January 05, 2012

Huge differences in personality between men and women

From the paper's conclusion:
In conclusion, we believe we made it clear that the true extent of sex differences in human personality has been consistently underestimated. While our current estimate represents a substantial improvement on the existing literature, we urge researchers to replicate this type of analysis with other datasets and different personality measures. An especially critical task will be to compare self-reported personality with observer ratings and other, more objective evaluation methods. Of course, the methodological guidelines presented in this paper can and should be applied to domains of individual differences other than personality, including vocational interests, cognitive abilities, creativity, and so forth. Moreover, the pattern of global sex differences in these domains may help elucidate the meaning and generality of the broad dimension of individual differences known as “masculinity-femininity” [11]. In this way, it will be possible to build a solid foundation for the scientific study of psychological sex differences and their biological and cultural origins.
From the press release:
The researchers used personality measurements from more than 10,000 people, approximately half men and half women. The personality test included 15 personality scales, including such traits as warmth, sensitivity, and perfectionism. When comparing men's and women's overall personality profiles, which take multiple traits into account, very large differences between the sexes became apparent, even though differences look much smaller when each trait is considered separately. However, the study indicates that previous methods to measure such differences have been inadequate, both because they focused on one trait at a time and because they failed to correct for measurement error.

PLoS ONE 7(1): e29265. doi:10.1371/journal.pone.0029265

The Distance Between Mars and Venus: Measuring Global Sex Differences in Personality

Marco Del Giudice et al.

Abstract

Background

Sex differences in personality are believed to be comparatively small. However, research in this area has suffered from significant methodological limitations. We advance a set of guidelines for overcoming those limitations: (a) measure personality with a higher resolution than that afforded by the Big Five; (b) estimate sex differences on latent factors; and (c) assess global sex differences with multivariate effect sizes. We then apply these guidelines to a large, representative adult sample, and obtain what is presently the best estimate of global sex differences in personality.

Methodology/Principal Findings

Personality measures were obtained from a large US sample (N = 10,261) with the 16PF Questionnaire. Multigroup latent variable modeling was used to estimate sex differences on individual personality dimensions, which were then aggregated to yield a multivariate effect size (Mahalanobis D). We found a global effect size D = 2.71, corresponding to an overlap of only 10% between the male and female distributions. Even excluding the factor showing the largest univariate ES, the global effect size was D = 1.71 (24% overlap). These are extremely large differences by psychological standards.

Significance

The idea that there are only minor differences between the personality profiles of males and females should be rejected as based on inadequate methodology.

Link

24 comments:

Auditor George Danos said...

I find the datum that "only 10%" overlap between the sexes is detected in the multivariate distribution. Now, who recalls which personality trait was supposed to be overestimated by a factor of at least two, when ten percent was claimed?

Grey said...

Hopefully the Berlin wall is starting to crack.

Karl Zimmerman said...

While it is true this study found big differences on personality between men and women, one glaring flaw I see is the sample size, while large, is exclusively American. Thus, it cannot show if there are innate differences between the personalities of men and women. To do that, you'd need a significant sample size taking the same test in a number of divergent cultures.

I say this as someone who leans strongly towards the nature side, rather than nurture. This is a great first step, but I'm sure people will jump to conclusions which simply have not been proven in this paper.

andrew said...

One doesn't need all of the assumptions in the paper to get most of the effect. The fine grained measures of personality used and multivariate simultaneous clustering kind of analysis will get you there, and both of those points seem legitimate. Indeed, one could even do the same analysis, if the data was available, broken down to individual question answers on the surveys much as you would individual SNPs in genomic data.

The study talks about different hypothesized sources for the differences, but doesn't actually substantiate or link to their data any of that context.

It would be nice to see them spell out what result they would get if they used the methodology they criticized. They do break down the multivariate v. univariate analysis effect, but if they also compress their fifteen factor data into five factors and compare it, I missed that.

To the extent that they get a smaller degree of gender overlap than past meta-analysis, even when their data is parsed with the same methodologies used in the older studies, which it looks like they do, then at least some of their result is due to a starker distinction in the data and not methodology.

