Gender differences in science, technology, engineering, mathematics and computer science (STEM) programs at university

View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Archived

This page has been archived on the Web.

by Darcy Hango

[Release from The Daily] [Full article in PDF]

Skip to text

• Overview of the study
• Introduction
• Most STEM university graduates are men
• Labour market outcomes not always the same for all STEM graduates
• Mathematical ability in high school and program choice at university
• Young men with lower marks in high school were more likely to choose a STEM program than young women with higher marks
• Conclusion
• Related material for this article

Text begins

Introduction

In Canada and elsewhere in the world,^Note1 encouraging university students to choose a program in science, technology, engineering, mathematics and computer science (known as 'STEM' programs) has long been a defining outcome of national innovation strategies.

The focus on STEM degrees can be related to how they would contribute to a country’s competitiveness and economic prosperity. Given the scope and nature of their labour market activities–for instance, by engaging in research and development or other knowledge-enhancing activities–STEM graduates are considered key inputs of the national innovation system.^Note2 At the individual level, previous reports have also suggested that a STEM degree would lead to better labour market conditions and higher earnings, especially for those with a background in engineering and computer science.^Note3

That said, and despite the advances made in recent years,^Note4 women remain less likely to choose a career in STEM areas, and more particularly in engineering, mathematics and computer science. This stands in contrast to nearly all other fields of study, where women now represent the vast majority of graduates—especially health and social science programs. Why are women staying away from STEM programs?

This article examines gender differences in program choice and graduation. First, it profiles differences between men and women among recent STEM university graduates aged 25 to 34, using data from the 2011 National Household Survey (NHS). It also examines whether STEM university graduates have better labour market outcomes than non-STEM university graduates.

In the second part of this article, the factors associated with university program choice are examined for both men and women, using longitudinal data from the Youth in Transition Survey–Program for International Student Assessment (YITS–PISA). Of particular interest is the relationship between mathematical ability at age 15 and choosing a university STEM program.^Note5

Most STEM university graduates are men

Young adults are more and more likely to obtain a university degree, regardless of gender. However, women have made gains much more rapidly than men in recent years. In 1991, the proportion of men and women aged 25 to 34 with a university degree was about the same (at 16%); by 2011, the proportion had increased to 37% among women and 27% among men. As a result, 59% of all university graduates aged 25 to 34 in 2011 were female.^Note6

The share of female university graduates increased in nearly all fields of study, including STEM programs. Hence, in 2011, 39% of STEM university graduates aged 25 to 34 were female, compared to 23% of STEM graduates aged 55 to 64.  However, women represented 66% of all non-STEM graduates aged 25 to 34 in 2011 (a proportion that was in the 80% range in health professions and related programs, and in education programs). Women were thus proportionately under-represented among STEM graduates, at least compared with other fields.

Furthermore, women with a STEM degree were particularly concentrated in science and technology programs. Specifically, women accounted for 59% of graduates who had a university degree in science and technology, but accounted for 23% of graduates aged 25 to 34 with a university degree in engineering, and 30% of those with a degree in mathematics and computer science. Thus, 39% of the 132,500 women aged 25 to 34 who had a STEM degree had a background in engineering, mathematics or computer science (Chart 1).^Note7 In comparison, 72% of the 206,600 STEM-educated men had a background in these fields (47% in engineering and 25% in mathematics and computer science).

Description for Chart 1

Labour market outcomes not always the same for all STEM graduates

The better labour market outcomes of young graduates with a STEM degree are often used as motivational factors to encourage students to choose a STEM degree.^Note8 Labour market conditions, however, may vary across gender, type of program, and indicators of labour market performance.

One of these indicators is the unemployment rate. In 2011, the unemployment rate for men aged 25 to 34 with a STEM university degree was 4.7%, compared with 5.5% for those with a non-STEM university degree. For women, it was the opposite: the unemployment rate of women with a STEM university degree was 7.0%, compared with 5.7% among those with a non-STEM degree (Table 1).

Table 1
Labour market outcomes of university graduates aged 25 to 34, by sex and major field of study, 2011
Table summary
This table displays the results of labour market outcomes of university graduates aged 25 to 34 total, women and men, calculated using percentage and dollars units of measure (appearing as column headers).

