In the June issue of ACM Inroads , we consider the phenomenal growth of computer science (CS) in both CS undergraduate degree programs and CS courses at doctoral-granting and non-doctoral-granting units.1 This article examines the impact of the undergraduate enrollment surge on diversity (i.e., women and underrepresented minorities) using two existing data sets (i.e., the CRA Taulbee Survey  and IPEDS ) and data collected from the CRA Enrollment Survey . We also highlight relationships discovered from the CRA Enrollment Survey between actions taken by units to manage the surge and their impact on diversity.
As discussed in , the goal of the CRA Enrollment Survey was to measure, assess, and understand enrollment trends and their impact on computer science units, including the impact of the enrollment surge on diversity (i.e., women and underrepresented minorities). A positive consequence of the current undergraduate enrollment surge is a significant increase in the number of women and underrepresented minority (URM) students in computer science, both in courses and as majors. In addition, there is also some good news regarding the percentage of women and URM students in aggregate; the good news, however, is not universal across all units that responded to the CRA Enrollment Survey.
The CRA Enrollment Survey asked units for women and URM student enrollment data, including data for four representative courses at five-year intervals beginning with 2005; the course level data is examined in this article. The survey also asked each unit's perception about trends in the unit's recruitment and retention of students from underrepresented groups. We will discuss the relationship between this perception and the hard data collected about course enrollments in this article.
The enrollment data results discussed in this article show that the median percentage of female students in the courses surveyed has increased overall at both doctoral- and non-doctoral-granting units. While some may consider the percentage increase to be significant, others would correctly note that we are still a long way from gender parity. The results presented in the rest of this article also show that the median percentage of URM students in our four representative courses is increasing; these data do not include students at minority-serving institutions (MSIs), as MSI data skew our understanding of non-MSI institutions. We note that, at doctoral-granting units, the percentage increase for URM students is larger in the intro course for nonmajors than in the intro course for majors.
While there is some good news here, the data do suggest a shrinking pipeline may continue to exist for both female and URM students (i.e., in the course data provided, the representation of these students decreased from the intro through mid-level through upper-level courses). Also, as mentioned in our opening paragraph, not all units are seeing an increase in the percentage of women and URM students participating in computer science. In short, much work remains for the CS community in the United States to look like our population.
Degrees Awarded by Gender and Ethnicity
To provide context for the CRA Enrollment Survey data, we look at degrees awarded from the CRA Taulbee Survey  and IPEDS  data for the period of 2006–2015 (which is the period of the current enrollment surge). Taulbee graduation rates for doctoral-granting units show that female students comprised 14% of the computer science bachelor's degrees in 2006. This percentage declined over subsequent years until it reached a low point just above 11% in 2009. Since 2009, we have seen a slow but steady increase, with female students earning almost 18% of the 2016 computer science bachelor's degrees awarded.
For URMs, we consider Black/African American, Hispanic/Latino, and other underrepresented groups (American Indian, Alaska Native plus, after 2007, Native Hawaiian, and Other Pacific Islander). The omission of the Asian and multiracial ("two or more races") categories is designed to avoid considering groups that are not necessarily underrepresented in computing. We begin with Taulbee graduation rates for doctoral-granting units. In 2006, URM students comprised 9% of the CS graduating students at doctoral-granting units. We saw the percentage of URM students earning bachelor's degrees in CS increase to 12% in 2008, which coincided with the shift to new categories for race and ethnicity, a decrease to 9% in 2010 (the low point), and a subsequent gradual increase to 11% in 2016.
IPEDS provides data for degrees awarded at both doctoral-granting and non-doctoral-granting units. We selected data for two Classification of Instructional Programs (CIP) codes: Computer and Information Sciences, General (11.0101) and Computer Science (11.0701). These two categories are typically where CS units report their degrees awarded. We use "CS" to represent reported degrees awarded collectively under these CIP codes.
Figure 1 shows IPEDS CS degrees awarded for four groups—women, Black/African American, Hispanic/Latino, and other underrepresented groups (American Indian, Alaska Natives, Native Hawaiian, and Other Pacific Islander)—as a percentage of total degrees awarded for doctoral- and non-doctoral-granting units. Figure 1 begins with data from 2009, which is just after the low point for the percentage of computer science bachelor's degrees awarded to women. IPEDS data for 2016 is not yet available.
