The gender-equality paradox describes how increased gender differences have been observed in countries that are more progressive in terms of gender roles. The gender-equality paradox most commonly refers to the findings of a 2018 study by Gijsbert Stoet and David C. Geary that, counter-intuitively, suggests that countries with a higher level of gender equality tend to have less gender balance in fields such as science, technology, engineering and mathematics, than less equal countries. This research originally claimed that within the study's sample, more gender equality in a country is linked with a lower proportion of women studying STEM fields. The study and its results received significant coverage in non-academic media outlets. However, separate Harvard researchers were unable to recreate the data reported in the study, and in December 2019, a correction was issued to the original paper. The correction outlined that the authors had created a previously undisclosed and unvalidated method to measure "propensity" of women and men to attain a higher degree in STEM, as opposed to the originally claimed measurement of "women’s share of STEM degrees". However, even incorporating the newly disclosed method, the investigating researchers could not recreate all the results presented. A follow-up paper by the researchers who discovered the discrepancy found conceptual and empirical problems with the gender-equality paradox in STEM hypothesis.
Stoet and Geary (2018) study
Methodology and findings
The study conducted an analysis of the 2015 results of the Programme for International Student Assessment, the largest educational survey of its kind, focusing on the results of questions based on science aptitude and attitudes. This was contrasted with the level of gender equality as defined by the Global Gender Gap Index. The study had a number of primary findings. These can be summarized as follows:
Girls performed similarly or better than boys in two out of every three countries, and were more capable of STEM tertiary education in nearly all countries examined.
Science or mathematics is much more likely to be a personal academic strength for boys than for girls
More girls entered STEM degrees then graduated.
The difference between both the performance of girls in PISA was inversely related to the country's GGGI.
This gap was found to be correlated with the STEM graduation gap, showing that there is a similar gap between the number of girls and boys that enter STEM university programmes compared to those that complete their degrees in more gender-equal countries.
It's important to note that the absolute size of the gap found was not shown to be significant. Rather it is the relative relationship between the two that was found to show an effect. In other words, no relation was found between the total number of girls who entered and completed STEM degrees and the GGGI of the country. Rather, the effect was between the relative difference in number of girls vs. boys who entered and completed STEM degrees, and the GGGI of their country.
Possible causes and criticism
The authors suggest two possible, related causes for the unexpected finding. The first relates to expectancy-value theory which suggests that students determine further education choices based on their relative strengths. Expectancy-value theory is often used to explain the difference in occupational choices between men and women. Thus, the difference would be explained by girls choosing subjects at which they are relatively stronger than STEM fields. In other words, when comparing individual students' aptitude in various areas, girls feel they are stronger in non-STEM areas. An additional explanation put forth by the authors is that the effect is further increased in societies with lower life satisfaction, as defined by the OECD Better Life Index. A cursory statistical analysis confirmed an effect between them. The rationale here would be that students make more economically motivated decisions when experiencing lower life satisfaction. Thus, in wealthier, and more gender-equal societies, students feel freer to choose study based on their interests, rather than economic motivating factors. Longitudinal studies of the same data set that was analyzed by Stoet and Geary investigated the results of implicit-association tests, which measured perceived differences in gender. No relationship was found between interest in STEM, as reported via PISA, and perceived differences in gender according to IAT. This suggests that the perception of gender-related stereotypes might have no effect on career choice. Contrary to this study, another study measured both stronger implicit gender stereotypes as measured by the IAT as well as explicit stereotypes measured by a simple questionnaire in different countries and showed an inverse relation to the representation of women in science in those countries. Other studies have challenged the idea that professed interest is a good measure of intrinsic interest. For example, one study found that the number of women already in a field predicts the stereotypes people have about that field. Related to this is another study which found a relation between the perceived sexism in a specific degree program and the expressed interest in the field amongst girls considering it. Still further studies have shown that there is a significant overlap between parent and teacher expectations around gender and STEM, and what these children express. For example, one study found that parents were less likely to think their daughters would be interested in STEM areas, and that this belief was a strong predictor of later attitudes and efficacy at science. Still further longitudinal studies found a similar effect between a mother's prediction of her daughter's success in STEM and the daughter's later career choices. Similar analyses of the effects of bias in teachers in associating STEM with boys, rather than girls, was shown to also predict future interest in STEM. Many of these longitudinal studies were performed on middle and high schoolers, showing this effect is present before the age where PISA would be taken.
Related studies
Charles and Bradley (2009) study on economic development and gendered study choices
In 2009, Maria Charles and Karen Bradley conducted data analysis of sex segregation by field of study in 44 societies, finding a higher level of segregation in more economically developed contexts. This result could be relevant to the debate on gender equality and gender balance in study choice, as economic development is linked to gender equality. The authors note that this result seems paradoxical, as it contradicts accounts linking socioeconomic modernization to a "degendering" of public-sphere institutions such as schools and universities.
Falk and Hermle (2018) study on the relation between gendered preferences and gender equality
In 2018, Armin Falk and Johannes Hermle looked at data on 80,000 people in 76 countries to find out what might influence gender-associated differences in preferences, such as the willingness to take risks, patience, altruism, positive and negative reciprocity. Their main observation is that the more equal opportunities there are for women, the more women differ from men in their preferences. These statistical results could help explain Stoet and Geary's finding that women's educational choices diverge from men's in countries with greater gender equality, because greater gender equality could make their preferences diverge from men's.
