Fresh from talking sense about the likely implications of a male birth control pill, qwerty decides to play logician and the results aren’t pretty:
Vox’s data isn’t quite this bad, but is very sloppy and is based on very poor logic. As a result it’s probably off by a factor of 5 or more. Hard to say. Being off by a factor of 2 is pretty lame.
First, he takes a number of murderers (about 3400) in a certain age group. Then, since he personally has only heard of 2 murderers who were homeschooled, he assumes all the rest of the 3400 must have had a public education.
The worse thing about Vox’s logic is in comparing a tiny number to a huge number, without consideration of it’s accuracy, then projecting this sketchy data over 3 years, increasing his error by threefold.
He concludes that a homeschooled kid is only 1/114th as likely to kill as a kid from the public school system.
It would only take the discovery of 2 more killers on the homeschooled side (be it crime of passion, drug crazed homicide, road rage, or whatever) to change the result by a factor of 2.
Very sloppy conclusion, all in all.
This is the sort of pedantic observation that is particularly embarassing, ignoring as it does the forest for the trees. For qwerty to be correct, one has to believe the following things:
1. In a nation full of compulsory public schooling, a significant percentage of children have never attended public school or been homeschooled.
2. In a nation where 45 murders are committed every day, primacy of publicity would be given to one over a period of ten days without there being man-bites-dog component given the absence of a celebrity aspect. qwerty also ignores that there are many homeschooled individuals over the age of 22; I left them out in order to simplify the analysis but any +22 homeschooled killers would be fair game for the media as well.
3. Given a media which consistently seeks to negatively publicize that which it opposes, and given that this media is overtly anti homeschooling, the media would neglect to report widely on a homeschooled killer.
4. There have been at least 8 additional homeschooled murderes under the age of 22 in the last three years. Given the publicity that accompanied the recent case, what the chances that we’ve simply missed the coverage of them.
Most importantly, qwerty fails to note that his estimate still leaves 218 of the 228 homeschooled killers indicated by the statistics missing. Thus, even if we accept his revision, this would mean that instead of being 114 times more likely to commit murder than a homeschooler, a public schooled individual would be only 25 more times likely to do so. In other words, my point about there being a significant difference linked to schooling still stands.
The true deleterious effect of public schools may in fact be even worse – my 1/114 estimate may be too generous in favor of the public schools – because ten percent of the 45 million students in America attend private schools. This is why I originally said 40 daily murders, not 45, but I do not know at the moment if the privately schooled are doing their fair share of the killing or not.
To be strictly fair, given the gaps in the data, we should collate the publicly schooled and the privately schooled as one collectively schooled group in asserting that murderous 1/114 ratio, although I expect that given the fact that most killers are of lower IQ and lower income status and the dearth of media stories devoted to “prep-school killers” I suspect that my original estimate will ultimately prove to be too low, rather than too high. In any case, any possible errors here lie in the admittedly incomplete data, not in the logical assessment of it.