African American Undergraduate Degrees in Computer Science: Last Five Years

This is the fourth in a series this summer on African American undergraduate degrees following the first three on chemical engineering, mechanical engineering and computer engineering. See also Rapid Degree Growth, Engineering Degree Trends, Filtered Pathways and Big Drop for more background. For a broad view of engineering degrees in the past year see ASEE's Engineering by the Numbers, by Dr. Brian L. Yoder.

Computer science (CS) as a discipline separate from mathematics or engineering at US universities is quite young. Five of the first US academic departments, Purdue University (est. 1962), University of North Carolina (est. 1964), University of Illinois (est. 1964), University of Iowa and Stanford (est. 1965) recently celebrated their 50th anniversaries along with many others since and to come. Much of the history shared by the Purdue program defines many elements shared as the discipline has evolved. Depending on the scale, financial model and the history of the academic institution wherein CS programs are found, CS departments appear in larger units defined as engineering colleges, science colleges, liberal arts colleges as well as colleges or schools of computing. In some institutions separate CS or CS-related research themes may have grown in several places with CS ultimately becoming more interdisciplinary while efforts to more strongly define and redefine a CS discipline tied to a rapidly evolving technology sector have taken place. CS and the identity of CS programs is also strongly tied to the history of professional organizations that CS professionals and faculty have defined across this time. The two largest computing professional organizations are the Association for Computing Machinery (ACM, est. 1947, reports 100,000 members) and the IEEE Computer Society (IEEE-CS, est. 1963, reports 60,000 members). These two organizations jointly founded what was originally called the Computing Sciences Accreditation Board (now CSAB) in 1984. CSAB is the lead organization that provides input in the ABET accreditation process for CS. As of 2017 there were 296 programs that were listed as accredited through ABET.  

The large range of contexts wherein CS programs can be found and different approaches to data make a full reckoning somewhat challenging. The American Society for Engineering Education (ASEE) collects comprehensive data on engineering and computing disciplines, but it is primarily connected with institutions that have an engineering degree program. For those institutions ASEE CS data can be listed as data “inside” or “outside” the engineering college. This dataset does not capture many of the programs that exist in universities without engineering programs or at institutions wherein the engineering programs do not participate in ASEE’s data submission. CS data is also reported to the Computing Research Association (CRA) with the two datasets being mostly complementary and latter focusing mostly on doctorates and research and programs at public doctoral institutions. CRA also conducts a survey of students called the Data Buddies Project.  When it comes to tracking diversity, and in the case of this post, a specific focus on African Americans, it is evident that among professionals and postsecondary education the participation of African Americans is low relative to their fraction of the US population. Considerable evidence suggests that K-12 experiences of African Americans are a significant contribution to the likelihood that they will study CS at a university. 

For this post I have used the ASEE database to document the status of CS for undergraduate degrees earned by African Americans. Overall, ASEE data presented here covers a total of 213 CS programs for which data was reported in each of the past five years. For these programs the number of CS undergraduate degrees has increased by nearly 80% in just five years. As mentioned above, this will not provide a full view, but it is a quite extensive documentation of CS degree completion for the US. From 2013-2017, ASEE data shown here includes a total of 160 public and private universities that reported CS BS degree completions inside of an engineering college or school in each of the past five years. That number includes 63 public universities and 18 private universities that also reported data to CRA for the Taulbee Survey. The larger number for the ASEE data is substantially comprised of institutions without CS doctoral programs. Data shown here from ASEE also includes a total of 53 public and private universities that reported data for programs outside of engineering. Nineteen public universities and ten private universities that reported CS BS degree completions for programs outside of engineering colleges also reported to CRA for the Taulbee Survey. Six universities reported data in both categories since, e.g. University of California, Davis or University of California, Irvine, had a computer science and engineering degree program in the engineering school or college and a computer science degree program in another school of college. CRA’s Taulbee Survey data is overall less detailed for undergraduate programs, but only 99 of the 149 programs that reported CS BS data to CRA for 2017 did so in each of the last five years.  