Onur Dincer said...

From my viewpoint, they are just stating the obvious. But for those who delude themselves by thinking that men and women have similar personalities, this research will be a headache.

eurologist said...

Another question is, whether different personalities matter in real life, and how they matter. We are all social beings, so being exposed to extremely different personalities (are the sex differences significantly larger than the in-sex variations?) changes us - but how? Surely, sometimes, for the better, and leading to better (more productive?) outcomes.

sykes.1 said...

If there is a nuture effect, then increasing the sample beyond Americans will almost certainly increase the measured differences. Almost all the non-Americo-European minority are strongly paternalistic and have, if anything, exaggerated sex role differences.

Charles Nydorf said...

"Sex differences in personality are believed to be relatively small"
I wonder what world these investigators are living in!

pconroy said...

@Onur,

I totally agree. This result is obvious to anyone with a brain, it's only unclear to feminists - both the female and male types?!

Anonymous said...

@George: I agree that the 10% figure is suggestive and that it may well represent—or correlate strongly with—the gay population. This might not be welcome news to either gay liberals ("no correlation between male homosexuality and effeminacy") or to religious rightists, who love to whittle down that ten percent figure.

Of course, the ten percent overlap might be miss some gays and include many atypical straights as well.

Didn't this occur to anyone else?

Matt said...

Lewontin's Fallacy works everywhere?

Roy said...

@ Eurologist They do matter, barely 1% of the political leadership in the last century has been female (rulers.org)

Andrew Oh-Willeke said...

"one glaring flaw I see is the sample size, while large, is exclusively American."

This actually is a plus, as it elminates cultural confounds in the data. It does mean that the data aren't universal, and that they can be attributed to a non-nuture cause. But, the point of this article is to illustrate the existence of a clear gender difference in personality in contradiction to previous studies that also used exclusive American studies. By drawing a 10,000 person sample from one culture, the statistical power of that part of the analysis is enhanced.

Belenos said...

Well, I'm amazed that this research needs doing. But it says nothing about nature versus nurture, "personality", as measured by this test, is merely a social performance, both in terms of the test and the wider social context.

Personality is only expressed in terms of relationship with others, and so it is necessarily a product of social conditioning. Studies like this will do nothing to show how much deeper biological basis there is for it.

Study marks:

Stating the obvious 10/10
Scientific value 0/10

Dienekes said...

The important thing about this research (no matter what one thinks about the construct of "personality") is that it shows that by combining a few personality factors, one can guess whether someone is male or female with relatively high confidence (that's what the 10% overlap means).

So, it doesn't matter what one thinks about "personality". Whatever one thinks of it, it can be used to guess whether someone is male or female, so it _does_ capture something significant that is different between the sexes.

On a related point, I would love to see this research repeated for different races. There are small differences in personality between different racial groups, and it would be interesting to see whether by combining these there is a large or small overlap between the races. My bet is on there being a small overlap.

Actually, a more interesting question is this: is a man of a particular race going to be (on average) closer personality-wise to a female of his own race or to a man of a different race?

Belenos said...

In terms of Dienekes' comment, yes I agree that this test is an excellent predictive tool to decide if the subject is an American male or an American female, but we have other tools which do this (eyes for example). However, in contrast with what Andrew said, no biological extrapolations at all can be made from this.

We have reams of data from social anthropology and sociology which shows that men and women have different codes of acceptable behaviour, and different "approved social responses". The questions in the survey will elicit different responses from men and women in all cultures, though there will also be differences between the answers given by men and women in each culture, as the code of male/female approved behaviour and emotional response differs by society.

I'd say this will show a "racial" difference, if one wants to call it that, but it can give no valid indication of any biological basis for this difference, except in circumstances so controlled as to be impossible in the real world.

Roy said...

@ Belenos You are probably wrong. Politics too is "expressed in terms of relationship with others" but is not 'necessarily' a product of social conditioning. Indeed the indications are emerging that political preference has a complex relationship with size variations of specific neural regions, the amygdala for one.

