2

Historical Degree Production in Computing

Before assessing current undergraduate computing enrollment trends at U.S. institutions, it is imperative to consider the historical context. The most authoritative data on the historical production of computer science (CS) bachelor’s degrees lie in the national statistics on degree completions tracked by the Department of Education’s Integrated Postsecondary Education Data System (IPEDS).

This chapter presents and discusses time series trends in degree production to provide context for discussion of current enrollment trends. Specifically, it examines degree production trends in CS and related fields relative to net production of bachelor’s degrees over time, distribution of postsecondary degrees produced in these fields, past economic and institutional factors that likely contributed to past CS bachelor’s degree production trends, and variations by institution type from 2009 to 2015 to highlight the range of experiences at different institutions.

The IPEDS data discussed in this chapter come from three different cross sections of CS and related fields, labeled and defined as follows:¹

1. CIS—Includes the entire 11.x series of IPEDS Classification of Instructional Programs (CIPs), referred to within IPEDS alternately as “computer science” and “computer and information science and support services.” This broad category includes core computer science, information science/studies, information technology, and related fields, some of which are more vocational than traditional computer science programs.

___________________

¹ See the section “Data Used in the Development of This Report” in Chapter 1 for additional discussion of the IPEDS classification system for CS and related fields.

2. Core CS—Includes only “Computer and Information Sciences, General” (11.0101) and “Computer Science” (11.0701). These two CIPs have been considered to be closest to the core computer science degree programs currently offered at institutions of higher education offering bachelor’s and higher-level degrees. However, this set of classifiers may not necessarily reflect temporal trends in degree production accurately, due to the changes in IPEDS CIP codes in 2000 and 2010.
3. Computer engineering (CE)—Includes degrees from the 14.09xx series, identified as “Computer Engineering, General” (14.0901), “Computer Hardware Engineering” (14.0902), “Computer Software Engineering” (14.0903), and “Computer Engineering, Other” (14.0999).

Details about the full set of classifiers of all plotted data from IPEDS are described in Appendix F.

HISTORICAL PRODUCTION OF CIS AND CE BACHELOR’S DEGREES

Since the emergence of CS programs in the 1960s, overall production of CS and related bachelor’s degrees at U.S. institutions of higher education has grown significantly. CE was first uniquely identified as a subfield of engineering in federal education statistics in 1987, and generates about 10 percent of the number of graduates produced in CIS. The number and share of CIS and CE degrees produced over time are illustrated in Figure 2.1, with total bachelor’s degree production in all fields indicated as a dashed line for context (vertical scale read from the right axis).² The plots for “All Institutions” include bachelor’s degrees at all public, private not-for-profit, and private for-profit institutions; plots for “Not-for-Profits” exclude the bachelor’s degrees conferred by for-profit institutions.

The significant growth in the 50 years since the emergence of computer science as a unique academic discipline is marked by two striking peaks, with subsequent valleys resetting at a higher absolute level than the presurge level. Specifically, the absolute number of CIS bachelor’s degrees reported annually rose from 89 in 1966 to 60,309 in 2015, with peaks in 1986 (42,195 bachelor’s degrees) and 2003 (57,926 bachelor’s degrees) and valleys in between (24,553 in 1994 and 38,496 in 2009). CE also exhibits historical fluctuations.

CIS BACHELOR’S DEGREES GOING TO FOREIGN STUDENTS

The total number and fraction of all U.S. postsecondary degrees awarded in CIS to temporary residents (international students) are illustrated in Figure 2.2.

___________________

² See Chapter 1 for a more detailed description of the associated data source (IPEDS degree completions data) and associated limitations.

Although a moderate increase in the share of CIS bachelor’s degrees earned by international students occurred between 2008 and 2015 (from 4.6 to 5.6 percent overall), this constitutes only a small portion of U.S. bachelor’s degree production, and suggests that little of the recent surge is due to increased participation among foreign students.³ The fraction of CIS bachelor’s degrees awarded to temporary residents in 2015 remained below past peak levels.

___________________

³ Note that there is some uncertainty in these data given that a nontrivial fraction (around 10 percent in recent years) of these students declared neither a specific race/ethnicity nor status as a temporary resident (which are treated as mutually exclusive categories within the IPEDS classification system).

VARIATION IN CIS BACHELOR’S DEGREE PRODUCTION BY INSTITUTION TYPE

While the overall picture of computing degree production is helpful, it represents only aggregate trends and not the variation of experiences at different institutions. In the following these data are broken out in more detail using the IPEDS classifiers of “Institutional Control” (to identify trends at public, private,

and for-profit institutions) and “Carnegie Classification” (to identify trends by institution type).

Public, Private, and For-Profit Institutions

The historical production of degrees (at all levels) in CIS and CE is broken out for the categories of public, private, and for-profit institutions in Figure 2.3, beginning with the earliest dates for which data with the corresponding classifiers are available.