Table 1 Computer Science Bachelor’s Degrees for Programs inside Engineering Schools or Colleges (ASEE Data- African American % is the fraction of domestic US degrees)

CS has below average gender diversity among all computing and engineering disciplines, although it improved from about eleven percent to nearly seventeen percent female degrees from 2009 to 2017. Table 1 shows that across the past five years the fraction of domestic CS undergraduate degrees earned by African Americans for programs inside engineering colleges has been fairly steady, and about the same as the 4.1% seen for engineering degrees, while the international fraction of all CS BS degrees has increased. As is true for most engineering and computing disciplines the female fraction of African American degrees is higher than for other groups. Table 2 shows the reported fraction of domestic CS undergraduate degree earned by African Americans for programs outside of engineering colleges at universities with a separate engineering college or school. These two tables suggest that the fraction of African Americans and the fraction of women earning undergraduate degrees for CS programs inside of engineering colleges has been slightly higher than for CS programs outside of engineering colleges. The 2017 Taulbee Survey reported that the African American fraction of CS BS degrees across the past five years as a bit lower than shown here at 3.1%.** 

Table 2 Computer Science Bachelor’s Degrees for Programs outside Engineering Schools or Colleges (ASEE Data- African American % is the fraction of domestic US degrees)

Because the overall numbers can be quite small for individual institutions, it is easiest to make degree production comparisons across several years at one time. Of the 207 universities* programs that reported CS BS degrees to ASEE in each of the five years from 2013 to 2017, 184 reported one or more degrees earned by African Americans. These programs reported graduating classes averaging 75 total students with the median graduating class numbering 49, although both the number of programs and the size of existing programs is growing rapidly across this time.

Table 3 shows the CS BS degree production for universities reporting more than forty or African American undergraduate degrees across the past five years. These thirteen institutions produced nearly forty percent of the African American CS BS degrees reported to ASEE across this period, but there are considerable differences in the attributes and scale of the institutions in this table. Three of the institutions, Prairie View A&M,^ North Carolina A&T, and Howard University are historically black institutions (HBCUs).

Table 3 Top African American Computer Science BS Degree Producers 2013-2017 (Five Year Total- from ASEE Data except where noted)

Table 4 shows the twenty CS BS programs at doctorate-producing public institutions with the highest fraction of African American CS BS degree recipients across a five-year period and the population fraction of African Americans in those states. It is notable that for nine public doctoral institutions the average number of CS BS degrees earned by African Americans was five or less across programs that produced an average of 226 CS BS degrees per year. Further, among the twenty-five public doctoral institutions producing the largest number of CS BS graduates, eleven had African American CS BS degree fractions of 2% or less. 

Table 4 Highest African American CS BS Degree Fractions at Public Doctoral Institutions (2013-2017 ASEE Data- African American % is the fraction of domestic US degrees)

Summary

There has been little progress in increasing the fraction of American CS BS degree recipients who are African Americans. Progress will likely only take place through a concerted effort by industry, professional societies, academia and government to foster change, including stronger efforts in graduate degrees. CS undergraduate programs fare poorly compared to many other engineering disciplines in the context of gender diversity and slightly worse than engineering overall in the fraction of African Americans earning undergraduate degrees. Many of the largest CS programs in the US are strikingly behind the national averages for CS BS degrees earned by African Americans.

*For the remainder of the document the data inside and outside of engineering was merged, including for the institutions that reported programs both inside and outside of engineering.

^Prairie View A&M data was obtained from their public institutional research pages at https://www.pvamu.edu/ir/program-review-dashboards/. There may be other programs that should be included that I have missed, but I chose Prairie View A&M due to the overall size as a public HBCU and that they have a readily accessible public-facing database from which I could extract the data.

**CRA's Taulbee Survey includes international students in their calculation, but subtracts the students with unknown ethnicity and unknown origin. A more common and simpler convention subtracts international students from the total to give the fraction of the domestic population. This approach assumes the unknowns do not break differently. Based on CRAs data the African American fraction using this calculation is 2.7%.

All of the data and opinions are solely my personal responsibility and not the opinions of any of the institutions for which I have previously or currently serve. I am happy to correct or clarify the data or listen to perspectives that would result in updates or improvements. I encourage sharing and discussing this with colleagues and feel free to send me your perspectives and opinions so that we all can work on the challenge shown here, together. Thanks to Dr. Marie desJardins for some corrections ?2018 Keith J Bowman