At our recent meeting of social psychologists, I had a few conversations about a particular facet of our data, the fact that liberals in our dataset score higher on measures of neuroticism than conservativism. The effect in our data is small, but not insignificant (d=.24 of the standard deviation). This surprised some people in that there is a fair amount of research about conservatives being fearful that people are aware of, even as there is some contradictory evidence. A recent meta-analysis + study yielded mixed results, with some research and samples showing liberals as being more neurotic (including the lone non-student sample, though with a very small effect size), and some research showing conservatives as being more neurotic. One conclusion might be that this is all statistical noise. An alternative possibility is that it depends on the types of questions being asked or that it depends on the group being sampled. I thought I’d explore this in our yourmorals data.
First, you can see that within the yourmorals dataset, liberals appear more neurotic than conservatives regardless of the question that is asked.
This even extends to asking about symptoms in the recent past. The questions here are how often the participant has experienced “Being suddenly scared for no reason”, “Spells of terror or panic”, or “Feeling fearful” in the past 7 days, though the effect is tiny.
It appears that the effect is robust across questions. Our sample is not representative of the broader US, but in this instance, this may be instructive. Liberals may be more neurotic than conservatives within certain groups. Our data is a large enough sample that it likely represents a sizable group of people, and it is possible that there is something peculiar to the kind of people who visit yourmorals that makes our liberals more neurotic than our conservatives. As a broader test of this idea, I thought I’d examine those participants who visit yourmorals from sites like the New York Times or Edge, versus those who find yourmorals.org via search engines (e.g. searching for ‘morality quiz’), with the idea that the NY Times and Edge readers are more like our core audience (people especially interested in social science).
Here is the graph by question for those who find us via search engines:
And here is the graph for those who find us via the New York Times and Edge.org.
It may be self-evident from the graphs, but put another way, the correlation between neuroticism and liberal-conservative identification (1-7) is -.03 (n=1634, p=.22) for those who find us via search engines, -.08 (n=7129, p<.001) in New York Times readers, and -.13 (n=2382, p<.001) for those who find us via Edge.org. Overall, the correlation is -.08 (n=35,793, p<.001).
To me, this supports the idea that there is something peculiar about the kind of liberal that reads the New York Times or visits Edge.org or a site like YourMorals.org. Perhaps the common thread here is the idea that these are people who are searching for answers in life. It somewhat converges with this paper by Napier & Jost, where they find that liberals are less happy than conservatives, a finding that replicates in our data and has been found by others, and they found that this relationship is explained by the liberal un-acceptance of inequality. It seems somewhat implausible that liberals walk around consciously thinking about inequality a lot. But perhaps the inability to accept inequality is part of a general questioning of the way things are and what the larger meaning of things is, which inevitably leads to anxiety about why things are not ‘better’. I cannot show that with data, but I can say that, as a liberal, this rings true for me. My search for meaning and desire to create change inevitably lead me to anxiety producing situations when I try to swim against a tide. And yet it’s a tradeoff I continue to be willing to make.
- Ravi Iyer
You completely missed the fact that rightwing knee-jerk reactionaries are neurotic in their fear of almost everything – hence the visceral reaction to issues rather than a thoughtful approach to understanding issues. Your research is so flawed.
Research would support your view, but there are many different kinds of neuroticism. Indeed, one of our future projects is to unpack these and I bet it is likely that our results will support your hypotheses, but that there are still some ways that liberals are neurotic too. I’m reminded by Tolstoy’s quote that happiness in families is alike, but families are unhappy in their own way. I think liberal and conservative neuroticism simply looks different, but we’ll see.
I wanted to first start off by saying I’m a libertarian that loves your research and gained a lot of personal insights from it myself.
Now for some commentary. It is possible that some studies are conflicting on conservatives and liberals due to small sample sizes. Take for example your link on conservatives being more fearful, a study that only looked at 46 people. Sometimes, with a sample that small, randomness will find a trend when there really isn’t one. Find another group of 46 people, and the results could be different. (no evidence of randomization either based on my skim of the article, which is disappointing if true)Larger sizes tend to eliminate this from happening. On conservatives being more fearful, perhaps a meta-analysis of randomized studies should be in order.
It is also possible, as you said, that your sample is not representative of the liberal population at large. It could be that liberals as a group attract a range of subgroups with varying personality traits. Listening to NPR for example, it seems that a lot of the hosts tend to be emotionally stable with a curious bent. (they seem like empathetic libertarians) However, we all know that artists and writers tend to lean-left as well, and the saying is that this subgroup is notoriously neurotic. (based on the current research on writers and my own personal experience meeting quite a few, I’d say the saying is at least partially true)
There is also the possibility that people’s self-perceived personality may not always be accurate. I find it interesting that libertarians are your most stable group yet report low life-satisfaction. What a contradiction!
Just my two cents anyway.
If you are going to censor my post, at least show the common courtesy to tell me why. That way, I don’t make an assumption that this is another example of censoring ideas that don’t fit into social psychology’s perfect round hole. I assumed that this was something we were working on fixing, but I guess I was wrong. You can moderate this post out too; I am past the point of expecting a response.
John, first let me apologize for leading you to think we were censoring your comment. The reality is that we get a lot of spam on the blog and it isn’t anyone’s full time job to approve comments. So it sometimes takes a day or three to approve posts. It’s an unfortunate reality in the internet age and I’ll try to do it quicker going forward so others don’t think we are censoring them. We publish _any_ comment that doesn’t seem like spam or abuse.
As to your comment, I don’t think sample size is a huge issue as that’s what statistics are for. If the differences are stark enough and the measurement is accurate enough, you don’t need a big sample. However, you are quite correct about generalizability and the kinds of people who read our research may be specific kinds of liberals. That is mitigated somewhat by the fact that our research often replicates in people who find our research via different means (e.g. search engines, the NY Times, Reason Magazine). Still, there is something in common in all of those methods that could indeed be causing error.
As for self-perception, I do think that self-perceived personality isn’t always accurate, but in aggregate, I do think people’s self-assesments are good enough to trust. Meaning that there is signal that can be separated from noise.
Thanks for your comments!