Are boys born better than girls at maths? Or are the differences some of us observe a result of our different upbringings? Perhaps a bit of both?
In this data-driven story I compare and contrast country data from the maths portion of a global education test known as PISA (taken by more than half a million 15-year-olds around the world), with country data from an index used to measure the gap between women and men in education, the economy and political empowerment (formally known as the Gender Equity Index).
The results may surprise you!
The closer the index is to 100, the more equally men and women are treated. Similarly, the higher the maths score the better.
So to recap, the further up a country is on the scatterplot, the better its girls are at maths, and the further right a country lies, the more gender equal it is.
What happens if we choose to ignore the first two (education and political empowerment gaps) and focus on economic gaps between men and women?
We'll talk more about labour force gaps a bit further down, but for now let's see what happens if we ignore labour force gaps and focus on income gaps between men and women.
The closer the score is to 100, the smaller the gap.
This time the vertical axis represents the average PISA maths scores for boys, and the horizontal axis the Gender Equity Index.
So values greater than 0 mean that girls scored better than boys on average, and values less than zero mean that boys scored better. The horizontal axis represents the Gender Equity Index.
While the precise nature of these societal norms is beyond the scope of this story, I will leave you with a couple of scattered thoughts taken from Jo Boaler's book Mathematical Mindsets (Jo is a leading maths education researcher at Stanford) :
In an important study, Sian Beilock and colleagues found that the extent of negative emotions elementary teachers held about mathematics predicted the achievement of girls in their classes, but not boys (Beilock, Gunderson, Ramirez, & Levine, 2009)...this gender difference probably comes about because girls identify with their female teachers, particularly in elementary school. Girls quickly pick up on teachers' negative messages about maths - the sort that are often given out of kindness, such as 'I know this is really hard, but let's try and do it' or 'I was bad at math at school' or 'I never liked math.'
In another study, researchers found that when mothers told their daughters "I was no good at maths in school" their daughter's achievement immediately went down (Eccles & Jacobs, 1986).
The body of work on 'stereotype threat', led by the work of Claude Steele, shows clearly the damage caused by stereotypical ideas. Steele and his colleagues showed that when girls were given a message that a maths test resulted in gender differences, the girls underperformed, whereas girls who did not receive that message performed at the same level as boys on the same test. Subsequent experiments showed that women underachieved when they simply marked their gender in a box before taking a test.
Maths is often taught in such a dry, procedural, abstract way...unfortunately though, that dry, procedural maths particularly puts off girls and we know that girls, for whatever reason more than boys, want a connected subject. They want to see the connections not only between maths and the world and life, but they want to see the connections within maths as well. They want to see maths as this connected, conceptual subject, which is the subject it really is. But it's not taught in that way and that is a big part of the problem!
The 2012 Gender Equity Index data can be found here.
The 2015 PISA Maths Scores data can be found here.
The purple line that appears in the first five scatterplots is a simple linear regression line. In plain english, it's a line of best fit.
The analysis was limited to 41 countries as only 41 countries were included in both datasets : Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Italy, Japan, Korea, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovakia, Spain, Sweden, Switzerland, Turkey, Great Britain, United States, Brazil, Chile, Estonia, Idonesia, Israel, Russia, Slovenia, Latvia, Singapore, Columbia, Peru.
Scatterplots built using d3.js
Source code can be found here.