Integrating Approaches Requires More Than a Division of Labor

In a recent article, Wölfer and Hewstone argue that evolutionary psychology—and, in particular, sexual-selection theory—is useful for understanding sex differences in same-sex aggression, while social-role theory is best applied to sex differences in opposite-sex aggression. Wölfer and Hewstone tested this proposal using a rich data set containing high school students’ reports of their peers’ aggression. They regressed classroom-level sex differences in same- and opposite-sex aggression onto five variables drawn from the two theoretical positions. Three variables (gender and masculinity norms, derived from social-role theory, and body dimorphism, derived from evolutionary psychology) did not differ in their association with the two forms of aggression. Another variable (operational sex ratio, derived from evolutionary psychology) was not interpretable because it was confounded with the number of available targets, but a fifth (social hierarchy in males, derived from evolutionary psychology) predicted sex differences in same-sex but not opposite-sex aggression. Our focus is not on the study itself but on the authors’ proposal that theoretical oppositions between evolutionary psychology and social-role theory can be resolved by assigning one form of aggression to the former and another to the latter. We believe that this argument mischaracterizes both theories, reinforces the divide between evolutionary and social theorizing, and falls short of integrating the two.

Evolutionary psychologists often argue that sex differences in aggression arise because of male-male competition, which in turn happens because variance in reproductive success is greater for males than females (Wilson & Daly, 1985). This does not make evolutionary psychology relevant only to male-male aggression. Selection has acted on morphological and psychological traits that facilitate aggression in men or inhibit aggression in women. The result can be viewed as men having a greater “taste for risk” (Wilson & Daly, 1985, p. 59) or women being more fearful (Campbell, 1999). Either way, the traits in question are domain general: They have implications for sex differences in many forms of risk taking (Cross, Copping, & Campbell, 2011; Cross, Cyrenne, & Brown, 2013). Consequently, there is a robust sex difference in risky physical forms of aggression, while there is very little sex difference in the use of low-risk indirect aggression (Campbell & Stockley, 2013). (In Wölfer and Hewstone’s study, aggression was operationalized as being “mean” to someone, which makes interpretation of sex differences difficult.) Aggression arises from escalated conflicts of all kinds. Males’ greater strength and willingness to take risks is relevant in all disputes that could escalate to aggression—not just those with a same-sex opponent. Evolutionary psychology has examined opposite-sex aggression with reference to mate guarding and paternal uncertainty in men, and more broadly in terms of the inherent potential for bilateral conflict in long-term pair bonds (see Archer, 2013, for a review). Wölfer and Hewstone’s supposition that evolutionary psychology is relevant only to same-sex aggression lacks a clear rationale.

Social-role theory (like evolutionary psychology) aims to explain a broad range of sex-typed behaviors and traits. According to this theory (Wood & Eagly, 2012), gendered division of labor causes gender stereotypes about personality to arise via correspondent inference (Jones & Davis, 1965). These stereotypes foster prescriptive norms that shape men’s and women’s behavior in a range of situations. In other words, “internalized gender roles produce gender identities that act as trait-like [emphasis added] determinants of aggressive behaviors” (Eagly & Wood, 2009, p. 276). The term “trait-like” implies consistency across targets, and, indeed, social-role theory explicitly encompasses same-sex aggression (Wood & Eagly, 2002). The extent to which an individual internalizes and conforms to gender norms should predict aggression across contexts. Wölfer and Hewstone’s proposed application of social-role theory to opposite- but not same-sex aggression, therefore, neglects the fact that social-role theory already offers an explanation for sex differences in same-sex aggression. In practice, Wölfer and Hewstone found that “masculinity norms” (p. 1289; measured as approval of male violence) were unconnected with sex differences in either same-sex or opposite-sex aggression.

Evolutionary psychology and social-role theory are viewed as competing frameworks because while both theories aim to explain sex differences in all forms of aggression (e.g., see Archer, 2009; Eagly & Wood, 2009), they use theoretical approaches that have been crudely characterized as “nature” (evolutionary psychology) versus “nurture” (social-role theory). Encouragingly, researchers from both camps have made attempts at integration over the last 20 years. Evolutionary psychology has increasingly acknowledged the role of culture: at the micro level, examining its transmission between individuals, and at the macro level, demonstrating culturally driven niche construction as a source of genetic selection (Brown, Dickins, Sear, & Laland, 2011). The rigorous investigation of content biases (Boyd & Richerson, 1985)—the shaping of cultural products by the structure of an evolved mind—represents a promising avenue for understanding the creation and transmission of the gender stereotypes so central to social-role theory. Likewise, proponents of social-role theory, who initially rejected any role for sexual selection in the genesis of psychological sex differences (Wood & Eagly, 2002), now acknowledge possible biological differentiation in infant temperament, on which gendered socialization acts (Wood & Eagly, 2012). The goal of integration is to find a satisfactory means of incorporating biological and cultural predictors within a single framework. Wölfer and Hewstone’s proposed “dual-theory approach” (p. 1293) is the very opposite of this.

In pursuit of genuine integration, we suggest focusing on the way in which risk sensitivity is responsive to social context (e.g., see Cross & Campbell 2011, 2014). Risk sensitivity is a trait that differs between men and women (Cross et al., 2011) and directly affects aggression levels (Campbell, 2013, 2015; Eagly & Steffen, 1986). Furthermore, the perceived risk associated with aggressive acts depends on social context (sex of the target) and cultural prescriptions (differential permissibility of aggression as a function of the perpetrator’s and target’s sex). For example, cultural norms that support male-male fighting as an index of masculinity can, for some men, make refusal to fight a riskier option than fighting (Hochstetler, Copes, & Forsyth, 2014). Sex differences in intimate-partner violence also correlate with cultural acceptance of male violence (Archer, 2006). Note that sex (of the perpetrator or target) is not treated as a causal variable in itself (see Maney, 2016). Instead, sex is a proxy measure for variables that affect the risk of aggression (strength, risk sensitivity, internalized beliefs about perpetration and victimization, etc.).

In summary, we welcome Wölfer and Hewstone’s use of social-network-based data to investigate sex differences in aggression as a function of the sex of the target. However, applying evolutionary and social theories in a piecemeal fashion, rather than furthering integration, reinforces an artificial and counterproductive dichotomy between them.