Belenos said...

@Roy. I find it very probable that there is a biological basis (in terms of male/female difference) to some aspects of behaviour which we represent under the non-scientific term "personality". At the very least, the different hormonal effects should produce some very different responses.

It's just that this research is so methodologically flawed that it can tell us nothing about the topic. It's so bad that anybody who doesn't believe in any gender based "personality" differences can simply write the study off.

But that's psychology for you.

Pascvaks said...

Thoughts-

As a father and 'old' man, the difference between X and Y starts at conception and goes all the way to old age.

The mind and personality are further changed during puberty and maturity.

Many 'characteristic' drivers of X and Y are hormonal. I can vouch for Y. Anyone care to try this experiment will experience it first hand, I don't think you need to have your prostate removed but it may help, just start chemical castration injections and continue for five years. Six months in, you'll be neither X nor Y; very different indeed.

Roy said...

@ Benelos I was at Uni with an exceptional woman who was being exhorted to pursue a career in pure math. She failed our engineering drawing class outright.
Perhaps a way forward could be to resolve the spectrum of 'personality attributes' in a model of gender constructs of personal and social spaces. More obviously the 'time domain' components of
hormonal activity and their relationship to 'personality' could be probed. If these come to pass, I'll expect financial recognition rather than being modded up or cited :-)

Belenos said...

Roy: I think your friend from uni's experience would come under the group of displayed behavioural characteristics which psychologists unthinkingly label "intelligence", rather than the group of displayed behavioural characteristics which psychologists unthinkingly label "personality".

Grey said...

"This result is obvious to anyone with a brain, it's only unclear to feminists - both the female and male types?!"

I don't think it's just feminists. The critical point isn't just the differences - which everyone knows is true - but the overlap.

The pre-feminist culture wouldn't make exceptions for the overlap. The feminist culture doesn't accept basic reality. The solution is to accept basic reality including the existence of an overlap.

Belonos
"However, in contrast with what Andrew said, no biological extrapolations at all can be made from this. We have reams of data from social anthropology and sociology which shows that men and women have different codes of acceptable behaviour, and different "approved social responses"."

Then what about the overlap?

Pascvaks said...

@Grey -

"Then what about the overlap?"

One would think, the 'overlap' would show itself in the capability/acquired knowledge of what works in this/that situation; knowledge gained by male children, boys, and men watching female children, girls, women and by girl children, girls, and women watching male children, boys, and men. The ability to perform common and/or uncommon tasks 'normally' performed 'better' (by convention over time) by the other sex but not restricted to the opposite sex by a physical limitation. Thus, a common genetic capability with 'advantage' (or overlap), but not unique to one or the other?

Always thought we had more in common genetically, than different.

Easy772 said...

In terms of Dienekes' comment, yes I agree that this test is an excellent predictive tool to decide if the subject is an American male or an American female, but we have other tools which do this (eyes for example). However, in contrast with what Andrew said, no biological extrapolations at all can be made from this.
Belenos:
"We have reams of data from social anthropology and sociology which shows that men and women have different codes of acceptable behaviour, and different "approved social responses". The questions in the survey will elicit different responses from men and women in all cultures, though there will also be differences between the answers given by men and women in each culture, as the code of male/female approved behaviour and emotional response differs by society.

I'd say this will show a "racial" difference, if one wants to call it that, but it can give no valid indication of any biological basis for this difference, except in circumstances so controlled as to be impossible in the real world."

I agree with you for the most part. "it is virtually impossible to understand how biology works, outside the context of environment." And all that good stuff.

And also "Epigenetic theory is an emergent theory of development that includes both the genetic origins of behavior and the direct influence that environmental forces have, over time, on the expression of those genes. The theory focuses on the dynamic interaction between these two influences during development."

I do however think certain populations are more prone to produce individuals with certain traits and/or talents
http://www.livescience.com/3636-boys-warrior-gene-join-gangs.html