Monday, May 09, 2011

IQ Wars: McCain voters win

The 2010 GSS data have been released. As Inductivist noted, this is a fun time for quant bloggers. There will be a lot of sifting through special modules for the 2010 battery of questions, but of broader interest are converted IQ averages of voters in the 2008 election. With the post-racial man who putatively put together an autobiographical masterpiece on one side and milfy trailer trash on the other, the last Presidential election should've added an exclamation mark to the trend towards higher IQ Democratic voters and lower IQ Republican ones, yes?

No. Converting the white mean on the wordsum vocabulary test to an IQ of 100 with a standard deviation of 15, the following table shows the average IQ of voters categorized by who they voted for, and then further broken down into groupings based on sex, race, party affiliation, and age (n = 921):

Voted for...
Obama
McCain
McCain's edge
All
99.9
102.5
2.5
Sex



Men
100.4
102.2
1.9
Women
99.6
102.7
3.0
Race



White
103.2
102.9
(0.3)
Black
92.9
n/a
n/a
Hispanic
96.6
n/a
n/a
Asian
107.1
n/a
n/a
Party



Republican
101.3
102.7
1.4
Democrat
99.9
96.0
(3.9)
Independent
98.5
104.3
5.8
Age



18-24
99.4
100.9
1.6
25-34
98.6
101.5
3.0
35-49
100.4
103.7
3.3
50-64
101.0
103.3
2.3
65+
99.2
100.5
1.3

The sample sizes for non-whites who voted for McCain were simply too small to even be considered mildly suggestive and are consequently not included. Of the 351 respondents who report having voted for him, a whopping five are Hispanic, three are black, and two are Asian.

It's not especially surprising that among all voters, McCain supporters have an IQ edge, since despite all the Senator has done to dilute their population share, whites are the only ones who vote for the GOP. More remarkable is the IQ parity among white Obama and McCain backers. That Obama won the $200,000+ income category has been pointed out by Half Sigma as evidence that smart whites are increasingly finding the Republican party repulsive, but Obama's edge among this affluent group was only 52%-46%, nearly identical to the nationwide popular vote (52.9%-45.7%). The partisan divide among wealthy whites mirrors that of the rest of the country.

Traitorous Republicans and Democrats are both revealed to be less intelligent than their truer compatriots, hardly constituting a novel observation. The spread among independents does come as a bit of a surprise. I'm not sure how to account for it. Maybe the "independent conservative" members of the Savage Nation explain the variance.

McCain's IQ advantage is maintained across the generational spectrum. The Obama youth are more vociferous than they are intelligent. Funny how the McCain advantage is slimmer on each end--among kids full of quixotic notions and the elderly who are losing the ability to digest complicated and detailed information, turning instead to presumed probity and aesthetics to guide them--and thicker among the stone cold middle-aged realists!

Notice how even though the white average is set at 100 and NAMs are included in these figures, the racially inclusive averages are at or above 98, which is generally considered to be the nation's mean IQ. Eligible voter participation rates and IQ correlate at .65 (p=0) at the state level, and both educational and income profiles show the average voter to be considerably more educated and wealthy than the average residents of the most educated and wealthiest states.

Parenthetically, the converted IQ average for third party voters (n = 30) is 105.6. That's fairly predictable, since merely seeking these candidates out requires some degree of intellectual curiosity.

GSS variables used: PRES08, RACECEN(1)(2)(4-10)(15-16), SEX, AGE, PARTYID(0-2)(3)(4-6), WORDSUM, BORN(1)

14 comments:

bgc said...

I would expect the standard deviation to be greater for Obama than McCain voters - ?

Off topic - is there a handy gadget for converting SAT scores to IQ, so I could look at US colleges interquartile range of IQ for data sets like this one:

http://www.satscores.us/

TwoYaks said...

