Can anyone give me a non-meme reason why I shouldn't fake my data...

Because when your experiment is repeated and your conclusion isn't consistent you will get a bad reputation

>a non-meme reason why I shouldn't fake my data?
So you don't produce false conclusions and set humanity back. Intentional falsification should be a crime punishable by hanging, drawing and quartering.

Ethics.

If humanity is capable of being significantly set back by people submitting falsified data then the real problem is the system being way too fragile, not OP being a predictably lazy shortcut seeker. If anything it'd be better for more people to introduce shit data so the systemic problem is more obvious and more robust checks are established.

>Chinks are going to do it.

Implying that everyone isn't doing it already.

Specially in Social Sciences, Biology & Statistics.

I have seen it happening even in Experimental Physics.

The theoretical physics (except String Theory) is the only branch of science you can trust without doubt.

I made up some data for my Masters thesis because I didn't have time to collect enough participants and I wanted a good grade. Never tried to publish it or anything but I still feel guilty about it. It will be a black spot on your consciousness.

The embarrassment of getting caught is soul-crushing.

This is what scientism looks like.

(OP)
So you'll get promoted fast -- and then have everything crashing down when you're caught.
There was a guy at Bell Labs. Very respectable solid-state physicist. Already in the senior ranks. Got some absolutely amazing results.
Which couldn't be duplicated.
What don't you understand?

If you're doing shit work then no one will care. It might be years before the truth outs, if it ever does.
If you're doing something important -- something good enough to get you the attention you dream will boost your career -- it'll be followed up immediately.
And discredited immediately. That 36 percent is an average. (I'm assuming you're in some "real" scientific field, not some wishy-washy area where even out-and-out fakery can be submitted to, and accepted by, the journals in the field.)

That guy from Bell Labs. He had a solid record. He was caught faking it and fired and never worked as a scientist again.
No one ever figured out why he did it. His legitimate work was fine and he didn't need to make it up.
And, of course, all the honest work he'd done had to be re-checked because no one knew just when he'd Turned to the Dark Side.

>I'm assuming you're in some "real" scientific field,
And that's where you're wrong. From what I've seen most people are caught because they either produce perfect results or fake data for something that would be so groundbreaking that it forces people to look closer. Would be interested if the Bell Labs guy fits into these categories.

I will personally tell on you.

The Bell Labs guy "produced" something groundbreaking. Since it was discredited, I've forgotten exactly what it was.

Can we ask what field you're in?

I'd also like to add that, if you're in science for fame and fortune, rather than hoping to discover something (useful or not), you're in the wrong field.
Also despicable.
If you go ahead with your nefarious plans, don't complain you weren't warned.

Social psychology.
Chill with the disparagement, I'm just a person asking a question.
And I'm definitely not in it for the fame and fortune. I just enjoy teaching and would like to make the research part of my job a little (a lot) easier

I can see getting away with it -- maybe forever -- in some social science fields.
Not sure how it applies to statistics. Faking the data or fudging the analysis?

In Biology, probably not. Not if it's anything significant. There was some Japanese guy, about a year ago, claimed he could produce stem-cells en masse with a simple procedure. That would be revolutionary. The hold-up in research is the difficulty in getting stem cells --- even ones that are destined to go in the wastebasket otherwise. It's like Burke and Hare. The laws blocked life-saving research.
The Japanese guy was famous for maybe a month, and then he was out on his ear.
Again, why?

I know a good use for statistics. Suppose that's right, and only 36% of studies are replicated. 64% are not replicated. Doesn't necessarily mean the studies were faked. People make mistakes or they have bad technique. But every "not confirmed" will draw suspicion and, after a certain number of failed experiments, people will just stop believing. You're either faking it or you can't be trusted to do the work correctly. Bye bye, promotions.

Things have been tightening up. Corporations are notorious for running studies and suppressing the ones they don't like. "Our study showed that smoking DOES cause cancer, boss." "Burn it and we will never speak of this again."
Journals (not the "we'll publish anything if you'll pay us" kind) have begun rejecting papers from any scientist who didn't file a summary of his theory and procedure BEFORE starting work. Too many people have been reporting only "favorable" outcomes and burying their failures.

That some people cheat is not new. The NUMBER of cheaters is, and something is being done about it. It's like athletes had a better chance of getting away with steroids before it became an embarrassment and testing became mandatory.

I still don't see the "advantage".
Do a study of whether playing violet videogames makes it more likely that teen-age girls will try to kill their friends.
It does or it doesn't. How is making up the numbers (as opposed to mailing out questionnaires) saving time & effort?

Faking plausible data is usually harder than doing it for real. You may have heard of the experiments where some students flip a coin 100 times and some just write down a "random" sequence of 100 Hs and Ts. The fakers can be spotted immediately.

Source please. Googling couldnt get me to it

Send your null result data to a mediocre journal. You'll have the moral highground in future discussions on the matter.

