Tuesday, March 28, 2006

Personality and Politics Results

The incomparable Chris of Mixing Memory kindly sent me his statistical analysis of 89 responses to my personality and politics survey. He found several interesting statistically significant correlations. But first let me describe the demographics.

Ages were fairly evenly distributed between 20 - 50 years. 67% of respondents were male. Despite explicit efforts to enlist conservatives, the results still ended up being dominated by economically left and socially libertarian bloggers [Update: Chris has posted a couple of graphs which make this very clear; his other remarks there are worth reading too]. From the 84 that indicated a location, Chris garnered the following information:
Six countries are represented, including New Zealand (8), Canada (3), Australia (2), the United Kingdom (1), France (1), and the United States (68). Within the United States, 31 states and the District of Columbia are represented. The states with the most respondents were Texas (9, including 4 from Austin), California (6), New York (6), and Pennsylvania (5).

Now for the actual results...

There was a very strong (0.75) correlation between the left/right and libertarian/authoritarian Political Compass axes. Most of the personality traits thus had significant correlations with either both or neither of the political axes. There were just three exceptions to this rule: agreeableness was weakly correlated with leftism, but fell just short of a significant correlation with libertarianism (if I recall my stats correctly, this means that there is greater than 5% probability that results like ours would have been obtained from chance fluctuations in the sample if there had been no genuine correlation in the broader population. [N.B. this is distinct from the actual chance of error.] For "significant" results this probability is less than 5%. Some of our stronger results were significant to p<.001, i.e. 0.1%). Trust, on the other hand, was significantly (albeit weakly) correlated with libertarianism but not leftism. And dutifulness was significantly correlated with authoritarianism, but correlated too weakly to qualify as significant in relation to rightism.

The other personality traits were either significantly correlated with both leftism and libertarianism, or both rightism and authoritarianism, or else were not significantly correlated with any of the four. So I will simply mention whichever member of the above pairs obtained the strongest correlation in each case.

There was a very strong (0.75) correlation between sympathy and leftism (hence, also, a just slightly weaker correlation between sympathy and libertarianism; I will leave out these parenthetical elaborations for subsequent results).

The next strongest correlations -- all of which held to p<0.001 -- were between liberalism and leftism (0.75); openness to experience and libertarianism (0.66); artistic interests and libertarianism (0.55); leftism and imagination (0.45); and leftism and emotionality (0.38).

Weaker but still significant correlations held between adventurousness and libertarianism (0.35); depression and leftism (0.33); intellect and libertarianism (0.27); and conscientiousness and rightism (0.23).

Other correlations which fell just short of official statistical significance - but I think still warrant a mention - were between right/authoritarianism and: friendliness, activity level, orderliness, self-discipline, cautiousness, anger, and immoderation; and between left/libertarianism and: excitement-seeking, altruism, neuroticism, and anxiety.

Curiously, there were also notable (but not quite "significant") correlations between older age and: artistic interests, also excitement-seeking; and between youth and: cautiousness, dutifulness, conscientousness, altruism and cooperation.

A note on interpretation:

Bear in mind the skew towards left/libertarianism. Correlations with 'rightism' here might be more about moderate leftists (who are "right" in comparison to those further left, of course) than about actual right-wingers. Those who understand statistics better than I do might be able to say more in comments. I just thought I should warn against drawing overly bold or partisan conclusions here. (E.g. if those on the far left tend to be less angry than moderate leftists, that might be enough to create a correlation between "rightism" and anger. It doesn't necessarily say anything about actual right-wingers.)


If anyone wants the results spreadsheet, just send me an email and I'll forward it along (if Chris doesn't mind). For the rest of you: have fun speculating about the deeper meaning behind these results...

(Do more sympathetic and imaginative people become left-wing because they better understand the plight of those less fortunate? Are they depressed because they just care too much? Or are they just too unconscientous to accept right-wing ideals of personal responsibility? Have at it!)

Finally, I should note that the original survey is still "open" (and will be as long as the internet is!), so if more conservatives feel like taking it, we might be able to get even more interesting and accurate results in future. Authoritarians: think of it as your scientific duty ;-)


Post a Comment

Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)