The gender-equality paradox, which argues that gender differences are more pronounced in more egalitarian countries and vice versa, has recently been extensively covered in the media. However, a new study refutes the idea.
Numerous studies in recent years have suggested that differences between men and women are more pronounced in countries that score highly on gender equality indices. This phenomenon has been dubbed the gender-equality paradox. But a new study completed at the University of Helsinki indicates that no such conclusions can be drawn.
"Previous research on the gender-equality paradox has been based on an apparently straightforward but highly misleading statistical method," says Postdoctoral Researcher in Psychology Ville-Juhani Ilmarinen.
He has worked with Professor of Social Psychology Jan-Erik Lönnqvist to reanalyse three studies previously published in prestigious, general-science journals. The first focused on PISA data and the differences between boys and girls in attitudes to STEM subjects (science, technology, engineering and mathematics), the second on differences between women and men in economic preferences (altruism, trust, reciprocity, etc.) and the third on gender differences in personality traits. A total of 15 variables were used to examine gender differences. All of the studies involved at least 50 countries and dozens or hundreds of thousands of women and men.
However, when the statistical method was deconstructed, it became apparent that the conclusions drawn in the previous studies were both grossly simplistic and exaggerating the strength of correlations. As country equality increased, the mean levels for women and men remained constant in all 15 variables measured. For example, boys were not more interested in science in more equal countries. Similarly, girls' general interest in science was not associated with gender equality in their country.
Instead, the mean levels for the variables could vary in line with gender equality. For example, both men and women were more open in personality as gender equality increased.
"Our results highlighted the study on economic preferences published in the journal Science. In it, our analysis found only three correlations between the level of equality and women's and men's preferences: men were on average less altruistic in more equal countries, and both women and men were on average more patient in more equal countries. No differences were discovered for the nine other preferences."
More harm than benefit for promoting social equality
The researchers state that gender equality in a given country does not generally correlate with differences between men and women. They find it noteworthy that the scholarly community has previously been unable to respond to such a clear deficiency in the statistical methods used in the studies, although the gender-equality paradox has been criticised from other perspectives.
"When considered in more detail, it's very difficult to show support for the 'gender-equality paradox' as a clear phenomenon, or more generally to summarise the various findings by giving them an umbrella term. In our view, the previous studies have caused more harm than benefit for both production of knowledge and promotion of social equality," notes Ilmarinen.
The researchers say that correlations between country-level mean scores and equality can be examined, but understanding phenomena requires exploring the correlations separately for women and men.
"However, there's reason to assume that further research won't uncover a single universal phenomenon: for some characteristics, women's mean scores vary by equality in their country, for others, the same is true for men, and for yet others, for neither gender."
In recent years, the increasing polarisation of girls' and boys' political attitudes has made headlines. Ilmarinen says no credible research has been published on this phenomenon either.
"So far, the idea is based on data journalism, which is a valid approach, but the phenomena it produces mustn't be confused with peer-reviewed scientific research."