Go back to China,

>Not sure how it applies to statistics. Faking the data or fudging the analysis?
Add a few data points to tip the p < 0.05

Data collection takes enormous amount of time. We can't usually just mail out questionnaires to random people. Senior staff usually delegate that part to grad students (who in turn fake data themselves) but I don't have that luxury
That's a good advice actually

Do people in your field really throw null results in the wastebin and publish only the p < 0.05 results?

>publish a paper
>nobody ever reads it besides the review board
feelsbadman

I remember that, the asshole who claimed he could revert blood cells to an undifferentiated state despite them not having any DNA.

Its actually what intellectual integrity looks like, even in the best case scenario because of the massive number of papers published every day the peer review process barely functions. Retards publishing imaginary data just makes it worse, eventually to the point of peer review meaning almost nothing and its certainly arguable we are already at that point. Which is a real fucking problem because its the only system in place currently to disseminate and curate scientific thought. In case you are a normie who doesnt actually know anything about peer review beyond the name, its destructive effect in conception, execution, and run on affect is similar to counterfeiting.

thats not real science you HAVE TO LEARN AND MAKE THE WORLD SMARTER DONT JUST FAKE SHIT

>peer review peer review peer review
Reproduction is more important than peer review in experimental work. Who is going to catch your loose cables just by reading the paper?

So long as it's not medicine IDGAF, fields like physics are long dead, no harm in keeping the illusion of progress going.

How are you going to learn about the original results, in order to replicate them should you choose, without a peer review process? How are you going to choose which experiments to reproduce given that resource is limited?

Physics is booming right now because many of the theories people subscribed to in the 60-90 and even early 00s are largely defunct and theoreticians are scrambling to develop new frameworks to explain and predict current and future results respectively. Experimental physics is never dead on account of the incredible complexity of many modern experiments.

This. If all your data turns up no meaningful conclusions, people won't necessarily think you faked it (although they might if you're a Veeky Forums posting autist). They'll just think you can't collect good data.

What area of physics are you talking about?

What is no meaningful conclusions? Null results are still meaningful.

>my field already has the lowest replication rate in science
Source?

I take it that if I say integrity you'll say that's a meme then?

Yes? If you get null results you can forget about publication, even if the null results are in the parts of the experiment that are direct replications of something else

science.sciencemag.org/content/349/6251/aac4716

Not a meme, but when you consider it on balance against all the positives on the other side it doesn't hold, at least for me

You're a chink right? It's just wrong. A gut feeling.

>muh -ism
Faggots abusing others' trust for personal gain are disgusting. Demanding intellectual integrity does not mean elevating science to a religion. People fudging data to get money and approval are mocking the scientific method and should be brutally punished to make them stay the fuck away from the field.

>Not a meme, but when you consider it on balance against all the positives on the other side it doesn't hold, at least for me
Fair enough, but if you ever find yourself old and grey and having to take sleeping pills every night to keep yourself from staring at the ceiling for hours and questioning your life's choices then remember this thread and remember this warning.

Will keep that in mind

>Can anyone give me a non-meme reason why I shouldn't fake my data?
Fate catches up with everyone. Noticed most civilisations have a concept of fate? The reason is that socoety relies on trust and ethics and if you break those there is a strong reason to hunt them down like crazed dogs.

>Getting published a lot early in your career is a huge boost,
True

>and the risks of falsifying data are practically zero.
Today yes but big data has already uncovered a lot of fraud. Suddenly a lot of researchers cannot find their original data.

I was in a group that uncovered someone who had falsified a major discovery. His career imploded during his presentation.

Also follow retraction watch.

>That guy from Bell Labs. He had a solid record. He was caught faking it and fired and never worked as a scientist again.
That is justice.

Sadly Sudboe is sneaking back in again, thanks to a large network of friends in the right places. It is disgusting. Meanwhile junior post docs cannot get tenure.

>there is a strong reason to hunt them down like crazed dogs

Agree. The question is how well-equipped and prepared is science to hunt us down.

Patent attorney reporting in here.

When we file a patent application we expect not only patent literature to be cited against the claims but also scientific papers. That is especially the case in the medical and pharmaceutical fields. And getting fraudulent papers cited can be a huge problem where there are major financial incentives to check the cited papers. If we can determine those are frauds we can have them dismissed as non-enabling.

This is not a hypothetical case, I have done it myself and also attacked citations based on the alleged authors being convicted for commercial and scientific misconduct.

>science.sciencemag.org/content/349/6251/aac4716
Crikey!! It is as if nobody there paid any attention to Cargo Cult Science by Feynman.
calteches.library.caltech.edu/51/2/CargoCult.htm

Of course they didn't. For every sperg who actually cares about the truth there are five cancerous normalfags like OP who just want to pretend to be scientists for some material gain.

We shall hunt them down, one by one. retractionwatch.com/

>Caught Our Notice: Brian Wansink issues correction that’s longer than original paper

Well, well, well.