	Total	Women	Men
	percentage
Unemployment
Total STEM	5.5	7.0	4.7
Science	6.2	6.6	5.8
Technology (except engineering technology)	5.1	3.4	6.7
Engineering	4.9	7.1	4.3
Mathematics and computer science	5.4	8.5	4.2
Non STEM	5.6	5.7	5.5
Skill mismatchNote 1
Total STEM	14.3	18.3	11.8
Science	18.0	18.9	16.8
Technology (except engineering technology)	22.2	20.5	23.5
Engineering	10.6	13.5	9.8
Mathematics and computer science	13.6	22.4	10.1
Non STEM	19.7	18.5	22.2
	dollars
Median wages and salariesNote 2
Total STEM	59,300	53,200	62,300
Science	51,700	49,100	55,300
Technology (except engineering technology)	51,700	49,700	54,600
Engineering	65,200	61,100	66,300
Mathematics and computer science	59,300	54,900	60,800
Non STEM	52,200	50,200	56,000
Note 1. Percentage of persons working in occupations requiring a high school education or less. Includes persons who were employed during the NHS reference week, or weren't employed but last worked in 2010 or 2011. Similar differences were found between groups when the sample was restricted to employees working full year, full-time in 2010. Note 2. Gross wages and salaries before deduction, employees working full year, full time in 2010. Note: STEM includes science, technology, engineering, mathematics and computer science. Source: Statistics Canada, National Household Survey, 2011.

Men with a STEM university degree were also less likely than non-STEM graduates to be working in occupations requiring a high school education or less—a form of 'skills mismatch'.^Note9 Among men aged 25 to 34 with a STEM university degree, 12% were in occupations generally requiring a high school diploma or less, compared with 22% among those with a non-STEM university degree. Among women, 18% of STEM graduates were in occupations requiring a high school diploma or less, a percentage that was very similar to that of women with a non-STEM university degree.^Note10

Examining skills mismatch within STEM groups is also important, as it may help explain the larger proportion of female STEM university graduates that were in occupations requiring a high school diploma or less, compared with their male counterparts. In 2011, 18% of university science graduates were employed in occupations requiring a high school diploma or less, compared with 11% of those with an engineering degree. Therefore, the relatively higher mismatch rate of female STEM university graduates can be explained, in part, by their larger share in programs with higher mismatch rates (for example, science) and their lower share in programs like engineering.

Similarly, the relatively lower mismatch rate of non-STEM female university graduates (19%) compared with their male counterparts (22%) can be explained, at least in part, by the fact that many women graduated from non-STEM programs in which mismatch rates are relatively low—for example, education (10%) and health professions and related programs (10%).

That STEM programs would also lead to better-paying jobs is often used as a selling point to encourage more youth to pursue such programs.^Note11 According to the NHS, this was true, but even more so for men: in 2010, men aged 25 to 34 who had a STEM university degree and were employed on a full-time and full-year basis earned a median of $62,300, versus $56,000 for those who graduated from non-STEM programs.^Note12 The gap was smaller among university-educated women: STEM graduates employed full-time, full-year, had median salaries and wages of $53,200, compared with $50,200 for non-STEM graduates.

If STEM-educated men are earning more than non-STEM-educated men, however, it is largely because they are concentrated in engineering, where earnings are significantly higher. Men with a degree in engineering earned $66,300, compared with men with a degree in science and technology, who earned about $55,000.

Conversely, women with a degree in a science or technology field—where they were more concentrated—earned just under $50,000, that is, about the same as non-STEM graduates ($50,200) and significantly less than women with a degree in engineering ($61,100).

Mathematical ability in high school and program choice at university

Past studies have attempted to understand why women were less likely than men to enter and complete a STEM university program. For some, gender differences in mathematical ability—a prerequisite to success in these fields of study—would explain why fewer women are found in such programs.^Note13 For others, the gender gap is rather a by-product of differences in values and preferences.^Note14 Which of these assumptions is supported by the data?

Longitudinal data from the Youth in Transition Survey and the Program for International Student Assessment (YITS–PISA) can be used to better understand the relationship between mathematical ability and the choice of a STEM university program among men and women. The YITS–PISA interviewed a sample of youth aged 15 in 2000, and re-interviewed them every two years until they were 25. With this data, some key educational characteristics measured during adolescence (including PISA tests conducted when students were aged 15) can thus be linked with other measures collected in subsequent years, including program choice in university (see Data sources, methods and definitions).