The IPEDS data presented in Figure 1 for women show a similar trend to the Taulbee data discussed previously. According to IPEDS data for doctoral-granting units, women comprised 13.6% of 2006 computer science bachelor's degrees, 11.3% of 2008 degrees (the low point), and 15.3% of 2015 degrees. For non-doctoral-granting units, the percentage of women has been higher overall, with 20.5% of computer science bachelor's degrees awarded to women in 2006, declining to 15.5% by 2011 (the low point), and rising slightly to 16.6% in 2015.
The IPEDS data for URMs at doctoral-granting units are a bit different from the Taulbee data, likely because of the difference in the definition of doctoral-granting,2 but the general trend shows 11–12% of computer science bachelor's degrees were awarded to URMs from 2006–2011; in 2015, URM students comprised 13% of the computer science bachelor's degrees at doctoral-granting units. In non-doctoral-granting units, IPEDS data show a higher percentage of URM students, with almost 20% in 2006, but remaining close to 18% for most of the 2007–2015 period. A decline in degrees awarded to Black/African Americans was offset by an increase in degrees awarded to Hispanic/Latinos. We note that the U.S. Hispanic population grew by 28% from 2005–2015, while the Black/African-American population only grew by 11% and the overall population grew 9% .
The percentage of degrees awarded to Black/African Americans declined from 2009 to 2015 in both doctoral- and non-doctoral-granting units. The percentage of degrees awarded to women increased in doctoral-granting units, while the percentage dropped and then recovered somewhat in non-doctoral-granting units.
Figure 1(a) shows that there was a growth in degrees awarded from 2009 to 2015 for three of the four groups (all but Black/African American) in doctoral-granting units. Figure 1(b) shows that there was a growth in degrees awarded for only two of the four groups—Hispanic/Latino and other underrepresented groups—in non-doctoral-granting units. The percentage of degrees awarded to Black/African Americans declined from 2009 to 2015 in both doctoral- and non-doctoral-granting units. The percentage of degrees awarded to women increased in doctoral-granting units, while the percentage dropped and then recovered somewhat in non-doctoral-granting units.
Due to the overall increase in CS degrees awarded, the number of degrees awarded to all four groups has grown from 2009 to 2015. Table 1 shows the increase in the number of IPEDS CS degrees awarded from 2009 to 2015, separated by non-doctoral and doctoral-granting units. For example, almost 4,000 more CS degrees were awarded in 2015 than in 2009 at non-doctoral-granting units, which represents a 48% increase. The growth in degrees awarded to women (except in non-doctoral-granting units), Hispanic/Latino, and Other Underrepresented groups (American Indian, Alaska Native, Native Hawaiian, and Other Pacific Islanders) has outpaced the overall growth across both doctoral and non-doctoral-granting units. While the number of IPEDS CS degrees awarded to Black/African Americans has increased, the rate of increase is less than one-half of the overall increase.
Representation of Women at Different Stages of the Curriculum
The previous section looked at degrees awarded for diverse students, from the CRA Taulbee Survey  and IPEDS  data for the period of 2006–2015. This section and the next examine the diversity of students enrolled in CS courses.
The CRA Enrollment Survey shows that the percentage of women has grown in all three of the CS major courses surveyed from 2005 to 2015 for both doctoral-granting and non-doctoral-granting units (Figure 2). For doctoral granting units (see Figure 2(a)), most of this growth occurred between 2010 and 2015; for non-doctoral-granting units (see Figure 2(b)), this growth occurred between 2010 and 2015 in only the mid- and upper-level courses. Doctoral-granting units also showed consistent growth from 2005 to 2015 for their nonmajor intro course. Unfortunately, there is not sufficient data on the nonmajor intro course for non-doctoral-granting units. That is, only 20 to 22 non-doctoral-granting units of the ∼700 provided course diversity data in the CRA Enrollment Survey for each of the 2005, 2010, and 2015 periods surveyed, and only 11 of these units provided diversity data for a separate nonmajor intro course.
For the non-doctoral-granting units that provided course enrollment data, the median percentage of women was higher than in doctoral-granting units. We also note that the median percentage of women at doctoral-granting units in the upper-level course in 2015 (17%) (Figure 2(a)) is close to the Taulbee percentage of women who graduated in 2015 (16%). For non-doctoral-granting units (Figure 2(b)), our course enrollment data at the mid- and upper-level shows a larger increase in the percent-age of women from 2010 to 2015 than IPEDS data (see Figure 1(b)), which could be due to either our small sample size or a recent pipeline increase that has yet to be realized as CS degrees.