>he ain't wrong

If you’re climate scientist go ahead. If you’re in a real scientist go look up Obokata

Is she our girl?
en.wikipedia.org/wiki/Clare_Francis_(science_critic)
Someone is out for revenge on a scale not seen since the Count of Monte Christo!

Whoa!

Wackypedia is silently attempting to kill off Retraction Watch, and if succeeding they can wipe all articles relying on this source. Sneaky!
>en.wikipedia.org/wiki/Talk:Clare_Francis_(science_critic)

>Physics is booming
Let's see
Superconductors: Stuck at 133 K since 1993
Particle physics: 13 billion euro accelerator found exactly one particle
Dark matter: Some guy said it exists, zero proof ever
Gravitational waves: So rare and hard to detect they're basically useless
String theory: May as well be pulled out of our asses because it's impossible to verify
Fusion: Trying to reinvent some failed 1950's design, achieved nothing just like all the other failed designs

You are brutal, user. However there have been a lot of unconfirmed superconducting objects and not all teams have been frauds. I vaguely remember one team found a transition at dry ice temperatures. Still a long way from room temperature.

Solid state physics is booming. Semiconductors are safe bets and nano tech is doing pretty well.

Dark matter, dark energy is junk science. Sure, we see there is something wrong with present standard model but that sure s no excuse to brew up all these hypotheses that defy Occam's Razor and common decency.

String theory is the moral collapse in science, garbage that claims "elegance" is a valid substitute for the accepted scientific process.

Fusion has some things going on but cold fusion is too far out.

>Someone is out for revenge on a scale not seen since the Count of Monte Christo!
Not really, whichever one or more person is using that pseudonym is apparently just running a program that automatically identifies aberrations for most of what they send to journalists. Sounds pretty low effort actually, probably just someone who figured out they could write a program to do this and proceeded to start using it anonymously instead of trying to cash in on it by selling it as a proprietary software product. I'm sure they have some negative feelings towards these frauds but the frequent use of machine generated finds suggests to me they aren't fanatical about it.

Do you have any idea how difficult that "automation" really is?

Morality ? Honor ? dignity ?

>Gorilla poster
Ah. I should have expected this.

Don't want to get into a dick waving contest here and I won't claim to be an expert, but I do in fact have a healthy familiarity and appreciation for it, I work as a software developer for a security company and I've written ANN programs in the past (as in actually writing them in C++, not just bringing in a reference to some prefab tensorflow API or whatever it is people do nowadays).
I don't see the effort of building a program like that as necessarily having anything to do with the effort you might exert specifically because you hate scholarly frauds e.g. this person could be like me and just interested in programming and/or machine learning with the application of their work directed towards this replication crisis business just being an afterthought of something to do with it.

>perform 20 studies
>one comes back with p < 0.05
>publish that result only
>replicators who get a null result don't get published
wew lad sounds like your entire field is a fraud

What would happen if you tried to use research others had faked? Come on guys this is basic.

>There was a guy at Bell Labs. Very respectable solid-state physicist. Already in the senior ranks. Got some absolutely amazing results.
Jan Hendrik Schön, yes. Also his co authors were badly tarnished by this huge stink. His work caused a lot of resources to be poured into dead ends.

Exactly. Scientific fraud can cause pursuit of ideas that lead to a dead end while dropping what would have been fruitful avenues. This can cost lives. And few care.

Experiments need to have outcomes that can be reproduced. If no one else can reproduce your same results, it will eventually catch up to you and you'll lose credibility in your field. Especially if your data is obviously falsified after no one else gets anything near what you did.

The easier aberrations is reusing the same diagram with different labels or scaling. A more insidious trick is simply to omit readings that do not match the preconceived idea of what it should look like.

I am genuinely interested to know how you can detect that and please do tell. Running a working systematic scan across all published works is sure to cause massive fallout of academia and bring some long needed justice into academic careers.

Statistical analysis has already caused fallout in soft and pseudo sciences like psychology, economics and the like but in physics and chemistry the frauds are more skilled in manipulation.

Around here several frauds have been the most prolific publishers and the universities have not really wanted to look too deeply into this. So from uncovering the frauds to actually having them kicked out is also a long journey.

>Statistical analysis has already caused fallout in soft and pseudo sciences like psychology
Statistical analyses have made psychology more pseud if anything, since any moron can p hack a value to < .05 and then it's "empirical evidence"

He's not, it's pretty rare for frauds to get caught. Even then, the "punishment" is almost always just "sry guys, we take the article back". Look at the ETH prof and the people around him. Few months after his articles got revoked, he has another one in Nature.

I fucked up a statistics run on some data when I published. I sent the paper off for review, then realized my mistake. I sent in a revised paper with the correct statistics.

The editor permabanned me from submitting to his journal. If you submit fake data, all I can say is, don't get caught. There are no second chances in science.

>There are no second chances
Demonstrably wrong. See Voinnet et al.

>doesn't fake data
>still gets caught

>he has another one in Nature
What? He must have some impressive connections in high places to carry that one off. Unbelievable.