Overall, 40% of men in the YITS-PISA sample who went to university in early adulthood opted for a STEM program as their first choice in university, while only 20% of women did so (Chart 2).^Note15 In fact, young women were much more likely to choose a first program in social sciences (50%, compared to 32% of men among those who went to university).

Description for Chart 2

Those who chose a STEM program had higher PISA test scores when they were in their mid-teens than those who chose other programs (Table 2). For example, women who chose a STEM program had obtained an average mathematics PISA score of 588 at age 15, compared with 565 among those headed for social sciences. Among men, those who chose a STEM program had achieved a score of 597 on average, compared to 585 among those who chose social sciences. Males also tended to have higher PISA scores than women on average (589 versus 569).^Note16

Table 2
Average mathematics PISA scores at age 15 among YITS-PISA respondents who attended university, by first program choice and sex
Table summary
This table displays the results of average mathematics pisa scores at age 15 among yits-pisa respondents who attended university total, women and men, calculated using average units of measure (appearing as column headers).

	Total	Women	Men
	average
Total	577.8	569.3	588.5Note ***
STEM	593.4	587.5	596.7
Social sciences	571.8	565.3	584.7Note **
Business	567.8	555.4	583.2Note *
Health	571.9	570.3	576.6
Other	570.4	572.2	569.1
Weighted sample size	132,350	73,567	58,783
Note * p   < 0.1; ** p < 0.05; *** p < 0.01 indicates significance of gender difference Note: STEM includes science, technology, engineering, mathematics and computer science. Sources: Statistics Canada and Human Resources and Skills Development Canada, Youth in Transition Survey (YITS); Organisation for Economic Co-operation and Development, Programme for International Student Assessment (PISA), 2000 to 2010.

In order to control for mathematical proficiency levels, the sample can be divided between those who had “higher” PISA scores at age 15 (mathematics proficiency levels 4, 5 and 6—a PISA score of 545 and over) and “lower” PISA scores (levels 1, 2 and 3). Those who had higher mathematics proficiency levels were more likely to choose STEM programs than their counterparts with lower proficiency levels. For example, 23% of women with higher mathematics PISA scores entered a STEM program, compared with 15% of those with lower scores (Table 3).^Note17

Table 3
First university program choice of YITS-PISA respondents who attended university, by category of PISA scores and sex
Table summary
This table displays the results of first university program choice of yits-pisa respondents who attended university stem, social sciences, business, health and other, calculated using percentage units of measure (appearing as column headers).

	STEM	Social sciences	Business	Health	Other
	percentage
Total	30.8	42.0	13.8	9.3	4.3
Higher mathematics PISA score
Women	23.2	48.3	12.6	12.4	3.6
Men	45.7	31.3	13.3	5.1	4.7Note E: Use with caution
Lower mathematics PISA score
Women	15.3	53.9	15.8	12.2	2.9Note E: Use with caution
Men	38.5	32.2	15.4	Note F: too unreliable to be published	Note F: too unreliable to be published
E use with caution F too unreliable to be published Notes: Higher PISA score is defined as 4th proficiency level and above. Lower PISA score is defined as 3rd proficiency level or below. STEM includes science, technology, engineering, mathematics and computer science. Sources: Statistics Canada and Human Resources and Skills Development Canada, Youth in Transition Survey (YITS); Organisation for Economic Co-operation and Development, Programme for International Student Assessment (PISA), 2000 to 2010.

That said, women with higher scores were less likely to choose a STEM university program than men with lower scores (23% versus 39%). In fact, among women, social sciences were preferred by most, regardless of mathematical proficiency. Men, in contrast, were always more likely to choose a STEM program, even among those who had a lower proficiency in mathematics.

Young men with lower marks in high school were more likely to choose a STEM program than young women with higher marks

When high school marks in mathematics were examined instead of PISA scores (Chart 3), similar results were found:

• Students with higher mathematics marks were generally more likely to choose a STEM university program. For instance, more than 40% of women with marks in the 90% to 100% range chose a STEM program, compared with 11% of women with marks under 80%;
• Men were more likely to opt for a STEM program, even in lower-marks categories. More than 30% of men who had marks under 80% and attended university chose a STEM program, compared with about 10% of women in the same category;
• Young women with higher mathematics marks in high school (at least 90% in grade 9 or 10) were less likely to opt for a STEM university program than men with marks in the 80% to 89% range.

