Andy Perrin responded on Scatterplot to a Twitter debate that happened yesterday and I couldn’t resist adding two cents. It started with a link posted by Nicholas Christakis to a review article on gender differences in personality across societies. The main claim of the article is that gender differences in personality are larger in more gender-egalitarian countries, providing support for evolutionary theories of gender differences and against social role theories, which would predict gender convergence in egalitarian countries.
Steve Vaisey commented, “I am genuinely interested to hear how sociologists who study gender would react to these findings.” Andy’s response, which is as usual worth reading, argued against the study’s interpretation from a social constructionist perspective.
This sent me down the rabbit hole of actually reading the article and a bit of the research it is based on. I am not impressed.
Let me qualify that I am not a gender scholar, nor am I deeply familiar with this literature. My priors are that gender differences in personality are both biological and social, but that the average person (obviously not everyone) is insufficiently skeptical of biological explanations because they fit our expectations and stereotypes. I am inclined to be doubtful about this kind of research, but open to evidence.
Here’s two reasons this article left me underwhelmed.
First, the lit review is sloppy in a way that makes me not trust the authors about other things — say, the quality of their data collection across relatively small samples in dozens of countries. It sets up social role theory as a straw man (“social role theories of gender development contend any and all ostensible differences between men and women are primarily the result of perceived gender roles” [p. 47], when the debate is really about relative importance, not “any and all ostensible differences”). It quotes from an article about social role theory (“men and women have inherited the same evolved psychological dispositions” [p. 47]), but the page is not part of the article, and the quote does not seem to appear in the article at all. Based on Googling, it appears to come from a misquote in an edited volume. This may seem trivial, but if you’re asking me to trust that you used good research methods on a study that involved data collection in 50+ countries, I’d like to know that I can count on you to represent the literature accurately.
Second, I dug into one of the more prominent empirical studies in the review, by the same lead author. The review describes the study like this:
More egalitarian gender roles, gender socialization and sociopolitical gender equity, however, were associated with larger gender differences. For example, the largest overall gender differences in personality were found in relatively high gender egalitarian cultures of France (d = −0.44) and the Netherlands (d = −0.36), whereas the smallest gender differences were found in the relatively low gender egalitarian cultures of Botswana (d = 0.00) and India (d = −0.01).
The examples here are somewhat cherry-picked. Yes, these are the top two and bottom two countries. But if you look at their whole chart (p. 173), there’s more to the story. The five countries with the biggest gender differences in personality (measured as mean gender difference in big-five traits) are France, Netherlands, the Czech Republic, Brazil, and Belgium. The five most gender-similar countries are Indonesia, Congo, Fiji, Botswana, and Finland.
If we look at the World Economic Forum Global Gender Gap indicator, France is the only country on the top five that makes it into even the top 20 countries for gender equality. Brazil and the Czech Republic are both basically at the global median. Finland, which has among the most similar personality scores by gender, is the third most gender-equal country in the world. So while correlations may be there across all countries, this is hardly dispositive unless you can come up some explanation for why gender egalitarianism leads to personality difference across genders in France, but similarity in Finland.
And when they actually dig into the relationships, they appear to be flimsy and overstated. The study examines nine country-level measures of gender equality. After controlling for the country’s development level, only four measures show a significant relationship to the gender-personality gap at the p < 0.05 level (p. 177). So by most measures, the study finds no relationship. Notably, neither of the UN composite measures of gender equality, the Gender Empowerment Measure and Gender Development Index, are significantly related.
The four measures that appear to be related to mean country-level gender difference in personality are 1) traditional values, 2) cultural trust (“can most people be trusted”), 3) the gender gap in smoking, and 4) “when the respondents are more inclined to agree with a question irrespective its content” [sic — I don’t even know what that means]. Two of these seem very questionably related to gender, and I’m going to discount those.
That leaves us with the following: countries with more traditional values have less sex differentiation in personality traits, and countries where women smoke a lot have more sex differentiation in personality traits. Both relationships are significant at the p < 0.05 level — but note that we have already run through several possible measures that turned out not to be significant.
From this, the study winds back up to its dramatic conclusion: “in more prosperous and egalitarian societies the personality profiles of men and women become decidedly less similar” (p. 178). More prosperous, yes: development level, not gender equality, is the best explainer of gender-personality differences. But more egalitarian? Color me unconvinced.
And from that, the abstract jumps to: “It is proposed that heightened levels of sexual dimorphism result from personality traits of men and women being less constrained and more able to naturally diverge in developed nations.” Wow. The word “naturally” is doing an awful lot of work there.
The review article covers a bunch of other studies, too, but based on what I’ve read it does not seem worth the time to dive in further. My priors — that personality differences are both biological and social, and that people are too credulous of biological explanations of gender differences — remain unchanged.