and you refuse to engage in academic misconduct, you will get pushed out of most projects and this might be the end of your academic career.

In contrast, there are few, if any, consequences for the senior scientists and principal investigators who perpetuate this culture. There's no oversight, no integrity checks, nothing. Universities tend to ignore even clear cases of data fabrication and fraud, where e.g. images derived from experiments are fake. Ultimately, this creates selective pressure for misconduct, as those willing to bend the rules are rewarded.

The advisor gets all the credit for publishing and none of the blame for retracting. That’s why some labs have been hit by multiple retractions (clearly a indicator of fraud culture) but only the low level researchers are punished

The problem is that academic publishing isn't tied to any real fitness function.

Unlike a startup or company or some other thing where you vote with dollars on utility, many papers can't be quantitatively evaluated on how much they contribute to science.

Some papers are obvious and groundbreaking, but a lot of research is at such a frontier that literally no one else can evaluate it. You just nod your head and trust it.

That, coupled with the fact that negative results just aren't published, is incredibly inefficient and disheartening.

That dynamic is deeply broken

The widespread belief that it is rare, is an invitation to fraudsters. Who have therefore wound up in such positions as president of Stanford. It is hard to catch them because people's a priori belief that there isn't much fraud makes it hard to accept that any particular researcher could be a fraudster. Which greatly reduces the risk of being a fraudster.

It’s also just not fraud.

Well meaning students have little alternative than to publish journal papers showing how their algorithm/method had a 0.001% improvement (measured by questionable metrics) because they are trying to get N publications to graduate.

While discourse about such methods can be beneficial, it also tends to create too much noise that drowns out more meaningful work.

Just to clarify, I think fabricated data in top biology and medicine journals is relatively rare because it's typically high-throughput data, and you must archive it. It's just too much trouble to fake that.

However, another equally damaging form of fraud, cherry-picking data and manipulating statistical inference is common. I think the reason is that it's much simpler to get the results you want this way, and you are much less likely to get caught.

Previously, fake data was common, e.g. in the form of fabricated Western blots, as it has been proven by large-scale image analyses conducted by sleuths. I imagine similar forms of fraud are still frequent in other disciplines or in domains that generate low-throughput data, such as some clinical trials.

That's all true but I think what you meant to say is that successfully detecting fabricated data is rare. I've been in the weeds on this and after a while it's tempting to believe you can always spot scientific fraud by reading papers, but of course it's not so. A paper is just a bag of words and pictures, it can say anything it wants. As long as a data fabricator doesn't make basic errors you won't detect fraud.

Unfortunately, when people do find ways to automatically detect inconsistent numbers in a paper, like SPRITE or GRIM, they find maybe 50% of all checkable papers fail the audit and the authors almost invariably refuse to share their raw data for further checks despite having previously agreed to do so. So it seems likely that fake data is actually not rare at all but very widespread.

Caveat: like I say in every comment on scientific fraud on HN, all this varies dramatically by field. Some fields are much more corrupt than others. That 50% number is from psychology. Computer science isn't that bad relative to others. But for example, in climatology it's taken as axiomatic that it's OK to fabricate data and then present it as "observations". Anyone who tries to call them out on this behavior gets attacked, censored or even sued, yet if you did the exact same thing in physics nobody would hesitate to call it fraud. There isn't even a consistent definition in academia of what data fabrication means.

So no, sadly we don't really know how prevalent data fabrication is. The only way to detect scientific fraud to the level of robustness we'd accept for any other kind of fraud is regular, randomized lab audits with jail time or ruinous damages for perpetrators, backed by rigorous field-independent written standards for what evidence must be provided, all checked by outside organizations. Think financial audits but for scientists.

Obviously not only is academia not doing this, it's culturally nowhere close to considering the possibility of even thinking about starting. Academia has spent so many years engaging in such deep ideological purges that everyone who remains is at minimum sympathetic to ideas like "defund the police". The concept of policing themselves to the level accountancy does will never come up, not even during discussions of reform.

To me it's quite simple:

  * If there's no clear way to get the raw-data 
  * If it's unclear how the data was gathered   
  -> the results are most likely fabricated

One strange psychological effect I've seen is that especially in medicine some people themselves spend a lot of time rigging data to get to their PhD, but in the end they absolutely trust in anything published in the field. "The science nowadays is so good - medical scandals today are completely impossible"

>In climatology it's taken as axiomatic that it's OK to fabricate data and then present it as "observations".

Do you have a source about this?

