Responding to Ideas on Sex Differences in Brain and Behavior

The following is a guest post by three co-authors who wished to respond to an earlier co-authored post on this blog, "Sex Differences in Brain and Behavior: Eight Counterpoints." What follows is their response in its entirety. It is a pleasure to host such collegial conversation.

A response to Del Giudice, Puts, Geary & Schmitt’s Eight Counterpoints: Agreements, Ghost Disagreements, and Points for Continued Debate

by Cordelia Fine, School of Historical & Philosophical Studies, The University of Melbourne; Daphna Joel, School of Psychological Sciences and Sagol School of Neuroscience, Tel-Aviv University; and Gina Rippon, Aston Brain Centre, Dept. of Psychology, Aston University, UK. (Authorship is alphabetical: the authors contributed equally to this article.)

Earlier this year, we published an article in which we set out “eight things you need to know about sex, gender, brains, and behavior” (Fine, Joel & Rippon, 2019). Based on our many years observing both the scientific and popular treatment of the topic of sex/gender differences in brain and behavior, our aim was to provide a brief, accessible guide “to help everyone parse the next ‘hot’ new biological explanation of female/male behavior differences by providing eight things to know, look out for, and ask, from the nitty-gritty of whether there even is a difference to the grand sweep of evolutionary explanations.”

We were pleased to see Del Giudice et al.’s response, “Sex Differences in Brain and Behavior: Eight Counterpoints”, engaging with this article, and to note so many points of agreement (Del Giudice, Puts, Geary & Schmitt, 2019). There were, however, also several points of ‘ghost disagreements’–that is, places where Del Giudice et al. argued against views that we did not express and do not hold. Our primary aim here is to highlight the many points of agreement between us, including those framed by Del Giudice et al. as disagreements, and identify and discuss helpful criticisms, observations, and tensions that arise out of their counterpoints.

Points of Agreement

We are pleased that, broadly speaking, Del Giudice et al. largely agree with our suggested ‘eight things’. We all agree on factors of study design that make empirical findings more reliable, including adequate sample sizes, and clear a priori predictions (‘False Positives and False Impressions’). We all think statistical indices that go beyond statistical significance and provide information (directly or indirectly) about overlap or indicate the probability of a randomly picked female scoring higher than a randomly picked male are helpful for understanding the theoretical or practical consequences of an observed sex difference (‘Size Matters’). We all agree that in considering sex differences (‘What Type of Difference Are We Talking About?’), “[a]ge-related, cross-cultural, and contextual variability are real and important.” (Del Giudice et al., 2019).

Moving deeper into the weeds, the points of accord proliferate. We all agree, for example, that sex differences in brain and behavior can reflect the direct effects of sex–namely the direct action of sex-linked genes and hormones on brain and behaviour–but also indirect effects of sex. That is, variables that correlate with sex category may underlie the observed difference. For example, sex differences in appearance can elicit different experiences that in turn influence brain and behaviour (‘Where Do Differences Come From?’). We also agree that it is important not just to look at sex differences in isolation, but also consider how multiple attributes combine (‘How Does it All Add Up?’).  Additionally, we all concur that caution should be exercised in inferring behavioural differences from differences in brain structure or activation (‘What Does a Brain Difference Mean?’). Finally, we all agree that care is required in comparing humans with other species (‘Animal Comparisons’), and that making the case that a sex difference is evolved requires the provision of robust evidence (‘Evolutionary Explanations’). 

These last two points, however, gave rise to several ‘ghost disagreements.’ Below, we, therefore, explain why our position with respect to Evolutionary Explanations, Animal Comparisons, and what counts as a ‘real’ sex difference, is not nearly as divergent to Del Giuduce et al.’s position as they claim.

Ghost Disagreements

What do readers need to know about the ‘Evolutionary Explanations’ that sometimes accompany biological accounts of sex differences in brain and behavior? In our original article, we stressed two points. The first was that evidence about associations between human sex-linked behavior and reproductive success is typically very difficult to acquire. As such, “speculations about what kinds of behaviour would have enhanced reproductive success can owe more to culturally influenced intuitions than scientific evidence”. We also (in ‘Animal Comparisons’) followed evolutionary biologist Marlene Zuk (2002) in warning that: “when it comes to the realm of sex-related behaviors, animal species provide a wide diversity of patterns. It is therefore important that researchers and commentators do not simply choose animal species that support their preferred arguments regarding the evolutionary basis of human gender stereotypes, while species that show different patterns are overlooked.”