From Figure 2(a), the largest percentage of women in doctoral-granting units can be found in the nonmajor intro course followed by the major intro course. We also observe that percentages drop from the intro major to the mid-level to the upper-level course in both doctoral-granting and non-doctoral-granting units. This could suggest an issue of retaining women into higher-level courses. We stress, however, that this question regarding retention requires further study. For example, one possible explanation could be that the apparent decrease is simply a byproduct of growth. That is, students in upper-level classes entered the program at a time when the percentage of women was lower. Furthermore, the data are consistent with this explanation—that is, the percentage of women in the upper-level course in 2015 is between the percentage of women in the 2010 and 2015 mid-level courses. Another hypothesis is that the larger percentages in the lower-level courses reflect an increase in female nonmajors in those courses. Of course, if the trend reflects a leaky pipeline, where women are dropping out at a higher rate than men as they advance, a concern about female student retention is valid.
For doctoral-granting units, data also revealed more rapid growth in the percentage of women in private schools, as compared to public schools, in all four courses surveyed, as shown in Figure 3. For example, consider the nonmajor intro course and the mid-level course. In both cases, public and private schools had similar percentages of women in 2005 and very different percentages of women in 2015.
IPEDS data on CS degrees awarded for public and private institutions, as shown in Figure 4, corroborate this observation. The percentage of female degree earners was 14% for public schools and 20% for private schools in 2015. While private institutions have a larger percentage of women graduating, public institutions still produce more women overall (approximately 2,600 vs. 1,800 in 2015). A study of what private schools have been doing, or not doing, compared to public schools could be useful to increasing the number of women in CS.
Of course, if the trend reflects a leaky pipeline, where women are dropping out at a higher rate than men as they advance, a concern about female student retention is valid.
While Figures 2 and 3 show the change in the median percentage of women in our representative courses across units, Figure 5 shows how much women's representation in the chosen upper-level course varies by unit among the doctoral-granting units. Across units, the percentage of women enrolled in the upper-level course ranged from a low of 0 to 2% (5 units in 2010, but none in 2015) to a high of 42 to 44% (one unit in 2015). In short, there is substantial variation across departments, but the distribution shifted up between 2010 and 2015.
Lastly, we note that approximately 50% of the units that responded to the CRA Enrollment Survey perceive the percentage of women in their unit is increasing. While this perception exists in both doctoral- and non-doctoral-granting units, doctoral-granting units were more likely to state that the percentage of females in their unit is increasing significantly. We understand that perception may not reflect reality but, in this case, hard data on the percentage of women enrolled in the introductory course for majors are consistent with the unit's self-reported perceived change in the percentage of women entering the major. Based on data from the CRA Enrollment Survey, the proportion of women in the intro majors course at doctoral-granting units increased by an average of 0.4 percentage points from 2010 to 2015 for units that self-reported the proportion of women was level or declining (n=18), by 5.1 percentage points for units reporting the proportion of women as increasing (n=14), and by 7.3 percentage points for units reporting the proportion of women as significantly increasing (n=8).
Representation of Underrepresented Minorities at Different Stages of the Curriculum
Regarding the perception of underrepresented minorities in CS, approximately 50% of doctoral- and 30% of non-doctoral-granting units that responded to the CRA Enrollment Survey perceive that the percentage of URMs in their unit is level, while approximately 20% of doctoral- and 40% of non-doctoral-granting units perceive that the percentage of URMs in their unit is increasing. Also, approximately 10% of doctoral- and 15% of non-doctoral-granting units perceive that the percentage of URMs in their unit is decreasing, while the remaining units (approximately 20% for doctoral and 15% for non-doctoral) stated either they don't know the percentage of URMs in their unit or the numbers are too small to determine. Again, while perception may not reflect reality, there is some evidence that this perception is also accurate. The 32 non-MSI units that reported a percentage of URMs enrolled in the introductory course for majors in both 2010 and 2015 is consistent with the self-reported perceived change in the proportion of URMs entering the major. Based on data from the CRA Enrollment Survey, the proportion of URMs in the intro majors course at doctoral-granting units decreased by 2.7 percentage points from 2010 to 2015 for units that self-reported the proportion of URMs was declining (n=5), increased by 1.4 percentage points for units reporting the proportion of URMs was level (n=19), and increased by 2.7 percentage points for units reporting the proportion of URMs was increasing (n=8).
Though the percentage change for URMs is small, the number of URM students has shown remarkable growth since 2005, mostly between 2010 and 2015. As shown in Figure 6, the total number of minority students reported in the course enrollment data of doctoral-granting units has grown from 708 to 808 to 1620 in the intro required course, from 299 to 443 to 992 in the mid-level course, and from 208 to 322 to 606 in the upper-level course. While this growth in the number of URM students is occurring at all units, we note that a significant number of URM students are added by the four minority-serving institutions (MSIs) in our sample. Because these MSIs have a very different profile of students, they could bias our data and are, unless noted, removed from the data and analysis of the following percentages and correlations.