There are many detailed discussions out there. Go to any skepticism website and search for terms like TOB adjustment, homogenization, "the blip", "the pause". I'd give you links but without fail HN commenters just engage in ad hominem attacks on any source given without bothering to read them. If you want more details and can't find any given this info, email me and I'll send you starter links.

tl;dr where does temperature data come from? Climatologists gather data from various sources and then aggregate them into time series which are then used for later research. Which would be fine, except that along the way they do things that aren't considered legitimate in other fields like:

• reporting "observations" in their "raw data" from weather stations that haven't existed for decades. They are actually software-generated guesses based on other stations that themselves may not exist or be in comparable similar surroundings. This isn't a rare thing, in some countries most of the data is attributed to weather stations that don't physically exist or have moved large distances, yet still show up in databases with readings at their old coordinates.

• adjust their "raw data" in various ways and for various reasons. That is, what gets presented as observational data in climatology - like graphs of temperature - are not actually readings from thermometers, although it's usually presented as if it is. Modern temperature time series have dozens of adjustments including many extremely questionable ones like homogenization (a form of spatial averaging).

• dropping confidence intervals or even reporting temperatures that are simply the max of the CI around the real reading. This is a problem because QA on weather stations is frequently non-existent. In the UK over 80% of weather stations are WMO grade 4 or 5, meaning they are junk tier with uncertainties of 2C and 5C respectively. As claimed warming is 0.1C/decade this sort of data is of no use for detecting it but is used anyway. It's a global problem not UK specific.

• they regularly rewrite temperature time series in ways that invalidate all prior published papers and claims based on that data, but they don't retract any of those papers.

Editing data points, making up readings from non-existent instruments, using garbage-tier instruments, ignoring confidence intervals, playing with the data until it comes good and retroactively deciding your data was wrong but not doing retractions of papers built on it, are all behaviors that would be decried in most other fields.

Until academia can settle on universal standards for what it calls science people's trust in it will continue to fall, because they won't be able to make any assumptions about what kind of methods the word science really means :(

> I'd give you links but without fail HN commenters just engage in ad hominem attacks on any source given without bothering to read them.

That's a red flag. Reads as evasive to me. Why are you bothered by internet randos disagreeing with you? If you have reputable sources link them instead of wasting time. If the responses are truly ad hominem that should only strengthen your argument in the eyes of any making an attempt to be objective.

I apologize if it seems uncharitable but my default assumption in this sort of scenario is that you've been the subject of well reasoned rebuttals and don't want to confront that reality.

> universal standards for what it calls science

> what kind of methods the word science really means

That's nonsensical. "Science" is a vague, very high level methodology. There's no single correct way to go about things other than to recursively apply evidence driven approaches.

Some fields have broad applicability of course. For example something like the foundations of statistical methods will presumably remain the same no matter what task you apply them to. But what the results of such methods "mean" is going to vary widely.

No, I've never seen a well reasoned rebuttal. The type of reply you get here on HN is always exactly like yours: an insistence on arguing the who instead of the what. Hence your requirement that any given sources are "reputable", without specifying what that means. Based on experience it's guaranteed to exclude 100% of all sources that discuss the issue honestly. That game has played out here 1000 times and is boring. If you disagree, check a fact or two. Go find sources that you think are reputable instead of expecting others to guess and read a few, poke holes in them if you can. That would be interesting.

> "Science" is a vague, very high level methodology. There's no single correct way to go about things

It's really not. But this is the kind of semantic dispute I am warning of. If your response to discovering that academics engage in data fabrication is to say that's ok because science can mean anything, then science doesn't mean anything. And eventually the general public will learn that, and vote to defund it.

> the foundations of statistical methods will presumably remain the same no matter what task you apply them to.

Overfitted models are not only common in many fields but academics often try to justify them as OK to use, so I wouldn't assume statistical methods remain the same no matter what task you apply them to. After all, science is really vague, right? Who is to say what the foundations of statistics are?

Maliciousness, incompetence, and accidents all look EXACTLY THE SAME from a replication perspective. We can't tell the researchers' intent.

Until the "industry" (defined vaguely as scientists, their institutions, universities, funding entities, etc, etc) cleans house and punishes those researchers, we're quickly approaching a time where we'll have to take EVERY study skeptically until it can be replicated.

* Punishment could range from "no, we won't publish your stuff without data+methodology" to ratcheting back funding to "we publicly document your lying/incompetence" (hardest and riskiest) to a variety of other things.

> we're quickly approaching a time where we'll have to take EVERY study skeptically until it can be replicated

I've always felt like this should be the norm. Why would you trust something before it can be replicated? Even if it's unintentional, people make mistakes.

Individually, yes I've been there a LONG time.

Unfortunately, we have a media and political structure that uses the most recent study/model/whatever to advocate for, design, and enact policy before that review.

> I've always felt like this should be the norm. Why would you trust something before it can be replicated? Even if it's unintentional, people make mistakes.

