The study is a bit more interesting than this title suggests.

>An intervention aimed to discourage participants from choosing the cheating-enabling environment based on social norm information did not have the expected effect; on the contrary, it backfired. In summary, the results suggest that people low in moral character are likely to eventually dominate cheating-enabling environments, where they then cheat extensively.

In other words: we tried to tell the cheaters "you wanna cheat? fine, you can only play with the other cheating pieces of crap." The intent was for this to be a punishment, but it turns out they like it. The strategy backfired.

Don't dismiss this as a "well duh" thing, there's cool stuff in here!

Let's substitute cheating with something else in your sentence.

"So you wanna compete as a road runner? Fine, you can only run with other competitive runners."

See where that is going?

Cheaters don't care about other cheaters being there; they either think they can out-cheat them, or else that the other cheats don't matter: there is enough of a bonanza there that all cheaters can win, and in fact more is left if non-cheaters are eliminated.

Not only that, but being aware of the environment as being "cheating-enabling" means that when they lose the cheaters can blame the competitor having better cheats, rather than having lost because they're a worse player.

I win? I'm the better player!

I lose? They had an advantageous cheat!

wait, but no. the goal of the runner is probably to get better at running not just winning whatever the cost.

someone who wanted to win no matter the cost even if it made them over the long run worse at running would probably find way to disqualify better runners or some other methods that doesn't necessarily involve enjoying running.

the difference between what you're describing and the the scenario is a difference between instrumental and intrinsic goals

The intent was to create an analogy between the motivation to cheat and some other motivation (not involving cheating).

Let's assume that the goal of the cheater is likewise to get better at cheating, and that cheaters intrinsically enjoy cheating.

That is because the "good people leave" is not punishment, it is often reward for cheaters. They got rid of who they see as risk-averse petty nitpicking loosers.

Sure it's cool but I don't see what's unexpected about it. So long as the reward is the same and the cheating is mostly effortless, I don't see why a cheater would prefer to play against honest people and have to compete against genuine skill versus a cheater competing against other cheaters.

If the reward were a function of how honest the environment/population was and cheaters still preferred the dishonest environment, then I'd find that to be interesting... but if the reward is mostly the same regardless of the nature of participants, then I don't see why cheaters would prefer to play against honest individuals.

I also don't think this finding generalizes very well, for example when cheating still requires a great deal of effort on the part of the cheater. For example take a competitive sport that is notorious for cheating, cycling. I'd be surprised if cyclists who take performance enhancement drugs would prefer to compete against other cyclists who use those drugs compared to competing against cyclists who don't use drugs.

Find me a study that comes to that conclusion and I'll be a lot more interested.

Oof. Based on examining the publically-available code[1], I have one major concern with the methology and conclusions of this study.

The problem is that the “tell us about yourself” questionnaire that determines whether the study classifies someone as being “low in moral character” happens after the experiment in which cheating would or would not have happened[2].

This means that the results of the questionnaire are going to be affected by cognitive dissonance as people are rationalising why they made the choices they did during the “actually cheating or not” portion of the experiment. For example, it seems intuitively likely that people who have just cheated on a task are going to be much more likely to answer affirmatively to sympathetic prompts like “Cheating is appropriate behavior because no one gets hurt.” (an actual statement from late in the questionnaire to which the participant is asked to indicate their level of agreement).

By my interpretation, the study isn’t telling us that people who think that cheating is okay are more likely to maneuver into situations where cheating is possible; it’s telling us that people who have gone into situations where cheating is possible are more likely to tell us that cheating is okay. Which might not necessarily be the same thing.

I’d have liked these results a whole lot more if they’d asked people to fill in the questionnaire before giving them the opportunity to cheat, rather than after, to exclude the confounding effects of rationalising on the results.

[1]: https://github.com/bahniks/Experiment-cheating [2]: https://github.com/bahniks/Experiment-cheating/blob/master/S...

I think that a similar criticism could be made if they were given the questionnaire beforehand, in that the questionnaire primes the participants one way or the other.

Some cultures have a much higher tolerance for cheating than others. The consequences do kind of worry me, because that sort of environment propagates. Now you've got people dying because you cheated to get emissions certification, or safety regs are flaunted, and whistleblowing has very real consequences so this sort of thing goes completely unreported for years.

