Hacker News Reader

Comment by vouaobrasil | Hacker News Reader

It may be a person's instinct to immediately blame the researcher, and of course, I definitely think data fabrication is never justified.

However, what is the societal problem here? It's the problem that in academia, there are such pressures to publish and be first nowadays that many people will do anything to get ahead. It doesn't help that social media and news sites have introduced a sort of gamification into it.

People have their natures, and the current state of academia encourages cheating, plain and simple. Just like the current state of finance encourages market manipulation.

If we want to fix it, simply outing these fakes will not be enough. While I don't have much sympathy for the cheaters, I also don't have much sympathy for the scientific community that has created this environment.

View on HN

>It's the problem that in academia, there are such pressures to publish and be first nowadays that many people will do anything to get ahead.

If you can not do honest work in academia you should leave and go work somewhere else.

Fraud is not the correct reaction to the problems of academia, as it actively hurts it even more. Nobody is forced to do academic research for a living, especially if you have a degree in physics.

> Fraud is not the correct reaction to the problems of academia, as it actively hurts it even more.

Not in the short term. And only for yourself personally if you are found out.

Of course there are perverse incentives which goes against the scientific spirit. It doesn’t even require malicious intent. Neglect and bias can be similarly powerful negative forces.

I think we’re getting to competitive sports territory, where people justify it to themselves because they’re convinced everyone else is doing it.

I never said fraud is the correct reaction. Like I said, I believe it is wrong unequivocally.

I did say that fraud is something that is essentially encouraged in academia. Nobody is forced to do it of course, but then again, any system will have its flaws that encourages bad behaviour, and it is also advantageous to restructure the system to minimize that.

If you asked your CISO about securing your system against hackers and their response was "If they can't do honest pen testing, they should leave IT and go work somewhere else." what would your feelings be? OP wasn't trying to justify bad actors, but explain how the system is selecting for them, and we have to accept that our system is running in a world where there will always be back actors present. Systems must be built to be resistant to them.

Any system that relies on individuals adhering to the honor system will be abused. History has shown this repeatedly. We have created an institution that has all the wrong incentives in place. This institution will continue to churn out bad actors who will just get better and better at hiding their trails.

It's not the correct reaction but it's a predictable reaction. Academia has some huge problems with how they measure success and how they police academic integrity. The incentives make this outcome likely.

There are no steroids in baseball.

Steroid users are few and far between, and we must weed them out of baseball.

Nearly everyone is using steroids, but it was their choice, so it's their own fault.

Another lofty comment about honesty in academia. No research job lets you spend 6 months redoing stuff just to make sure it’s 100%. It’s about publish what you have with enough disclaimers. It’s about convincing the reviewers to get the paper in. Rest will be resolved later.

Classic cases are the Bell labs guy finding organic semiconductors and fermi lab guy finding new elements. They both were just making up stuff which they believed existed but just needed work to be found and published. They would make up experimental data to support existence of theoretical things. The assumption was just publish this first, get the accolades then use it to get money to get someone to eventually do this properly. But they just predicted stuff which weren’t true. Nobody knows the amount of bullshit that was made up that eventually turned out to be true.

All of this because scientists these days are not expected to work on things. They are expected to produce results. Nobody is giving passionate people time and space to explore things. It’s all about results now.

The irony is that the reward for running the rat race is freedom from rat race. Postdocs publish random crap to boost numbers for their tenure which they think will relieve them of this stupidity and let them focus on pure research. but once they catch the tigers tail they gotta keep running.

> No research job lets you spend 6 months redoing stuff just to make sure it’s 100%.

These jobs do exist[1], and we as a society need to figure out how to stop organizing research around a single manager (professor) with zero meaningful oversight.

[1]: https://www.nature.com/articles/leu2016193#

This is a fine position to hold, while excluding more-useful ones as "condoning bad behavior" or "excusing bad behavior" or whatever, if you want to judge people but not solve any problems.

It's true, and also entirely unhelpful.

Remember when Lance Armstrong got in big trouble for taking performance-enhancing drugs for the Tour de France, and he was lambasted in the media?

That might have been right and proper. But years later, I read that of the top 30 finalists that year, every single one was taking performance-enhancing drugs. Presumably the same was true in previous years. Armstrong just happened to be the first-place finalist in a year where they finally had the ability to start testing.

In a case like that, it's not enough to say "if you can't bike fast without steroids, then biking isn't for you!" If cheating is pervasive, those who can't or don't cheat will fall behind and quit, and all that's left will be those willing to cheat. From the POV of Science writ large, if the culture is rotten then the incentives must be changed. You can't rely on the independent integrity of every individual not to cheat, or you'll just end up with only cheaters.