If you are close enough with a scientist, generally they will admit they don’t trust a single study

Some fields also have guardrails, such as the LIGO being two separate detectors a with two independent teams

There isn’t really a way to “clean house” on a large scale, everyone has to somehow be more virtuous and not lie to themselves when p hacking or publishing data they know someone wouldn’t be able to reproduce.

no they don't. sometimes incompetence can stumble on the correct answer. like you could repeat the experiment, get very different results, be convinced of the charlatanism and incompetence of the original reporter, and then sigh a huge sigh of disgust because the data may have all been wrong but it points to the same overall/big-picture conclusion.

> Maliciousness, incompetence, and accidents all look EXACTLY THE SAME from a replication perspective.

Why do you think that? If something fails to replicate it, you can investigate the original paper, and then there may be very clear evidence of fraud.

I do think that you can tell apart these cases. Outright fabricated data is very different from p-hacking, which is very different from a meaningless garbage paper.

On the other hand, convenient laboratory errors might be hard to tell apart from fabrication (though not always, sometimes people get caught using photoshop), and statistical incompetence might be hard to tell apart from p-hacking. So I agree this is why it's hard to punish fraud.

I just recently found an incredibly cool paper that solved a problem I was having. It was published in a tier A conference. They claimed they would make the data and code public at a later point.

Problem is, I could find nothing of it when I googled. O tried to get ahold of the authors and I even wrote to their department. The only Mail back I received concluded that neither would be available any time soon.

So there‘s that. A paper that could be completely made up, with no data to support it. It might as well be make believe. I would lie if I said this didn't make me lose faith in my own work, since it justifiably takes me ages to gather data and write code for my experiments. Great.

Would you consider the Reinhart-Rogoff paper: https://en.wikipedia.org/wiki/Growth_in_a_Time_of_Debt#Metho...

Where a mistake in the Excel spreadsheet was used by many politicians to justify austerity measures to be a #2 or #3a problem (or both)?

Problem 3 has another effect. There's a whole industry of conspiracy theorists who will point to that paper to give credence to their own BS, while they create Youtube videos and wild pop-sci articles to influence the general public on whatever their agenda is.

I have a friend who recently has 'taken an interest in science' and is now passing Youtube videos all of the time that are total crap, but make reference to 'scientific' papers. I 'learned' how earth is already doomed because Betelgeuse went supernova, and that there are deep shafts below the pyramid of Gizeh that only an alien race could have dug. Stuff like that.

Net outcome where it comes to science may be that this helps erode people's trust in science.

It really doesn't matter what the motive is when the effect is the same. Treating it as an individual moral failing isn't going to work when the incentives to fake papers (wilfully or not) are so strong. Any fix that works for 1 is going to work for 2 as well and vice versa.

You can’t legislate morality.

Why not? Set up the laws to incentivize morality.

You don't even have to change the laws. Just changing the social norms would be enough. At the moment our social norms are that the ends justify the means. If you make money, or produce something that looks like a breakthrough, that is all that matters. How you did it, or whether the thing you did has actual value, matters not at all. As long as you've provided yourself with plausible deniability, that's good enough.

I know this because I've been the beneficiary of this system on multiple occasions, and it makes me sick. This is not the kind of world I want to live in.

Most if not all legislation originated in morality at some point. We've just collectively forgotten that's still an option today.

No you can't legislate morality. But operationally we want accountability commensurate to control. As others have pointed out, there must be consequences to bs. With that in place, we don't have to legislate other than point out consequences. Now to be a pedantic, consequences are a kind of legislation, but at least it's about behavior and outcomes not intent. It would elides certain issues out of scope.

Unfortunately the line between 2 and 3 is very blurry. Like paper mill output submitted to legimate medium-to-low impact journals. Some of it gets inevitably published, others just puts a strain on peer reviewers.

About 3a: Never touch MDPI, the amount of fake data I've seen is ridiculous

Is 3/3a about to be more serious with LLMs in the mix?

"about to" it is.

In a recent conversation with the editor-in-chief of a journal I am on the editorial board of, a substantial bulk of the submissions we get are LLM written papers that essentially randomly look for associations in accessible data, which are then sold to faculty (primarily in China).

Arguably "randomly look for associations in accessible data" is like 90% of academic publishing even before LLMs though ;)

No, probably Not. Nobody is reading these Journals anyway. It's only good one resumes. i think even 3a is Not a problem because fake papers will follow a specific pattern in Meta Analysis. Should be catched in one of the "filter" stages during Meta Analysis.

probably not. the writing quality will improve but LLM-generated papers will still be ignored.

to the extent they aren't ignored, but seem so plausible that they are taken seriously, eventually people will want to talk to the researchers about their results, invite them to give talks, and so on. at which point it becomes problem 1.

4. Plagiarism

Or cheating in other ways, like collusion rings [1], reviewers using LLMs to review, or authors putting text in their papers so reviewers using LLMs to review will give them a favorable review [2]

[1] https://cacm.acm.org/opinion/collusion-rings-threaten-the-in...

