If you actually look at the data, most of the common claims of psychological differences between men and women are unfounded. For example, mathematical ability, verbal skills, temperament, emotionality, aggression, leadership effectiveness, moral reasoning, conscientiousness, and self-esteem. Only in very few instances are commonly cited sex differences corroborated by empirical data, but even in these cases - such as male superiority in mental rotation or abstract reasoning - the data show the true differences to be extremely small. These insights do not come from isolated and hard-to-replicate individual studies. Male and female differences have been studied for more than a century: the evidence refuting them is drawn from entire bodies of research, where datasets have been assembled, pooled, and analysed in their totality.

Two good meta-analytic reviews to start with: The Gender Similarities Hypothesis and Gender Similarities and Differences. After scrutinising the cumulative research evidence on psychological sex differences, Hyde conclude that assumptions of gender differences should give way to a “gender similarities hypothesis” – i.e. psychologists should begin with a working assumption that they will more frequently not encounter gender differences in their research. Bear in mind, Hyde’s reviews cover research that was originally conceived what were assumed to be likely sources of gender difference, and yet still found very little.

One example is Mathematical Performance:

In an early meta-analysis of data drawn from over 3 million people, the computed gender difference in mathematical performance equated to a total effect size of d=0.05 (Hyde, Fennema & Lamon, 1990). This is virtually trivial. The effect size statistic ‘d’ represents the difference in 2 scores divided by the standard deviation of the data as a whole, and a d of 0.05 equates to a shared variance of just 0.06%. In other words, 99.4% of the variability in maths scores resulted from factors other than the participant’s genders (e.g. individual differences in ability or education).

A more up-to-date meta-analysis, covering over 7 million boys and girls, found differences that were similarly minute: in this sample, ds ranged from 0.02 to 0.06 depending on the participants ages (Hyde, Lindberg, Linn, Ellis, & Williams, 2008). The non-existence of a sex difference is seen in adults too, and in high-level mathematics (Lindberg, Hyde, Peterson & Linn, 2010). In short, on the basis of studies of literally millions of people, it is highly misleading to claim that men and women differ in mathematical ability in any meaningful way.

Possible Exceptions:

Men typically outperform women in mental rotation tests, with one meta-analysis reporting an effect size of d=0.57 - i.e. a shared variance of 7% between gender and mental rotation ability (Maeda & Yoon, 2013). However, we also know that mental rotation improves with repetition; therefore, the gender difference seen in the data may just reflect different participation levels in extracurricular activities that involve the practicing of mental rotation for fun (e.g. video games).

There is an empirical gender difference in self-reported sexual behaviour, but whether self-reports equate to actual behaviour is something we’ve gone over a thousand times so I won’t get into it here.

Emotional stability. Large-scale personality studies often report sex differences in neuroticism and agreeableness (Costa, Terracciano & McCrae, 2001). However, while most of these findings arise from data gathered from US adults, the sex differences are not consistent across countries and are virtually non-existent in some (Hyde, 2014). The very fact that these sex difference vary cross-culturally suggests that they are the result of cultural contexts, rather than of intrinsic natural differences between male and female human beings.

One further thing to think about: statistical studies have shown that the most commonly cited sex differences refer to variable that cannot reasonably be compared across gender lines – they are dimensional (range continuously from low to high) rather than taxonomic (involve non-arbitrary categories). Because they refer to characteristics that occupy ranges instead of categories, these variables are distinct from biological gender itself, which is primarily a two-category construct. This means that such characteristics do not easily lend themselves to separation into binary groupings. As such, when conceptualising sex differences is it seldom statistically justified to declare that men are ‘like this’ while women are ‘like that’ (Carothers & Reis, 2013). In the end, people vary in all sorts of ways as individuals.