Tuesday, March 28, 2006

Richards Politics and Personality Meme

As some of you may remember, way back when, Richard of Philosophy, etcetera started a meme in which bloggers took both the IPIP-NEO personality test and the Political Compass quiz, and reported their results (if you want to participate, take the tests now, before reading on). Richard has posted the results, but I thought I'd say a little bit more about them. Naturally, none of these results should be treated as scientific, because, well, they're not. In addition to the fact that neither the online IPIP-NEO or the Political Compass are great tests, the sample was pretty skewed towards the Left and (Social) Libertarian ends of the Political Compass scales, as indicated in the two histograms below.
The scales both go from -10 to +10, with -10 representing the most extreme Left/Libertarian score, and +10 the most extreme Right/Authoritarian score. As you can see, there weren't many bloggers on the Authoritarian end of the Libertarian-Authoritarian scale (16 out of 89, or 18%), and there were relatively few on the Right end of the Left-Right scale (23 out of 89, or 26%). Furthermore, on both scales, those on the Left and Libertarian ends tended to be more extreme than those on the Right and Authoritarian ends. The mean for those on the Left end was -5.6 (median = -5.88), while the mean for those on the left was 4.4 (median = 4.0). The difference is even bigger for the Libertarian-Authoritarian scale. The mean for those on the Libertarian end was -5.3 (median = -5.3), while the mean for those on the Authoritarian end was 2.3 (median = 1.7). In both cases, the differences in extremity are statistically significant. The one good thing the political compass results have going for them is that they seem fairly consistent with the bloggers' self-identification as either liberal or conservative.

The point of all that is simply to make it clear that, if you wanted to compare the relationships between the scores on the compass to the scores on the personality test, this isn't the sample to do it, because it's so heavily skewed. The fact that 67% of the bloggers who gave their results were male doesn't help either. But because this is a blog, and not a scientific journal, I'm going to report the results anyway. Just remember, "bullshit" would be a good word to describe these results. But it's interesting bullshit.

Since the interesting correlations were between the personality variables and the two political compass scores, I'll focus on those. I will quickly note, however, that there were no statistically significant correlations between the personality variables and either age or sex, though the correlation between age and artistic interest (.21) was pretty damn close. It indicates that as people got older, their artistic interest grew. Don't ask me what that might mean (young people and their terrible music! bah!). Also, the correlation between the Left-Right and Libertarian-Authoritarian scales was high (.75) and statistically significant.

For simplicity's sake, I'll report all statistically significant correlations between the compass scores and personality variables as being positive, with those that were negative (i.e., personality scores went up as individuals became more Left or Libertarian) being reported as positive correlations with Left/Libertarian. On the personality test, there were 5 personality categories (Extraversion, Agreeableness, Conscientiousness, Neuroticism, and Openness to Experience), the Big Five. Within each category there were 6 additional variables, each with its own score. First the correlations for the categories. Conscientiousness was positively correlated Right (.23) and Authoritarianism (.22), Agreeableness was positively correlated with Left (.22), and Openness to Experience was positively correlated with both Left (.63) and Libertarian (.66). For the variables within the categories, only one was significantly correlated with either Right or Libertarian. Dutifulness, which is in the Conscientiousness category, was positively correlated with Authoritarian (.23). There were a bunch of significant correlations with Left and Libertarian among the variables within the categories, so instead of writing them all out, I'll just let you read this table:

As you can see, all six of the variables in the Openness to Experience category were positively correlated with both Left and Libertarian, with some of the correlations in the moderate to high range. I suppose it's a good thing that the personality variable Liberal correlated with Left, and that the correlation between Liberal and Libertarian was about the same as that between Left and Libertarian.

Before signing off, I should make a few comments (because I like to hear the sound of my own typing). I think it's interesting that these results are somewhat consistent with the two highly controversial studies I mentioned in a recent post. The correlations between the Left and the category Openness to Experience, as well as Imagination, Artistic Interests, Adventurous, and Intellectual are consistent with the Jost et al. results. I suppose the correlation between Left and Sympathy could be consistent with those results as well, as could the correlation between Authoritarian and Dutifulness. I suppose the fact that this sample was hardly representative of the population as a whole is another point of consistency. Also, it might be interesting to note that the correlation between depression and Left and Depression is consistent with the literature showing that conservatives/Republicans tend to rate themselves as being happier. I don't know what the hell the correlation between Emotionality and Left means. Maybe it's why they call 'em "bleeding heart liberals."

So there you have it, the results of Richard's meme. They're kind of interesting, somewhat consistent with what I think many of us would expect, and entirely meaningless. I don't know if you could ever get a truly representative sample through blogs (though it can't be any less representative that a sample of college undergrads), but it would be nice to have a more even distribution on both the Left-Right and Libertarian-Authoritarian scores. I suppose it's skewed because Richard and the other blogs that picked it up are mostly read by us lefties, though it's mighty tempting to say that the skew is because Richard's blog is a philosophy blog, and Intellectual is positively correlated with Left and Libertarian (that's really meant as a joke, folks).

8 comments:

Brandon said...

Well, as you say, the results are interesting, even if it's only interesting as a sign of what's going on among bloggers who do memes like this and are related by their blog reading to Richard's blog. As one might say, it's an odd population, but it's still a population....

Clark Goble said...

It's unfortunate that more conservatives and moderates didn't contribute results. It seems, as you said, that there were so few that there wasn't enough data to say much about conservatives.

Richard Y Chappell said...

I vaguely recall reading that the personality test scaled results within age and gender groups. So I'm not sure whether it would be possible to get results from that test showing that "as people got older, their artistic interest grew." (Though that claim does sound independently plausible to my ear.) The results might instead indicate that our older bloggers were likelier than younger bloggers to be more artistic than their (respective) peers? But perhaps I've misunderstood something here...

Chris said...

Richard, I think you're right. If I remember correctly, what they do is normalize your score based on an age range. So you're right, the best way to interpret it is that, relative to members of their age group, the older bloggers (and none of them were "old," since 51 was the highest age) had greater artistic interest than the younger bloggers.

Bora Zivkovic said...

Long, long time ago I took a jab at the Political Compass, pretty much concluding that the Left/Right and Libertarian/Authoritarian axes are not independent of each other, so I am not surprised that there is high correlation between Left and Libertarian, etc. So, the whole coordinate system graphing of the Political Compass is bogus - is the questionnaire bogus, too. Is comparing something else to it, then also bogus?

Clark Goble said...

I tend to find the questionnaire for the compass fairly problematic. I admit.

Chris said...

The political compass does suck.

Genius said...

IPIP NEO looked like it would take too long and I'm not entirely convinced it is a good measure anyway.
Although I might have done it before once...