[2] https://www.theguardian.com/technology/2025/jul/14/scientist...

I did a meta-analysis for my masters thesis and I was convinced that 2 of the 39 papers in my dataset had fraudulent data.

> Problem 2 is amenable to reform for the most part (fields are already doing this)

how are they?

Problem #3 is relatively easy to address: shift from journals to conferences. Organizing fake conferences is a lot harder than setting up paper mill journals.

Conferences aren’t double blind. They aren’t even single blind. Fundamentally they can’t be

Peer review should be but sometimes isn’t. It always can be though, unlike conferences

Not really. I see many, many fake conferences adjacent to my field each year. I don’t really see why setting up one would be harder than creating a fake journal.

I think they mean predatory journals and conferences. Setting up a predatory journal costs nothing, whereas actually holding an in-person conference of any kind costs money.

A good time to repost about the famous VIDEA conference:

https://users.cg.tuwien.ac.at/wp/videa.html

https://users.cg.tuwien.ac.at/wp/videa-paper.html

These conferences have fees, just like predatory journals. They tend to be held in dodgy places in otherwise touristy areas, and most of them are partly or fully online anyway. The cherry on the cake is that attendees can use them as holidays funded by their institution.

I don’t dispute that there are probably orders of magnitude more fake papers than fake conference presentations, just that shifting from journal articles to conference talks won’t help at all. If the incentives shift, current paper mills will just shift with the market.

I can tell you from personal experience that organizing a "fake" conference is beyond trivial. All you need is a topic that is narrow enough that all of the attendees know and like each other. I was an attendee at several such conferences back when I made my living publishing papers, and I was astonished at how easy it was to publish bullshit as long as it was the right venue, and how borderline impossible it was to publish anything that I considered to have actual value. (In my career I published dozens of conference papers, but only one journal paper and one book chapter.)

In actual practice what I found was that the principle driver of publishing success has absolutely nothing to do with the quality of the work, it has to do with how much your reviewers think that you might some day be in a position to review a paper of theirs. This is the fundamental problem with peer review when: career success is measured by quantity of papers published, the resulting dynamic is governed by game theory, not scientific merit.

I get "invited" to fake or semi-fake conferences all the time. Also fake or semi-fake "awards". I just have a programming blog. I would not expect any other field to be any different.

Thanks for the breakdown.

If we can't rely on journals like Nature or NEJM to catch blatant fraud, then what hope is there further down the chain?

You can't rely on journals because they have bad incentives. Their customers are the universities they're supposed to monitor for fraud, and their prestige comes from the same place news/magazine prestige does - publishing incredible claims.

Catching fraud further down the chain is easy. Other professions know how to do it. Just adopt the same framework as used in financial controls. External auditors, written standards, randomized spot checks, lots and lots of paperwork to provide audit trails back to real experiments with hard-to-forge evidence. And severe punishments for those found to have engaged in intentional fraud.

That sounds expensive and difficult. Science is not a single field with a single set of practices. Regulating it is more like trying to draft a single set of regulations for finance, agriculture, pharma, and software.

You can get a glimpse of that when applying for grants or publishing in more prestigious journals. There are often endless checklists and questions about regulatory and ethical compliance, data availability, reproducibility, and whatever else. And more often than not, many of the questions are category errors or otherwise make no sense in the context of your work. Even the people who are supposed to regulate science lack the ability to do so in a way applicable to the entire breadth of science.

I don't think the differences between fields in (academic) science are as big as the differences between agriculture and software. They all take grants to publish papers, in journals, and they are measured via similar metrics like citations, papers published, impact factors and so on. Academia is highly homogenous across institutions too.

Yes, for sure a lot of controls implemented via audit aren't perfect fits for what is being regulated, that's the nature of trying to fix bad incentives via regulation. It's better to just have aligned incentives to start with, but that can't be done in academia, it's regulation or nothing.

Also, my experience has been that sometimes academics like to claim their field is special and shouldn't be expected to play by the same rules as other fields. But that's often not true. They should in fact be held to the same standards as everyone else is and their justifications for why they're an exception don't hold water.

The academia is far from being homogeneous. Institutions in the same country are usually similar, but things vary a lot between countries.

Science is fundamentally an activity, not a field. You can do scientific research in most fields. Agriculture and software are nice examples, as academic scientists are doing both. And in many cases, the outputs are used by the general public.

Quantitative metrics such as citations, paper counts, and impact factors are mostly used by bureaucrats who don't know what they are doing. Actual evaluations in reputable institutions depend heavily on the subjective judgment of other people in the field.

The role of grants varies a lot between fields and insitutions. In some cases, a professor is expected to become a manager, win grants, and hire other people to do their research. In other cases, grants are nice but optional. Such professors can do research with limited external support, both personally and with students funded from other sources.

Reply on news.ycombinator.com