Short Form: People Cheat

People cheat, across many domains, surprisingly often to me.

Academics

At Middlebury, 35% of 377 students self-reported an honor code violation in 2014, 97% of which were not punished.
Some other random links suggested similar numbers (and up to twice as high for colleges without honor codes).
It’s interesting how some professors think cheating is very rare. I wonder how much of this is a selection effect (good students are less likely to cheat when undergrads), vs. incentivizes for colleges to not research/publish stats on this (leading to people falsely assuming it’s rare), vs. norms changing over time for cheating to be more acceptable (because more people are doing it / because school seems more like busy work that no one actually cares about), vs. something else.

I highly recommend the movie Icarus. [Spoiler alert: Russia doped (took performance-enhancing drugs) for the 2014 olympics across sports, and it appears to be surprisingly easy to get away with doping.]
Lance Armstrong got away with doping for a long time, and it sounds like cyclicsts were very into it as a collective.

MLB went from 0 cases –> many when they actually started a good testing regime.
Screwball is an entertaining documentary on a quasi-doctor who helped MLB players get steroids (and also high-schoolers / anyone that asked for them…).
A-rod claimed he never doped, got caught doping, then claimed he was clean, then got caught again, then continued to claim he didn’t take drugs.

I’d bet a good amount of money that male sports with <1-5% of players getting caught have similar base rates compared to other sports, and just have bad testing regimes.
For example, Soccer.
Icarus claimed the Russian world cup soccer team was doping. I don’t think FIFA followed up on that.
Documents suggest Real Madrid and Barcelona players (two of the most famous clubs) were doping back in 2006, I think it was basically never followed up on.
When I googled NBA doping a while back, I found similar things as with the soccer searches (some players suggesting it was common, but no strong testing regime / not many players getting in trouble for it).

I’m guessing the base-rate for professional male sports is something like 20-70% (median 40%), and maybe 2-5x lower for female sports (median 15%).
If you’re a male sport and the rate of players you catch is below 1-5%, and you don’t have intense testing policies / have never had a public scandal, it seems likely you just aren’t doing good testing and are in the 20-70% regime.

Cheating on spouses: a random link from a google search suggested 20% of married men admitted cheating and 13% of women.
Evading taxes.
Other crimes.

Is there a good name for the effect “sometimes when test numbers are very low, the true number is way higher than you think, because that suggests your tests are very bad?” In addition to the examples above, figuring out accurate estimates for how many people were infected with covid at the start of the pandemic was made tricky by a similar effect, where one had to guess at how many people were being tested, and many people incorrectly focused just on the raw testing numbers.
It’s interesting to think about the base rate for willingness to do unethical things, and also how different things affect willingness to cheat (benefit, cost via chance of getting caught).
Someone’s maybe done this already, but it seems like it’d be pretty easy to make a quantitative model of how likely there is to be a tragedy-of-the-commons type failure mode given x actors, y % benefit, and z % chance of getting caught.
Lots of problems in capitalism look like “well, there are a bunch of actors, and definitely a few of them are going to cheat in that way that is hard to catch, so I guess all the competitive ones have to now.”
It’s always mind-boggling to me when people like A-rod or Trump seem to not care at all about the truth.
Oftentimes it seems like the only way to get a regulating agency to actually do good testing is to create a public scandal, by e.g. publishing the details of just how many people brazenly cheat on a thing.

June 24, 2021 ·