Please note that some of the graphs there start with 50% as original y-axis. It creates an appearance of males drinking twice as much as females, for example.
Nah, they just score higher on a vocabulary test because they're articulating extra crisply.<p>Working in a public library is a good way to dispel illusions about drinking being classy, by the way. We could always tell when some of the veterans got their monthly checks...<p>Also: Bike thieves <i>do</i> suck, and these graphs would send Tufte on a crying jag.
Ironically, I learned a new word: 'confound' (<a href="http://en.wikipedia.org/wiki/Confounding" rel="nofollow">http://en.wikipedia.org/wiki/Confounding</a> ). And then I realized that the last chart has confounds. Intelligence and vocabulary are correlated, but worth differentiating in this case.<p>A higher vocabulary is pretty strongly correlated with being well-read. I think we can trust that. And being well-read implies being exposed to different things (though it is by no means the only way to be exposed to different things -- i.e., those who are well-read [subset operator] those who are exposed to a wide variety of things.<p>But those who are exposed to a wide variety of things are highly correlated with those who wish to take risks <i>to be</i> exposed to new things. And the latter group seems pretty well correlated with those who drink.<p>Note, intelligence only 'lurks' near vocabulary.
From another article: "In the 1998 to 2004 data, each point higher on the Wordsum test causes a $1,200 decrease in income." [<a href="http://www.halfsigma.com/2006/07/higher_intellig.html" rel="nofollow">http://www.halfsigma.com/2006/07/higher_intellig.html</a>]<p>So where does "smartness" fit in? I hate to imply causation, but maybe:<p>"Wordsat smartness" -> annoying person who uses big words other people don't understand all the time -> social isolation -> less money -> alcohol!<p>Whew!
I don't know about the word test scores but the rest of the demographic info doesn't look right.
For instance, the second chart. The data suggests that statistically, the odds of a girl drinking are higher than an african-american drinking.<p>By the way, very very misleading graphs. I was totally shocked by going through them. I had to read the comments here to realize what was being shown.
Apparently, here is the vocabulary test used at the GSS and referenced in the last graph:<p><a href="http://inductivist.blogspot.com/2010/04/gss-vocabulary-test.html" rel="nofollow">http://inductivist.blogspot.com/2010/04/gss-vocabulary-test....</a>
I was curious where the data came from, since the article didn't really explain it. Here's the details:<p><a href="http://en.wikipedia.org/wiki/General_Social_Survey" rel="nofollow">http://en.wikipedia.org/wiki/General_Social_Survey</a>
I thought this was a linkbait title, but it actually is the title of the original article.<p>One of many thoughts appropriate to this dubious claim: often, questions like, "Do you smoke pot?" or "Do you drink?" have more to do with identity than with the actual frequency of consumption.<p>I'd bet that people "of class"-- well-to-do, educated, liberal-- are more likely to say "yes" to the question, "Do you smoke pot?" However, a large number of such people are those who haven't smoked for 20 years-- those who wouldn't be averse to smoking pot if it came their way, but haven't looked for it or been near it in a long time. (I have no problem with people who use marijuana, but the fact is that to get regular access often requires association with "unclassy" characters, because most drug dealers are creepy and the process of asking for access is degrading.)<p>If you look at who is actually smoking pot, and weight by frequency of consumption, I'd guess that the correlation goes away. Same with alcohol.
People always shoot the messenger for pointing this out, so I'll expect about a million downvotes for it, but in case you case didn't know it already GNXP is the cryptoracist camp along with Charles Murray and Steve Sailer. Seriously, check out some of the comments.<p>Okay, now you can downvote me.
> I knew that blacks were more likely to be teetotalers, and expected that women would be as well.<p>Is there anything scientific about this at all? The whole article sounds fishy.