Figure 7 shows the median percentage of URM students for each course, for both doctoral-granting and non-doctoral-granting units (not including MSIs). While the percentage of URM students does not show the consistently increasing trend seen in the percentage of women students (see Figure 2), we note the percentage of URM students has increased when comparing 2005 to 2015 data for both doctoral- and non-doctoral-granting units, which is consistent with the Taulbee and IPEDS data on degrees awarded. While our enrollment data do not consider different URM populations, the IPEDS data appear to show a decline in Black/African American students which is offset by a rise in Hispanic/Latino students.
Although the percentage of URM students seems to be growing, we note that the percentages are still extremely small (e.g., only 9% URM students in the 2015 upper-level course data for doctoral-granting units). In fact, in 2005, the median percentage of URM students reported by non-doctoral-granting units for the upper-level course was 0%. Clearly, there is much room for growth among the URM student population, especially when one considers that 34% of college students are URM in the United States . Like the trends we see in the course enrollment data for women, we typically see a decrease in the URM student representation from the intro through mid-level through upper-level courses. As previously mentioned, further study is needed to determine whether a leaky pipeline exists.
Lastly, we look at IPEDS CS degrees awarded to URMs at public and private institutions, including MSI institutions. Figure 8 shows that the IPEDS CS rate for URMs (Black/African American, Hispanic/Latino, and Other Underrepresented) from 2009 to 2015 has decreased slightly at private institutions and increased slightly at public institutions. The figure also shows that private institutions have a greater representation of URMs in CS than public institutions, like the representation of women in CS discussed previously (Figure 4).
The Relationships Between Unit Actions and Diversity Growth
The CRA Enrollment Survey included several questions about actions (e.g., train faculty in scalable class management) that units were taking in response to the surge. In this section, we highlight a few statistically significant correlations that relate growth in female and URM students to unit responses (a composite3 of several different responses).
Units with an increase in minors have an increase in the percentage of female students in mid- and upper-level courses.
- Units that explicitly chose actions to assist with diversity goals have a higher percentage of female and URM students. We observed significant positive correlations between units that chose actions to assist with diversity goals and the percentage of female majors in the unit for doctoral-granting units (per Taulbee 2015, r=.19, n=113, p<.05), and with the percent of women in the intro majors course at non-doctoral-granting units (r=.43, n=22, p<.05). A similar correlation was found for URM students. Non-MSI doctoral-granting units showed a statistically significant correlation between units that chose actions to assist with diversity goals and the increase in the percentage of URM students from 2010 to 2015 in the intro for majors course (r=.47, n=36, p<.001) and mid-level course (r=.37, n=38, p<.05). Of course, units choosing actions to assist with diversity goals are probably making many other decisions with diversity goals in mind. Improved diversity does not come from a single action but from a series of them (e.g., ).
- Units with an increase in minors have an increase in the percentage of female students in mid- and upper-level courses. We observed a positive correlation between female percentages in the mid- and upper-level course data and doctoral-granting units that have seen an increase in minors (mid-level course r=.35, n=51, p<.01; upper-level course r=.30, n=52, p<.05). We saw no statistically significant correlation with the increased number of minors in the URM student enrollment data. The CRA Enrollment Survey did not collect diversity information about minors. Thus, it is not possible to look more deeply into this finding from the collected data. Perhaps more women are minoring in computer science, which would then positively impact the percentage of women in mid-and upper-level courses. However, units that reported an increase in minors also have a higher percentage of women majors per Taulbee enrollment data (r=.31. n=95, p<.01). Thus, we can't be sure of the relative contribution of women minors and majors to an increased percentage of women overall in the mid- and upper-level courses. In short, more research is needed to understand this finding.
- Very few units specifically chose or rejected actions due to diversity. While many units (46.5%) stated they consider diversity impacts when choosing actions, very few (14.9%) chose actions to reduce impact on diversity and even fewer (11.4%) decided against possible actions out of concern for diversity. In addition, only one-third of units believe their existing diversity initiatives will compensate for any concerns with increasing enrollments, and only one-fifth of units are monitoring for diversity effects at transition points.