Description for Chart 3

YITS also contains self-perceived measures of mathematical ability in high school: “excellent”, “very good”, “good”, “fair”, and “poor”. With respect to their mathematical ability, young men typically had a better opinion of themselves as 50% perceived their ability to be “very good” or excellent”, compared with 37% of young women. Again, those who perceived their mathematical skills more positively were more likely to choose a STEM program later at university.

Among youth who attended university and perceived their mathematics skills to be good, 36% of males opted for a STEM program, compared with 15% of females. Furthermore, 66% of males who perceived their skills to be excellent chose a STEM program, compared with 47% of females. Therefore, there was still a gender difference, even among those who had a better perception of their mathematical ability.

In summary, students with stronger PISA scores, higher marks, and better perceptions of their academic abilities in high school were more likely to opt for a STEM university program. Nevertheless, young men were always more inclined to choose STEM programs than young women for a given level of mathematical ability (regardless of measure). Are such findings the same when all factors that could influence program choice are taken into account (including immigration status, parental influence variables, and reading scores)?

By and large, the answer is "yes". When all factors except marks and self-perceived mathematical ability were accounted for in a model including all YITS-PISA respondents who attended university, both men with higher and lower PISA mathematics scores were still more likely to enter STEM fields than women with higher scores (Table 4, model 2). Specifically, men with high scores were most likely to choose such a STEM program by a margin of 22 percentage points over women with equally high scores. In addition, men with lower PISA scores were also more likely to choose a STEM program than women with higher scores, by a margin of 15 percentage points. Conversely, men were always less likely to choose a program in social sciences, by a margin of 12 percentage points among men with lower PISA scores and 16 percentage points among men with higher PISA scores.

The inclusion of marks and self-perceived mathematical ability (Table 4 model 3)—both positively correlated with the choice of a STEM program^Note18—reduced the effect of gender on the probability of choosing a STEM program, but only slightly (from 22 percentage points to 18 percentage points among students with higher PISA scores).

Table 4
Marginal effect of choosing a STEM university program across mathematics PISA score results and gender, based on three multinomial logit model scenarios
Table summary
This table displays the results of marginal effect of choosing a stem university program across mathematics pisa score results and gender stem, social sciences, business, health and other, calculated using average marginal effect units of measure (appearing as column headers).

	STEM	Social sciences	Business	Health	Other
	average marginal effect
Model 1 PISA score interacted with sex only
PISA score by sex (ref: higher-women)
Higher - Men	0.224Note for table 4 ***	-0.170Note for table 4 ***	0.007	-0.073Note for table 4 ***	0.011
Lower - Women	-0.080Note for table 4 ***	0.056	0.032	-0.001	-0.007
Lower - Men	0.153Note for table 4 ***	-0.161Note for table 4 ***	0.029	-0.057Note for table 4 **	0.036
Model 2 PISA score interacted with sex, plus demographic, reading scores and parental controls
PISA score by sex (ref: higher-women)
Higher - Men	0.217Note for table 4 ***	-0.156Note for table 4 ***	0.007	-0.077Note for table 4 ***	0.009
Lower - Women	-0.073Note for table 4 **	0.093Note for table 4 *	-0.002	-0.009	-0.010
Lower - Men	0.145Note for table 4 ***	-0.117Note for table 4 **	-0.006	-0.057	0.034
Model 3 All variables in model 2, plus mathematical marks and self-assessed mathematical ability
PISA score by sex (ref: higher-women)
Higher - Men	0.178Note for table 4 ***	-0.126Note for table 4 ***	0.004	-0.074Note for table 4 ***	0.018
Lower - Women	-0.010	0.012	0.007	0.006	-0.015
Lower - Men	0.194Note for table 4 ***	-0.159Note for table 4 ***	-0.010	-0.053	0.028
Weighted sample size	132,350
Note * p   < 0.1; ** p < 0.05; *** p < 0.01 indicates significant difference from reference (ref) category Note: STEM includes science, technology, engineering, mathematics and computer science. Sources: Statistics Canada and Human Resources and Skills Development Canada, Youth in Transition Survey (YITS); Organisation for Economic Co-operation and Development, Programme for International Student Assessment (PISA), 2000 to 2010.

