Political Orientation — Why the gap?

(For a related discussion, see Meng Hu’s on libertarianism and AFQT scores.)

In the Add health survey, when restricting consideration to non Hispanic Whites, liberals were significantly more verbally intelligent than self identified conservatives as self defined in both waves 3 and 4. The difference was in excess of one half of a population SD. The results were mixed when looking at wave 4 digit span backwards scores, however. Interestingly, liberals were more verbally intelligent in adolescence (age 12 to 18) presumably before political orientation (PO) coalesces. This suggests that liberalism attracts more VIQ individuals rather than that liberally oriented individuals become more VIQ via education, that is, liberal ID is consequent to VIQ, not vice versa.

This is shown below:


There was a good deal of longitudinal variation in self-ID PO. This had some effect on scores. Selecting only individuals who consistently IDed as liberal or conservative in both waves gave the following results:


I was forced to condense liberals and conservatives into one category each because otherwise the samples were unreliably small. While there was a liberal-conservative gap, there was no democrat-republican gap. I would surmise that this is the result of republicans running left. That is, republicans close the VIQ — and with it, the influence — gap by appropriating liberal-left ideas and attracting VIQ individuals. Below shows the point-biserial correlations between dichotomously coded groups and wave 1, wave 3, and wave 4 cognitive ability.


The VIQ gap clearly explains the left-ward drift of the nation. Not only are liberals more VIQ than conservatives, but liberal liberals and liberal conservatives are more VIQ than, respectively, conservative liberals and conservative conservatives. In the present mediacracy, the former can dominate the latter.

But why the gap? To explore this question I looked at numerous variables. When placed in a regression equation, the most explanatory of them only jointly accounted for 1/3 of the difference. Below I will review variables that are commonly noted.

First I looked at PO and novelty seeking. A number of variables loaded on a common factor which could be called a novelty or change factor. The factor strongly correlated with PO but was uncorrelated with cognitive ability. Embracing the mantra of change will help neither conservatives nor republicans close the gap.

Correlation matrix for novelty indexes and PO.


Novelty factor, PO, cognitive ability.


Next, I looked at the relation between PO and a preference for abstract ideas and self reported creativity. Both variables correlated with PO and cognitive ability. This is shown below:

Abstract-Imagination, PO, cognitive ability.


More importantly, a preference for abstractness non trivially mediated the PO-VIQ relation. If conservatives want to increase their mean VIQ, they should put aside the Buckley-esque commonsensicalism and work on creating a more theoretical conservatism.

VIQ-PO taking into account Abstract-Imagination


Next, I created Wave 3 and wave 4 “feeling factors” out of a number of related variables. The correlation matrices of the variables used (before recoding them to show the same signs) is shown below:

Correlation matrix for feeling indexes


I created two factors as the wave 3 and the wave 4 feeling factors differently related to PO. The relations between PO, Feeling 3, Feeling 4, and cognitive ability are shown below:

Feeling factors, PO, cognitive ability


While wave 4 feeling factor significantly correlated with cognitive ability, it didn’t significantly correlate with PO. While Wave 3 feeling factor did correlate with PO, it showed a relatively weak association with VIQ. My interpretation would be that a more mature other-understanding contra misanthropy is unassociated with PO, while a more naive, immature let’s-change-the-world age 20-26 understanding is associated with PO. But this understanding only weakly correlates with VIQ and, moreover, does not appreciably mediate the PO-VIQ association (analysis not shown). The upshot is that more “compassionate conservatism” will not win many brains.

Next, I created a religious-transcendent factor. The correlation matrix for the variables used is shown below:

Correlation matrix for religious indexes


This factor strongly correlated with PO and with party identification. But contra to expectation based on e.g., Lynn et al. there was no correlation between it and cognitive ability. I presume that the difference in findings had has to do with the definition and operationalization of religiousness e.g., I didn’t contrast religious with atheist, but rather factor analyzed a bag of religion/spirituality related variables.

Religious factor, PO, cognitive ability


Anyways, if the above is correct, it would suggest that both conservatives and republicans don’t need to lose religion to get (V)IQ. (To note, I’m sympathetic with many religious folk because I accept their basic mental frame; I just reject the existence, not the idea, of a transcendent — and so class myself as a nihilist of sorts).

Of the variables discussed, then, the one with the highest R-squared was openness to abstraction. This then might be an area to focus on.

This entry was posted in Uncategorized. Bookmark the permalink.

2 Responses to Political Orientation — Why the gap?

  1. JayMan says:

    Great work!

    I don’t believe these things are amenable to change. Parties change, positions change, people, not so much. You can change the stance of the parties and change the people in the mix, but you’re really just shuffling the chairs on the deck.

    I have wondered if the liberal edge in IQ is in g or just in verbal ability. I wonder if math ability would balance things out.

    As for the personality measures, seems you captured the trait “openness to experience”, which is the hallmark of liberalism.

    I’ll have more to say when I’m not on my phone. 🙂

    • Chuck says:

      At best, it’s work in progress. If you want, I could send the data file and a bootleg copy of SPSS so that you could take a look for yourself. …It’s funny that you commented since I started by looking at the rates of heart disease by European ethnicity — to check if you were onto anything. But then I lost interest and decided to take a look at the ideological stuff…. As for the musical chair analogy, agreed. But parties are coalitions of people. And ideologies are coalitions of ideas. So you can arrange the chairs partially to bring together more or less successful groups…..The nature of the difference is definitely worth investigating. I checked self reported english, math, science, and history grades; the general grade factor correlated with verbal IQ (at about 0.3) but there were no general grade factor difference. Ditto with self perceived intelligence and self-reported ever-held-back-a-grade and so on. On the other hand, I constructed a general longitudinal factor and that showed a substantial difference. So the difference taps into a common longitudinal component which is what would be the case were the differences in g…Unfortunately, the NLSY doesn’t have a “conservative/liberal” variable. I suppose that one could construct an index of social conservatism, though, by using responses to personality questions (e.g., How important is tradition? Are traditional gender roles important?) and then look at ASVAB subtest scores. I don’t know. Maybe I’ll try it…I would probably want to redo this analysis first, though. There were a lot of other variables that I looked at but didn’t note.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s