ABSTRACT
Time and again, research has shown that men are less accepting of homosexuality than women. Studies on such attitudinal sex differences have been overwhelmingly conducted in Western democracies, however, with a special focus on the U.S. Whether the sex difference in attitudes towards homosexuality is a worldwide phenomenon has not yet been investigated. Using data from the seventh wave of the World Values Survey (2017–2021), this article provides evidence that the sex difference is not universal, but limited almost exclusively to Europe and the Americas, indicating the need to replicate studies conducted in these societies in global cross-country comparisons. Contrary to predictions of the social role theory or biosocial construction theory, but in line with predictions from evolutionary psychology and a growing number of empirical studies in this field, the sex difference in attitudes towards homosexuality widens with rising gender equality and development, especially when the two coincide.
Disclosure Statement
No potential conflict of interest was reported by the authors.
Supplementary Material
Supplemental data for this article can be accessed online at https://doi.org/10.1080/00224499.2023.2198500.
Notes
1 An exception is the study by Kite and Whitley (Citation1996), who clearly stated the reason for their choice of countries. They included only studies with respondents from the U.S. and Canada in their meta-analysis “to control for possible cultural differences in attitudes” (Kite & Whitley, Citation1996, p. 340). Instead of relying on the study design to control for potential cultural differences, our analysis investigated if and to what extent such cultural differences do exist.
2 The term weird has been criticized for being unclear about the extent to which the individual elements actually belong together (Syed & Kathawalla, Citation2020). Recently, Sakaluk and Daniel (Citation2022) empirically showed considerable differences with respect to these elements (Education, Industrialization, Richness, and Democracy) both within the group of Western and non-Western nations. Instead of a classification into weird and not weird countries, we differentiated between Western and non-Western countries and focused on two macro-level predictors that play an important role in existing research: development and gender equality (see below).
3 The greater parental investment of women does not only relate to their greater responsibility for child rearing in most societies, but also to pregnancy and a limited source of only around 400 ova, while men produce millions of sperm cells every day.
4 An exception is disgust due to an alleged health threat that – according to Pirlott and Cook’s (Citation2018) study – only gay and bisexual men (but not lesbian and bisexual women) are considered to pose. In this case, the researchers did not differentiate between heterosexual men and heterosexual women feeling disgusted, so we could not derive a hypothesis for our analysis from their proposition.
5 See the note to for the construction of the Mean Gender Role Attitudes Index.
Table 2. Descriptive information on variables.
7 I.e., it is not easy – often virtually impossible – to see from a regression table in which range of the X–variable a variable Z has a significant influence on Y. Even the interpretation of the significance of the coefficient on the interactive term X*Z itself depends on the values of Z. Thus, interpreting an interaction just from a regression table is not advised. As Brambor et al. (Citation2006) showed, “it is perfectly possible for the marginal effect of X on Y to be significant for substantively relevant values of the modifying variable Z even if the coefficient on interaction term is insignificant. [… This] means that one cannot determine whether a model should include an interaction term simply by looking at the significance of the coefficient on the interaction term. […] The point here is that the typical results table often conveys very little information of interest because the analyst is not concerned with model parameters per se; he or she is primarily interested in the marginal effect of X on Y (β1 + β3Z) for substantively meaningful values of the conditioning variable Z. While it is often possible to calculate the marginal effect of X for any value of Z from the typical results table using a little algebra, the problem is that it is not possible to do the same for the standard errors. This is because the relevant elements of the variance-covariance matrix necessary to calculate the standard error […] are rarely reported, i.e., cov(β1β3)” (Brambor et al., Citation2006, p. 74).
8 With being the mean of the country means of the acceptance of homosexuality.
9 Since the results from the two-way interaction model M4 are largely similar to those from the three-way interaction model M5, and the latter exhibits the better model fit in terms of AIC and R2 at the individual and country level, only M5 will be presented and discussed here.
10 This model uses the GNI p. cap. as a measure of development and not HDI, since multicollinearity between GII and HDI is too big (r = 0.86) for a meaningful modelling of the respective interaction.
11 An explanation for this non-finding could be that the GII also contains a strong female health component, which might be less relevant for attitudinal aspects than the areas of politics, economy, and education addressed in the other two gender equality indices.