The issue that the OP pointed out and that your comment completely failed to address is that the way the system is set up is that the pressure to produce either pushes academics either towards dishonesty or pushes them out.

People who are committing fraud aren't trying to have a correct reaction to the problems of academia, and leaving fixing the problems of academia to them changing their behavior will have the outcome you should expect.

Telling people to be good is a waste of time unless they're your kid or loved one and they really don't want to disappoint you. Fixing your system so it doesn't reward people who commit fraud is the only thing that even has a possibility of working.

The president of Stanford was caught pressuring his team to fabricate data for years, so this is likely pervasive everywhere in academia nevermind much higher pressure environments with lots of competition like elite institutions in India and China.

Especially because many of them have even less of a stigma against academic dishonesty than in many Western systems.

Edit: not to suggest that is anything inherent about people from those cultures.

I think it's more to do with people in those environments not having good alternatives that creates incentive to do anything to get ahead. If we do some self reflection, similar things can be said about Uber & Airbnb et al. Painting this as cultural phenomenon is not productive IMHO.

> However, what is the societal problem here?

I would say yes. I think we need to place a lot more value in publishing research that isn't surprising. Research that simply replicates (or fails to replicate) other research should be just as valuable to one's career as finding something unexpected.

I scientist that has replicated a bunch of studies and identified several that they couldn't replicate is doing something very very valuable for the human race and we (universities, grants, etc.) need to recognize that.

A fundamental problem with research academia is that we scaled it up massively without introducing more division of labor.

Interesting. I'm not sure I fully understand what you are saying though. Can you expand on that?

It's all throwing PhDs at novel research. Boring reproducing known things is just one type of support work that is neglected. There is a vast "supply chain" of work that allows novel research to be done, and we've neglected it.

I can see how that can be the case if it's for survival, but no one pressured this person to try to claim to find one of the most sought-after technologies out there. It also doesn't apply cause he voluntarily chose to go to a prestigious university where this kind of pressure exists. Competition can be healthy thing as long as it doesn't threaten the researcher's career, which clearly wasn't the case here. This guy had the balls to do it twice in two years after getting caught, clearly he's set in his ways.

Of course. I am not trying to remove the blame from him. On the other hand, we can also learn from these problems because they often point out nearby flaws in the general system that should be ameliorated. That was my main point.

I agree! I'm just saying that this case in particular does not really show the problem with pressure in Academia. If anything, it shows how lenient the peer-review process is.

Yeah that's true, that's something to think about. The peer-review process is indeed lame and weak.

"When a measure becomes a target, it ceases to be a good measure."

The publish-or-perish mentality combined with the grant system introduces a lot of bad incentives into the scientific ecosystem.

> I also don't have much sympathy for the scientific community that has created this environment.

Scientists didn't create this environment: you did. We, as a society, demanded that some sector produce new scientific research at a rate and level that is capable of supporting our economy, and then we allocated a very restricted set of resources for them to do it. And they've done miracles, too! We would literally not be alive today if the scientific community hadn't developed the improvements that allow us to feed, house, and provide energy to 8 billion people. And that's before we get to strictly optional stuff like the computers and networks that power this website.

It's a serious business and obviously that kind of pressure leads to bad results sometimes. But the correct response is to be grateful that we have an engine in our society that produces this basic research (since industry has dramatically lowered spending on basic R&D over the past decades) and try to improve it, not to be pointlessly angry at scientists for living in the system that society's incentives have built for us.

> It may be a person's instinct to immediately blame the researcher, and of course

However as they teach the little kids, trust your gut instinct. This person didn't just plagiarize this work, but also his PhD thesis: https://www.science.org/content/article/plagiarism-allegatio...

> If we want to fix it, simply outing these fakes will not be enough.

Yes, but without outing these fakes and harshly and punishing them as the first step, we're immediately communicating that this is "ok". "Just fake the data and throw it over the wall, if you get caught, just blame it on pressure and the general academic culture".

This post seems almost completely vacuous to me. I guess you could read it to mean that if a person cheats, there must be some "societal problem" causing it. That would seem to imply you believe society could in principle be 'fixed' s.t. no one is ever incentivized to cheat. Is that what you mean? Or simply, s.t. fewer people are incentivized to cheat. In that case, duh. But fewer to what extent? And are there any actual constructive suggestions forthcoming?

Academia is zero sum. All zero sum endeavors result in this kind of behavior.

Can you elaborate what you mean with zero sum?

There is a finite number of grant dollars and faculty positions. Your colleague's success hurts your chances of success.

Reply on news.ycombinator.com