This article considers the impact of the recent enrollment surge in computer science on the diversity of the field. While units want to increase the diversity of their students, unit leaders need to be aware that enrollment management can impact the students they want to retain and attract. In other words, there is concern that the actions departments take to manage increased enrollments might have a side effect of reducing diversity. At this point, we see substantial increases in the number of women and underrepresented minorities enrolling in computer science courses, and median percentages of both groups at all course levels has also grown. This result holds for all courses for which there is sufficient data from our survey, and in both doctoral-granting and non-doctoral-granting units. In other words, growth in female and URM students appears to be greater than the overall growth in students. Yet, we must continue our efforts to attract and retain these populations, as the percentages are still nowhere near where they should be. Our course data for both female and URM students show, in any given year, decreased representation as the course level increases. Further study is needed to determine whether a leaky pipeline exists, or whether there is another explanation for this trend.
This period of unprecedented growth in the field may be an opportunity to increase the diversity of computer science undergraduates. The CRA data shows that very few units are specifically choosing or rejecting actions due to diversity, but those that do have a higher percentage of female and URM students. It is likely that many of these units have been considering diversity in their actions over the long term, not just with respect to the enrollment surge. We strongly encourage units to track the diversity of their majors, minors, and non-majors and to consider the effect of the actions that they take on the diversity of their programs. We need to ensure that the diversity gains we have recently seen can bring the computing community closer to representing the population that uses the technology we create.
We acknowledge everyone who has assisted with the survey, data, analysis, or report in .
2. Computing Research Association (2017). Generation CS: Computer Science Undergraduate Enrollments Surge Since 2006; http://cra.org/data/Generation-CS/. Accessed 2017 May 17.
3. Computing Research Association, The Taulbee Survey; http://cra.org/resources/taulbee-survey/. Accessed 2017 May 17.
5. National Center for Education Statistics; https://nces.ed.gov/ipeds/. Accessed 2017 May 17.
6. National Science Foundation, National Center for Science and Engineering Statistics, TABLE 2-1: Undergraduate enrollment at all institutions, by citizenship, ethnicity, race, sex, and enrollment status: 2004–14; https://www.nsf.gov/statistics/2017/nsf17310/data.cfm. Accessed 2017 May 17.
7. Pew Research Center. Modern Immigration Wave Brings 59 Million to U.S., Driving Population Growth and Change Through 2065: Views of Immigration's Impact on U.S. Society Mixed. (Washington, D.C., 2015).
Department of Computer Science
Colorado School of Mines, 1500 Illinois Street, Golden, CO 80401 USA
W. Richards Adrlon
College of Information and Computer Sciences
University of Massachusetts Amherst, 140 Governors Drive, Amherst, MA 01003 USA
Computing Research Association
1828 L Street NW, Suite 800, Washington DC 20036 USA
Department of Computer and Information Science
University of Pennsylvania, 3330 Walnut Street, Philadelphia, PA 19104 USA
School of Computing
University of Utah, 50 S. Central Campus Drive, Salt Lake City, UT 84112 USA
Department of Computer Science
Purdue University, 305 N. University Street, West Lafayette, IN 47907 USA
Computer Science Department
Hiram College, 11730 Garfield Road, Hiram, OH 44234 USA
Department of Computer Science and Engineering
The Ohio State University, 2015 Neil Ave., Columbus, OH 43210 USA
2. IPEDS doctoral-granting units are institutions that grant at least 20 research/scholarship doctorates in any field (not including professional practice doctoral-level degrees such as JD and MD), following the Carnegie Classification of Institutions of Higher Education. Taulbee and NDC survey data define doctoral- and non-doctoral-granting units by whether the unit grants doctorates in computing.
3. Composites that were used for correlations include "Any Diversity Action Taken" and "Diversity Considered in Decisions." Specifics of how these composites were determined are given in the Methodology section of .
Figure 1. IPEDS CS Diversity Data (CIP 11.0101 and 11.0701) for doctoral-granting (a) and non-doctoral-granting (b) units. Almost 47,000 CS degrees were awarded in 2015. The data is for all Carnegie-classified, not-for-profit, 4-year public and private institutions in the U.S., Guam, Puerto Rico, and the U.S. Virgin Islands.
Figure 2. Median percentage of female students in the courses surveyed. The numbers in parenthesis denote the number of units that provided data for the course in all three years. Numbers above a bar indicate the percent change from 2005.
Copyright held by Authors/Owners. Publication rights licensed to ACM. $15.00.
Request permission to publish from email@example.com
The Digital Library is published by the Association for Computing Machinery. Copyright © 2017 ACM, Inc.Contents available in PDF
View Full Citation and Bibliometrics in the ACM DL.
To comment you must create or log in with your ACM account.