Bryan Caplan  

From Cheating to Signaling

The Political is Personal... Outside Offers, Matching, and ...
Alex Tabarrok cleverly notes that cheating on exams would be pointless if the human capital model were the whole truth:
Cheating works best if the signaling model is true. If education were all about increasing productivity and if employers could measure productivity then cheating would be a waste of time.
If students perceive the situation correctly we also have an interesting hypothesis: students should cheat more in those courses that offer the least productivity gains.
Interesting.  But my actual classroom experience makes me skeptical.  As far as I can tell, cheating, like crime, does not pay.  The reason isn't that grades and credentials are unimportant.  They assuredly are.  The reason, rather, is that most cheaters are incompetent cheaters.  They tend to cheat off students who know no more than they do, and end up failing anyway. 

Update: Alex reminds me that he's made my point in the past.

COMMENTS (12 to date)
Mike K writes:

So, are you saying that there aren't enough students "good enough" to cheat off of?

John Thacker writes:

Are you arguing that cheating is hard enough that students who would do a good job mostly find it more rewarding to simply learn the material?

FredR writes:

"The reason, rather, is that most cheaters are incompetent cheaters."

How do you know? Could it be that you just don't know about the competent ones because they are competent?

It also assumes that one can't learn anything by cheating. Which probably isn't true.

Gene Hayward writes:

As a High School teacher, I tell my students before each test "If you are going to cheat sit next to someone smart, not your friend you are sitting next to now". The cheaters usually dont get it but the smart ones give an acknowledging smile.

drobviousso writes:

FredR - that's exactly correct. Cheaters you know about are, tautologically, bad cheaters.

Emily writes:

Yeah, but what if education is mostly about productivity but employers can't measure it very well, at least not when they're choosing among who to hire, and they're also firing-averse? In that scenario, I'd expect cheating. Education being about building human capital vs. signaling, and employers being able to measure productivity easily vs. not, are separable.

Steve Sailer writes:

"The reason, rather, is that most cheaters are incompetent cheaters."

At least, the ones you've caught ...

It's a big world out there, and there are a whole bunch of people, a lot of them East of Suez, who are currently devoting a lot of IQ points and energy to figuring out how to game American academic tests. At this point, my money is on the gamers, not on the American academics, who mostly grew up in a more naive time and culture.

blink writes:

As other commenters have mentioned, there is an obvious selection bias when considering cheaters who have been caught. Only a small proportion of cheaters need be successful in order for the top ranks to comprise mostly cheaters. This applies to academics, sports, politics, and many other arenas.

Also, if cheating were rarely successful, the righteous indignation we observe would be very difficult to explain.

Bryan Willman writes:

Is it that many cheaters are incompetent cheaters, or that only incompetent students attempt to cheat?

Ccheating is a more complicated topic than you might think.

In college, in classes I thought then (and now) to be irrelevent except they were required for a degree, I took lots of short cuts. None "cheating" in the formal sense. But if you read only 10% of the assigned reading and attend only 2/3 the classes, it's hard to claim you've applied yourself. And when you still get an A, it implies:
1. The class was a waste of time.
2. The competition was incompetent, period.

Cryptomys writes:

The cheater also has to assume the risk of getting caught and receiving an F.

Plagiarism is much easier these days than it was when I was in college. Detecting plagiarism, through commercial services like TurnItIn, has also gotten much easier. In courses that require term papers or a lot of writing, the instructor doesn't necessarily even need TurnItIn. Simply take a few random sentences out of the paper and google for them.

rpl writes:
...cheating on exams would be pointless if the human capital model were the whole truth.
This statement is absurd. If "signaling" includes the case where the human capital putatively imparted by education is real, but employers must rely on grades to measure it, then the "signaling model" is so broad as to be useless. Nobody disputes that employers rely heavily on grades and test scores to measure the capability of prospective employees. The question of interest is whether the thing they believe they're measuring actually exists.

Now, note that the existence of cheating doesn't tell us anything at all about the question of interest. Whether or not education actually imparts any human capital, it's equally useful to spoof the scores that (people believe to) measure it.

Comments for this entry have been closed
Return to top