Ah... math and pornography are seldom as close as they are on Slashdot. Enter math lover Bennett Haselton. Bennett has a hypothesis: self professed math experts will be better at interpreting the applicability of a certain child pornography law to a recent case than self professed English experts. To test this hypothesis Mr. Haselton posts a survey presenting the facts of the case and the relevant line from the statute, and asks the respondents to rate their mathematical and English abilities. The results are predictable in that the respondents who seem to match Haselton's profile agree with him; the group that rated their math skills as "Excellent" was the only one to believe the law was not broken as written. Haselton demonstrates that these results are significant (in the statistical sense) and then the real fun begins.

The problem is that the data has a very annoying quirk: though people who rated themselves as "Excellent" in math were the most likely to think the law was not violated, there actually isn't a correlation with math skills and that opinion in the numbers. Below are the percentages of people in each math group who thought the law was broken:

Excellent: 44%
Very good: 74%
Good: 69%
Fair: 75%

At this point the "the better you are at math the more likely you are to assume the law wasn't broken" hypothesis must go out the window. We would expect the "very good" group's percentage to be somewhere in the mid 50s, but it's sitting at 74%. There's something odd going on with the "very good" group that can't be modeled with a simple correlation and you must either: a) decide that the data is inconclusive or b) rationalize the quirk away. Mr. Haselton is an industrious one, and he has already come so far, so you can guess which option he chooses.

To explain why the "very good" group is significantly more likely to assume the law was broken he has this to offer:

"I suspect that many people with self-reported "very good" math grades were probably just good students who studied hard and did the practice problems and got good grades in math, but without necessarily having the insight that makes someone an "excellent" math student"

Damn that "very good" group! They're just a bunch of liars! They're not really good at math, they just think they are. The obvious problem with this approach is that if you're not going to trust the "very good" group's answers, why would you trust the answers of any other group? But, having neatly brushed away that problem, Haselton moves on to the flip side of his argument: people who can write suck. Ok, that's not really his point, it's more like this:

"you can use excellent prose to defend an illogical idea, or you can use poorly crafted prose to defend a good idea, and so if you care about the quality of an idea and its impact on the real world, you have to look at the substance of an argument, not the style"

That's all fair and good, but completely irrelevant. His survey demonstrates that those who rate themselves as excellent in English interpret the given law differently than those who rate themselves as excellent in math. That's it. Everything else is Haselton using "excellent prose to defend" his opinion.

The problem is that Haselton thinks he knows the answer. He backs up his position (and at this point you should really read his essay if you wish to continue following along) with this from "Mark Rasch... the former head of the Department of Justice Computer Crimes Unit":

"The government is trying to apply this theory to real sex but simulated minors... Under federal law, you typically draw the line at the use and posing of real kids."

But Rasch is contextualizing the case significantly more than Haselton did for his respondents. The survey makes no mention of the courts' prior interpretation of the statute, it doesn't even link to the entirety of the law, just a one line snippet. That snippet mentions nothing of "real kids", just "material that includes a minor." I'm not a lawyer, but I imagine that without further context that one line could be interpreted in any number of ways. Good thing Haselton didn't post the 2nd Amendment.

I find it sweetly ironic that the guy who goes out of his way to complain about people with no logic but a trove of words ends up holding a bag of numbers with no logic in sight.

Finally, I'd like to point out that using self-assessed qualifications in surveys like this is inherently dangerous because of the Dunning-Kruger effect. For example, note that no one labeled themselves as poor in either English or math. Haselton suggests providing a math problem would improve accuracy, but I think it'd be easier to just ask the respondents for their math and verbal SAT scores.