Our second point was that ‘evolved’ doesn’t necessarily mean ‘inherited via genes’. Adaptive traits can also be passed on via non-genetic, including behavioral and symbolic, channels of inheritance (Jablonka & Lamb, 2006).

From this, Del Giudice et al. interpret us as claiming that investigations into the evolutionary origins of human sex differences, and seeking insights from non-human animal patterns of sexual selection, are not worthwhile endeavours. This is unfortunate, as it is not the case. For example, together with philosopher of biology John Dupré, two of us have made suggestions of productive ‘outstanding questions’ for future research investigating evolved sex-linked behavior (see “Outstanding Questions” box in Fine, Dupré & Joel, 2017, p. 671). Fine’s book Testosterone Rex discusses sexual selection research from a range of species, as a way of illuminating contemporary insight into the many factors (beyond gamete size) that influence both between-species and within-species variation in mating and parental care. The book further describes the implications of this for our own species, as well as the difficulties in generalizing from non-human animals to ourselves (Fine, 2017). Our point was not that all evolutionary perspectives are misguided, but that it is easier to make an intuitively plausible claim than to provide rigorous evidence. (It’s worth recalling, too, that our primary focus was to assist readers to interpret reports of sex/gender differences in the brain, which are not generally accompanied by any such evidence.) While we may be less forgiving of how compelling such evidence and comparisons are in many cases, we thank Del Giudice et al. for pointing out that these sections of our original piece did not convey the potential of such approaches and comparisons, when properly done.

A second important ghost disagreement is Del Giudice et al.’s claim that we think that any sex differences that are not stable and context- and age-independent are not ‘real’ differences, and therefore unrelated to evolution. This is simply incorrect. Elsewhere, two of us have extensively discussed the point that the development and inter-generation transfer of adaptive traits can be contingent on particular environmental conditions (or ‘developmental systems’) and are neither necessarily fixed nor universal (Fine, 2017; Fine et al 2017; see also Fine & Elgar 2017). Thus we too acknowledge the importance of understanding the different ‘ecologies’ in which sex/gender differences emerge. Contra the claim that we think evolutionary psychology is “based on a rigid genetic determinism”, a statement and implication we made nowhere in our original piece, Fine (2017, p. 185) explicitly observed that Evolutionary Psychologists propose that the expression of adaptive traits can be contingent on environmental context.

Finally, it’s true that we focused on examples of sex/gender differences from the more common small-to-modest end of the size spectrum. We certainly don’t disagree that some sex differences in behavior, such as in sexual preferences, or at the extremes of distributions (such as physical aggression) are substantial. We would, however, note that measures of sex/gender difference (small or large) may themselves be influenced by gender schemas (e.g., Morgenroth et al., 2017). For example, the large sex/gender difference in interest in ‘people vs. things’ to which Del Giudice et al refer has been extensively criticised for this issue (Valian, 2014; Rippon, 2019).

Points for Continued Debate

In many places, Del Giudice et al. seem to interpret our paper as promoting a particular position on the nature and origins of sex/gender differences in brain and behavior. As such, many of the points they make don’t so much provide “balance” with respect to our original aim–to help readers critically interpret reports of sex/gender difference in the brain–as make the case for skepticism regarding reports of social influences on sex/gender differences in brain and behavior. We agree this is a worthwhile goal, but it goes well beyond the scope of our article. Del Giudice et al. thus introduce many new issues, such as the predictions of social role theory, cross-cultural sex/gender differences in personality, and the extent to which we are influenced by gender socialization. The constrained space of a counter-counter-response is not the place to debate these complex topics, although we would welcome further exchange on them in the future since the interpretations that Del Giudice et al. offer relies on contested assumptions. However, we do think it is worth addressing some points of disagreement directly arising from claims that were made in our article.

One major point of disagreement regards the question: are there female and male brains? While we all agree that it is important to consider how sex differences add-up in individual brains, we disagree on the conclusion to be drawn from such a consideration. On the basis of the large multivariate effect size Del Giudice calculated for brain structure (Del Giudice, 2019) and the un-contested observation that it is possible to predict with about 80% accuracy whether one is female or male on the basis of one’s brain structure (e.g., Chekroud et al., 2016; Del Giudice et al. 2016; Joel et al., 2016, 2018; Rosenblatt, 2016), Del Giudice and colleagues claim that “male and female brains may be almost as anatomically distinct as male and female faces”. In contrast, on the basis of animal studies showing that the effects of sex on specific brain features may be opposite under different conditions, and that these sex-by-environment interactions may be different for different brain features, Joel has argued that sex differences do not add-up consistently but rather ‘mix-up’ in individual brains (Joel, 2011, 2012). This hypothesis was later confirmed in an analysis of over 1400 human brains, which found that many brains have both features typical of females and those typical of males (Joel et al., 2015). Subsequent analyses, which directly tested whether the brain types typical of males are different from those typical of females, concluded that they are not (Joel et al., 2018).