I really think WORDSUM isn't a good statistic to use for IQ in political matters, because of how dimorphic politics appears to be. And I don't need remind you that WORDSUM shows sexual differences. I'll readily concede the trend is there, though - I've no reason to think the bias would change over time, for purposes of comparing it to past elections.

Audacious Epigone said...

BGC,

I would've expected as much, too, but the SD difference is marginal (it is slightly wider for McCain voters).

I've used this one before (it does not have score conversions for the newer 2400 point version of the SAT).

TwoYaks,

Right, when it's all-in, the differences are going to tend to be moderate if they're detectable at all, and this is no exception. The sexual difference is interesting in that women do better on wordsum than men do, yet men were more likely to vote for McCain than women were, and the McCain voters still have the edge.

read it said...

Another IQ conversion chart for SAT, ACT, Miller Analogies, GRE, etc.

http://www.assessmentpsychology.com/iq.htm

bgc said...

Thanks!

bgc said...

Presumably this implies for the intellectual class that the overwhelmingly Democrat 'representation' in the mass media, education, law, NGOs and public administration is balanced by equally overwhelming Republicanism in other areas where intellectuals work.

Audacious Epigone said...

BGC,

A lot of those people (high IQ rightists) are in executive and professional positions in the private sector. There is little reason to think that the extreme right end of the bell curve is predominantly leftist, yet academia and the major media are. There must be expositions as to why that is out there, though I'm not sure I could articulate an explanation very well.

bgc said...

My preffered explanation is that a tendency towards Leftism relates to an abstracting intellect; an inbuilt tendency to privilege abstraction over experience.

This compulsive abstraction is mostly found among people of high IQ, but quite a lot of high IQ are not natural abstracters - are more practical, and experientially orientated. Have more 'common sense'.

This is what I once called Clever Sillies.

I am one of them - the above analysis is a classic example of Clever Silly abstraction!

(Which does not mean it is *necessarily* wrong.)

silly girl said...

"A lot of those people (high IQ rightists) are in executive and professional positions in the private sector"

Well, heck yeah. Investors want results, not excuses or abstractions. It is about the bottom line.

CSI w/o Dead Bodies said...

've seen this flawed reasoning before. Epigone is a fan of the book the Bell Curve which revived old eugenic theories about IQ and race. There are many alternative explanations for racial differences on standardized intelligence tests. These explanations include racial differences in educational experiences and cultural biases in test questions.

The whites who voted for Obama were virtually equal in average IQ (0.3 points higher) but there weren't enough minority McCain voters to do a comparison in the sample.

Anyway what is the meaning of the overall difference of 2.5 points? We attach so much importance to minuscule differences in things like batting averages but how often do we ask if it really means anything? The IQ scale is set up so that 100 is average and 15 is one standard deviation.

http://csiwodeadbodies.blogspot.com/

Audacious Epigone said...

CSI,

Have you read TBC? Since you seem to be under the impression that it is about race, I'll assume you haven't.

As far as "flawed reasoning" is concerned, I'm not sure what you're referring to. Like most posts here, it is empirically-based and designed to be primarily informational.

Stopped Clock said...

http://en.wikipedia.org/wiki/Differences_between_conservative_and_liberal_brain

According to the ASA, IQ data from the "Add Health" survey averaged 106 for adolescents identifying as "very liberal", versus 95 for those calling themselves "very conservative". An unrelated study in 2009 found that among students applying to U.S. universities, conservatism correlated negatively with SAT, Vocabulary, and Analogy test scores though there was a greater correlation with economic differences.

Anonymous said...

with the exception of the two outliers of Blacks and Asians, there is really no significant difference in these scores. What it does prove is that most people are average. It is likely that as your sample grew, all the scores would come closer to 100.

Anonymous said...

The correlation between adult IQ and WORDSUM = 0.71.

from Razib Khan http://blogs.discovermagazine.com/gnxp/2010/05/wordsum-iq/

You cannot take a .71 correlation and then use that to make detailed analysis of a few differential point spreads