Friendly amendment: s/flaunted/flouted

Yes I suppose if it was flaunted it wouldn't have been a problem :^)

Broken Windows theory. People tend to do what they consider normal in an environment. Change the environment and the people change. Either because they, in this case the cheaters, change their behavior or go to another environment.

What is very interesting because the original Broken Windows theory is mostly discredited nowadays. So, this variation standing is informative.

I think the discredited part is mainly that presence of small crimes can increase likelihood of major crimes. There's still pretty good evidence that more vandalism, run down buildings, littered streets, etc. can increase the amount of nuisance crimes. So sure it's not going to cause more murders, and policy-wise it doesn't make sense to over punish minor crimes, but community cleanup could still be helpful for reducing things like petty theft.

Which I think is consistent with this, having shitty surroundings leads people to treat their surroundings shittier is pretty analogous to being surrounded with cheaters making people cheat harder. The leap is that this will somehow escalate to worse/more violent crimes.

Not sure if the article mentions them, but a couple environments I regards as "cheating-enabled":

* Almost any social media platform

* Online product reviews

Advertising? Politics? Regulations?

Advertisers are better at advertising their products then actually advertising, though teasing out how much of that is caused by awful companies with garbage products is tough.

American politics is a pay for access game. The majority of politicians have no power, instead the powers been concentrated in party leadership who steer the ship swherever their backers point.

Regulations can be great. Smog, asbestos, lead, BPA, PFOA, are truly awful. The problem is we need to fill cemeteries for every major piece of legislation because those causing the harm are allowed to fight in bad faith against regulations with teeth or that could reduce profits.

Our world doesn't need to be this harmful. Profit and growth are characteristics of cancer, and that's exactly how humanity is acting towards the earth and each other.

I always thought it would be fun to have an "escape room" style fake testing center where cheating is allowed and encouraged and the goal of the game is to pass the test.

The "tests" administered would be impossible lorem-ipsum style tests, but you would have access to the answer key beforehand, and the goal would be to pass the test without your cheating method being detected. Obviously memorizing the entire answer key is the only foolproof way to cheat, but if you limit the amount of time you can see the key, it would force more innovative methods. For example having a partner with the key transmit the information to you somehow. The testing center will try to thwart you by making you take the test in a faraday cage, no phones, etc.

Here's some work on "security mindset" teaching with a cheating-expected test where students have to write down 100 digits of pi.

http://www.rumint.org/gregconti/publications/KobayashiMaru_P...

And a remote teaching version during covid-times https://daveeargle.com/2020/09/11/kobayashi-maru-proctorio-v...

wasn't this the premise in Naruto's written exam test if I remember thing correctly?

The teachers expected them to cheat since the exam was made very hard on purpose and the only ones who got eliminated were the one who were caught cheating. they were testing their spying abilities not academic knowledge (without the participants being told so beforehand).

and if my memory is not failing me, they even placed fake participants in the exam room to be the source of the correct answers.

That was the first thing I thought of, if anyone is interested the test is episode 23 and 24 of the anime [1].

[1] https://naruto.fandom.com/wiki/The_First_Exam!!

To change the society, change incentives.

“Never, ever, think about something else when you should be thinking about the power of incentives.” — Charlie Munger,

https://fs.blog/2017/10/bias-incentives-reinforcement/

The Psychology of Human Misjudgment https://fs.blog/great-talks/psychology-human-misjudgment/

Rosy ways to explain this result:

1. Some participants are able to predict the result but when they are forced to act on it beforehand the entanglement collapses.

2. The negative correlation with charity spending is because the cheating participants know a better charity that urgently needs the money. Defrauding researchers is morally justified in this case.

Seriously though, that was a very interesting read.

So cheaters tend to gravitate towards situations where they can get away with cheating. What a shocker.

Yes, it's the system, not the people. Same why laws matter and not corporations "playing nice".

I see it slightly differently. It’s cheaters feel that even among other cheaters they feel they have a better shot and will avoid the meritocratic choice.

So it’s a bit of “why we can’t have nice things”. It’s not the system as much as people taking advantage of social permissivity for individual benefit.

> It’s cheaters feel that event among other cheaters they feel they have a better shot and will avoid the meritocratic choice.

Cheaters think they are the meritocracy. They think they are smart and resourceful.

Maybe? To me it seems they either don’t have enough $ability to do things the accepted way, or they want to do things an easier way, so one way for them to do well is to use a shortcut (cheating), either for lack of ability or out of laziness (not wanting to put in the hard work). But, it’s not all bad. ‘Cheaters’ the lazy type are innovators, they look to make things easier to achieve, so they will employ automation and the like to get out of drudgery, which leads to discoveries and progress, right?