The disagreement lies not in the evidence, but rather in the interpretation of what one means in claiming that there are male and female brains. The following example illustrates this point. With a very simple rule–if a brain has more female-typical features than male-typical features, predict female, else predict male–one can accurately predict the sex category of the brain’s owner in ~80% of the cases. (Better prediction accuracy can be obtained by using more sophisticated decision criteria–though such algorithms trained on one population of brains may do much worse when applied to a new population.) The reverse is also true–knowing that a brain belongs to a female allows you to predict, with 80% accuracy, that her brain has more female-typical features than male-typical features (and the opposite for a male). But this doesn’t give you much information about the structure or function of her brain. Specifically, it gives no information about the exact number of female-typical features and male-typical features, nor about which feature is in which form. (By analogy, knowing that someone is female does not enable you to predict their facial features.) Yet it is the precise composition of a brain that determines its structure and function as well as its similarity to other brains. Indeed, Joel et al (2018) have shown that the chances that a woman and a man would have the same brain type are very similar to the chances that two men or two women would have the same brain type (for a thorough discussion see Joel et al., 2018, Joel, Garcia-Falgueras & Swaab, in press; Joel, 2020).

The second point of disagreement concerns the impact of overall brain size on patterns of sex differences. We were a little surprised by Del Giudice et al.’s commentary here. Of the studies they cite, only Ritchie et al. (2018) assessed sex differences with and without taking into account brain size. This study clearly showed that once brain size is taken into account there are marked changes in the pattern of observed sex/gender differences–some sex differences become much smaller, others vanish altogether, and still others reverse. The questions that are still unresolved are how best to control for brain size (as it has been shown that the relation between brain size and the size of specific brain features is not only non-linear but also different for different brain structures, e.g., Liu et al., 2014), and which differences will remain once the best strategy is agreed upon and tested. A very recent paper demonstrates that the number, size and direction of sex differences in grey matter volume varied dramatically as a function of which of five different methods of controlling for total intracranial volume was used (Sanchis-Segura et al, 2019).

Whatever the preferred method, Del Giudice et al. are correct that we need to debate whether the best background assumption is that the average female/male difference in brain size is an extraneous factor to be ‘controlled out’, or a causally important ‘sex effect’ (see Cossins, 2015). Of course, even if we resolve this question to everyone’s satisfaction, the question will still remain of what any remaining differences mean for behavior. In this respect, we remain more concerned than Del Giudice et al. regarding the influence of gender stereotypes, given extensive documentation of their pernicious influence on scientific reasoning from the eighteenth century to the current day (e.g., Bluhm, 2013; Fausto-Sterling, 1992; Fine, 2010; 2013; Joel & Vikhanski, 2019; Jordan-Young, 2010; Maney, 2016; Rippon, 2019; Russett, 1991; Schiebinger, 1991).

Last, we take Del Giudice et al.’s point that the term ‘sexual dimorphism’ is used in biology to refer to sex differences that are not binary, but are nonetheless theoretically or functionally significant. This usage has, however, been criticized by those who work in the area of sexual differentiation of the brain because of the mismatch of the literal meaning with the typical nature of male/female differences in the brain (e.g., McCarthy & Konkle, 2005). As such, it may be misleading for general audiences. But in addition, its catch-all use collapses all of the nuances of the forms, patterns, and contingencies of sex/gender differences discussed by ourselves and del Giudice et al. into a single category, and as such may have outlived its usefulness.

Concluding Remarks

The many points of agreement between ourselves and Del Giudice et al. are not a trivial matter, given a context in which our own work is misrepresented by some as politically biased, ‘blank-slate-ism’ or ‘anti-sex difference’, and the work of Evolutionary Psychologists is sometimes misrepresented as being rigidly biologically determinist and driven by conservative political ideology. Exchanges such as the present one, when focused on evidence and claims, are valuable – and rarer than we would like. They help researchers overcome the confirmation bias to which we are all susceptible. And they help demonstrate to onlookers that disagreement about scientific conclusions is healthy and constructive – science working as it should – rather than evidence of an academy riddled with ideological bias. We thank Del Giudice and colleagues again for engaging with our article and look forward to future exchanges with them.