As a result, if more men are found in STEM programs, it is not because they have better PISA scores than women. In fact, even when all measures of mathematical ability are combined in a model, gender differences remain significant. This suggests that the gender difference in the selection of a STEM program at university is due to other, unobserved, factors that go beyond academic achievement, parental interactions and influence, and immigration status.

Conclusion

Over the past few decades, women have made significant advances in university participation, including program areas that had previously been more populated by men. One area, however, remains male-dominated: science, technology, engineering and mathematical (STEM) degrees. And among women who choose to pursue a degree in STEM, most do so in biology or science programs, resulting in even fewer women in engineering, computer science and mathematics programs. These choices have consequences, as fields of study such as engineering and computer science lead, on average, to better outcomes in the labour market in terms of employment, job match and earnings.

For some, aptitude for a particular subject is a factor in university program choice. Although mathematical ability plays a role, it does not explain gender differences in STEM choices. Young women with a high level of mathematical ability are significantly less likely to enter STEM fields than young men, even young men with a lower level of mathematical ability. This suggests that the gender gap in STEM-related programs is due to other factors. Other possible explanations might include differences in labour market expectations including family and work balance, differences in motivation and interest, and other influences.^Note19

Darcy Hango is a researcher in the Centre for Education Statistics at Statistics Canada. The author would like to thank Martin Turcotte of the Labour Statistics Division and Sarah-Jane Ferguson of the Centre for Education Statistics.

---

Notes

1. See OECD (2012).
2. See Beckstead and Gellatly (2006).
3. See Conference Board of Canada (2013).
4. See Ferguson and Zhao (2013).
5. A longer and more detailed version of the analysis presented in this section is available in the report Ability in Mathematics and Science at Age 15 and Program Choice in University: Differences by Gender, by Hango (2013).
6. Source: 2011 National Household Survey and 1991 Census of Population.
7. Because university-educated immigrants were more likely to hold a STEM degree, the proportion of women choosing STEM programs was even lower among the Canadian-born. See Proportion of immigrants who are STEM graduates.
8. See Conference Board of Canada (2013).
9. In this section, the people who were not working during the reference week, but who were employed at some time in 2011 or 2010, are included in the universe. For those who did not work during the reference week, the skill match corresponds to the most recent occupation held by the respondent.
10. Even among those who were not working in occupations requiring a high school education or less, not all STEM graduates worked in STEM fields. According to Beckstead and Gellatly (2006), individuals holding “science and engineering degrees” outnumbered individuals employed in science and engineering occupations by a ratio of 2.2 to 1 in the early 2000s.
11. See Conference Board of Canada (2013).
12. Earnings are defined as gross wages and salaries before deductions of paid employees (self-employed workers are excluded).
13. In 2005, Lawrence Summers—then-President of Harvard University—suggested that males were more adept at pursuing careers in fields requiring advanced mathematical ability. See Summers (2005).
14. See OECD (2012); Kane and Mertz (2011).
15. Note that the definition of STEM programs is slightly different in the linked YITS–PISA data than in the National Household Survey (see Data sources, methods and definitions).
16. This result was consistent with other studies comparing all males and females 15 years of age (and not only those entering a university program). See Bussière et al. (2007).
17. For men, the difference between those with lower skill levels and higher skill levels was statistically significant at the p < 0.10 level only.
18. This is likely the case because more objective measurements of mathematical proficiency might be positively related to school marks at age 15 and self-perceived mathematical ability. In other words, the positive and direct relationship between mathematical proficiency (as measured by PISA scores) and the choice of STEM as a university program may be mediated by both math marks and self-perceived or subjective mathematical ability.
19. Some researchers have suggested that factors related to differences in preferences and labour market expectations could explain part of the gender gap in program choice (Turner and Bowen 1999; Diekman et al. 2010). Others have found that social class, family background and cultural capital in the home were also very important factors (Van de Werfhorst et al. 2003). Finally, a recent OECD (2012) study suggested that “gender disparities in subjects chosen appear to be related more to student attitudes (such as motivation and interest) towards a particular subject rather than to ability and performance at school” (p. 104).