Wednesday, February 1, 2012

Trends in Race/Ethnicity and Gender Representation in the Mathematical Sciences

Each year, the US Department of Education’s National Center for Education Statistics publishes its Digest of Education Statistics. Among its many annual tables, easily accessible back to 1990, is one that describes the number of Bachelor’s degrees by sex, race/ethnicity, and field of study. I’ve put together my own longitudinal table for the science, engineering, and mathematical majors. These have been the source of my tables on trends for women (see We Are Losing Women from Mathematics, September 2009) and for members of various racial or ethnic groups (see MAA Speaks Out on Capitol Hill, October 2009).

The disturbing trends that I identified then have continued. In particular, the number of women majoring in the mathematical sciences has continued to stagnant while the number of men has seen very healthy growth (see Graph 1). I will present the graphs of the number of Bachelor’s degrees in the mathematical sciences by race/ethnicity and gender and then conclude with some observations.

By racial or ethnic group, we see the same pattern reflected among Asian and Pacific Islander majors and, to a lesser extent, among Hispanic students (Graphs 2 and 3).

The situation is very discouraging for Black non-Hispanic students. The absolute decline over the past decade in the total number of Bachelor’s degrees in the mathematical sciences earned by Black non-Hispanic students is almost entirely the result of a decline in the number of Black women earning math degrees.

The following graph shows what has been happening to the number of non-resident aliens majoring in the mathematical sciences in the United States.

Observations: The most striking feature of Graph 1 is the sharp decline in the number of math majors during the 1990s, followed by an equally sharp rebound by men that is not reflected in nearly so steep a recovery for women. There are several factors that have been at play. One is that the college-age population decreased during the 1990s, bottoming out around 1997. Another is that the second half of the ‘90s was a time of economic prosperity. Unemployment rose in the early 2000s, peaking in 2003. By the fall of 2008, it was clear that we were heading into an economic disaster. As I showed in A Benefit of High Unemployment (October 2010), the number of students choosing to enter mathematically intensive majors is highly correlated with the economic situation: the harder students expect it to be to find a job when they graduate, the more likely it is that they will choose a mathematically intensive major. The year 2000 also saw the collapse of the bubble and with it a huge decline in the number of students heading into computer science. It is not unreasonable to assume that many of them switched to mathematics.

During the last decade, the growth in mathematics majors has occurred entirely in universities with graduate programs in mathematics and has decreased sharply at colleges only offering a bachelor’s degree in mathematics (see Good News from CBMS, November 2011). It is these undergraduate institutions that have traditionally had the largest representations of women and students from under-represented groups. I believe that this has contributed to both the increasing discrepancy between men and women and to the decline in the number of Black mathematics majors. Less clear is what has caused the decline in mathematics majors at undergraduate colleges.

It is noteworthy that the growth in the number of Hispanic mathematics majors has been very strong throughout the past two decades. This reflects the tremendous growth in the number of Bachelor’s degrees earned by Hispanic students: from 33,000 in 1990 to 130,000 in 2009, from 3% of all Bachelor’s degrees in 1990 up to 10% in 2009.

Graph 5 is interesting because it illustrates very clearly the post-911 effect on student enrollments from outside the United States. The first two years of sharply stricter student visa requirements, 2002 and 2003, were reflected in a sharp drop in the number of non-resident alien math majors 2007. The good news is that by 2009 we had almost fully recovered. But the most interesting feature of Graph 5 is that there has not been a widening gap between the number of non-resident men and the number of non-resident alien women choosing to go into mathematics. This may be entirely explainable by the fact that women are still catching up to men in this category (see Graph 6).

Note: I have not included data for Native American/Alaskan Native students because the numbers are so small, usually between 60 and 80 combined per year. As a result, there is very high year-to-year variation, obscuring clear trends. The NCES tables do not separate Mathematics from Statistics majors. Mathematics Education majors are not separated from other Education majors, so are not included.


  1. It seems to be the case from my limited observations that the majority of math education majors are female, so their exclusion from the data examined above might have a significant impact on the conclusions.

    1. Unfortunately, NCES does not report the number of Math Education majors because it does not break down Education majors by area. CBMS does, but the most recent CBMS data are from 2005. In that year, there were 3369 Math Ed majors (as tracked by the Math Department), of whom 60% (2028) were women.