Define "cheating".

Is it cheating or just being savvy if a farmer wires his new building to the 1999 NEC instead of the 2021 NEC since a copy of the 1999 one is $20 on eBay and omits some grounding requirements that are onerous for outbuildings.

Is it cheating or just following the spirit of the rule if a middle manager avoids a problem be skipping the process-required approval for a time sensitive change to something when it's exactly the kind of thing the higher up would approve in a heartbeat?

Cheating vs non-cheating isn't a binary thing. People have degrees to which they're willing to fudge stuff.

I don't think there is dichotomy between ability and cheating. The dichotomy is just too make us feel better if we lose against cheater.

Highly skilled people cheat to get little bit more advantage against other highly skilled people. And it is harder to catch them, because the ability demonstrated in some situations means everyone will go out of way to excuse them.

Highly skilled people cheating also means that only cheaters can rise to the top of competition.

I think cheating is a response to a sort of exhaustion with a gamified system... Like when you get tired of going through the grind on Gran Turismo and you enter a cheat code so you can get the R8 without doing the 24 hour le mans.

When you shift to things that aren't gamified, such as project based education, I there is little to no "cheating" because cheating is an artifact of the game design.

A good reply, it goes to a deeper level and I agree.

Maybe also related to gambling. People who feel they don't otherwise have a shot at getting something better or competing.

You can't change people, but can change a system. So it makes sense to blame the system instead of people.

Blacklists vs whitelists. Blacklists are better for society. I’d rather have default accept rather than default drop when it comes to laws.

It's the system, but not the system of laws. If leaders and elites constantly shit on laws because they think their status allows them to – then the laws won't make people honest. Because there are thousands of ways for dishonest behavior that doesn't break any laws.

I wonder how well this transfers. Instead of cheaters, bullies? Liars? General opportunists?

My gut is it transfers easily. Not clear on what to do about it.

always nice to have the data to confirm the hypothesis.

Please don't do this. Substantial comments are much appreciated.

The word you're looking for is "substantive", not "substantial."

subtantive: important, serious, or related to real facts

substantial: of ample or considerable amount, quantity, size

IMHO, it is substantive to point out that a study produced an obvious result that everyone already knew to be true.

Don’t put your words in my mouth, please.

I said substantial. As in “weighty, solid, meaningful”.

I'd say the comment runs afoul of the "Please don't post shallow dismissals, especially of other people's work." guideline.

Fair enough. So here's a less shallow dismissal:

This study seems to me to produce an obvious result that no reasonable person would doubt even in the absence of data, and is therefore a complete waste of time and resources. Am I missing something?

One could, of course, provide you with a stack of obvious ideas that no reasonable person would doubt in the absence of data...that are nonetheless wrong.

I doubt that very much. Can you give me a few examples? I can't imagine what such examples would even look like because in the absence of data I don't see how you could establish that they are false.

"...an obvious result that no reasonable person would doubt even in the absence of data...":

Aw, c'mon, you can do it, too. The Sun (and other astronomical bodies) go around the Earth. Time (and space) passes at the same rate for everyone, everywhere.

Here's another good one I just ran across:

"Let me tell you about some findings from the literature on divorce. People who start their marriages with the most love and affection for their partners tend to have a decline in satisfaction over time, but are unlikely to divorce, while people who had poor relationship quality at the outset tend to eventually get divorced, especially once their children are grown. Does that seem obviously true? Wait, that’s not right. That’s not what the research showed. The results really showed that those super-happy couples quickly come crashing down in the early years of marriage, and if they divorced, they tended to do so within the first six years, while those miserable couples were equally likely to divorce early or late. Oh wait, I got it wrong again. These are the real results: Couples who were especially unhappy early in their marriage were the most likely to get divorced early on (within the first six years), whereas when later divorces (seven or more years into the marriage) were examined, it was the most blissfully happy newlyweds that were most likely to get divorced.1 Sure enough, when I wrote about these findings on my blog, one commenter stated "So prevalent and obvious, it's hard to believe anyone wasted time and money on yet another study telling us what we already knew." Ouch, makes you glad you're not the researchers who spent 20 years collecting that data." (https://www.psychologytoday.com/us/blog/close-encounters/201...)

Highly relevant in the context of open source project infiltration.

Elaborate.