Between 2000 and 2005, for-profit institutions emerged as a significant source of CIS degrees, roughly matching the number awarded at not-for-profit private institutions. It is noteworthy that CIS accounted for between 8 and 28 percent of bachelor’s degrees awarded at for-profit institutions since 1987, but accounted for only between 2 and 4 percent of bachelor’s degrees at public and private institutions. Again, this likely reflects both the breadth of the CIS classifier in IPEDS and the nature of the computing labor market. Since 2013 the number of CIS bachelor’s degrees awarded at for-profit institutions has been falling, as has the relative fraction of bachelor’s degrees conferred in CIS (compared to the number conferred in all disciplines) at for-profit institutions.

CIS degree production at public and private not-for-profit institutions has surged since 2009. The increases have been most pronounced at public institutions, which produced more degrees in 2015 (33,930) than during the dot-com era peak in 2003, representing an increase of 87 percent since the post-dot-com

low of 18,189 in 2009. At private institutions 15,361 degrees were produced in 2015 (compared to a peak of 17,846 in 2003), a 52 percent increase since 2009.

As previously noted overall bachelor’s degree production has also been increasing, as illustrated in Figure 2.4. Nonetheless, CIS bachelor’s degree production has been increasing at an even higher rate, as illustrated in Figure 2.1. The number of CIS bachelor’s degrees produced annually at all not-for-profit institutions increased by 74 percent between 2009 and 2015, compared to an overall increase in bachelor’s degree production of only 16 percent at not-for-profits during the same period.

Carnegie Classification

American institutions of higher education are very diverse, and it is important to consider how CS production has varied at the different institution types to get an accurate picture. To illustrate which institutions have experienced the most growth in recent years, these data are broken down further by survey-specific (2010) Carnegie Classification⁴ and by for-profit status.⁵ Degree production trends by institution type are illustrated in Figure 2.5 for the following categories of institution:

• Research includes institutions classified as not-for-profit doctoral-degree-granting institutions with “very high” or “high” research activity, both public and private.
• Doctoral includes all other not-for-profit doctoral degree granting institutions, both public and private.
• For-profit doctoral includes all for-profit doctoral institutions.
• For-profit, other includes all other for-profit institutions.
• Master’s includes not-for-profit institutions whose highest degree offered (in any field) is at the master’s level (many of these have a professional character).
• Bachelor’s includes not-for-profit institutions at which a bachelor’s is the terminal degree, including many liberal arts colleges.
• Associate’s includes institutions that primarily offer associate’s degrees, but also have some bachelor’s degree programs.
• Other/unknown includes not-for-profit tribal colleges, special-focus institutions, and institutions that did not report a Carnegie Classification.

Combined, research and master’s institutions have long produced the majority (67 percent in 2015) of all CIS bachelor’s degrees. Research institutions reached a peak in CIS bachelor’s production in 2003, then fell until 2008, after which the number climbed again, surpassing the 2003 peak in 2015. Doctoral, master’s, and

___________________

⁴ For the Carnegie Classification definitions and methodology, please see The Carnegie Classifications of Institutions of Higher Education (2017).

⁵ Via the IPEDS “Institutional Control” classifier.

bachelor’s institutions follow a similar trend, but only nearly reaching their 2003 peaks in 2015. Production at for-profit institutions has exhibited a significant falloff beginning in 2013 (for-profit doctoral institutions) and 2014 (other for-profits).

Since 2009, growth in CIS bachelor’s degree production has occurred at research, master’s, doctoral, and bachelor’s institutions, though at different average rates. For-profit institutions experienced some growth, and then dropped precipitously. The recent (between 2009 and 2015) growth is having the largest impact on research institutions (97 percent increase), followed by master’s institutions (69 percent increase), other doctoral institutions (50 percent increase), and bachelor’s institutions (33 percent increase), on average.

For a closer look at relative trends in research institutions, this category was further separated by level of research activity. Figure 2.6 illustrates the total number of CIS degrees produced over time for those institutions with very high research activity (The Carnegie Classifications of Institutions of Higher Education, 2017). Active research institutions display much more pronounced growth than nonresearch doctoral institutions. Thus, production has shifted toward leading research institutions. As with the other categories, time will determine whether this reflects concentration at the top or a lag in growth.

To provide a sense of the nature of the recent onset of growth at not-for-profit institutions, the average number of degrees per institution in 2009 and 2015 is listed in Table 2.1 for all institution categories that have displayed increased

TABLE 2.1 Change in Average Number and Institutional Share of CIS Bachelor’s Degrees Produced per Institution by Not-for-Profit Institution Type, 2009-2015

NOTE: Average number of degrees per institution rounded to nearest integer. Only institutions reporting CIS bachelor’s degrees are included in each category. SOURCE: Data from IPEDS completions survey.

degree production since 2009. Average relative increases in program size are clearly higher for the more research-intensive institutions and for master’s institutions as compared to bachelor’s.

Some caution should be exercised when using these data to make assumptions about the nature of any particular CIS program, as program size does not follow a normal distribution. Overall, the top 10 percent of CIS degree producers are responsible for more than 50 percent of the degrees, and the smallest 50 percent of the institutions produce less than 10 percent. Less than 10 percent of the institutions experiencing growth are responsible for more than 50 percent of the growth. Furthermore, degree production is but one aspect of the experience in computing programs.

National degree production data do not provide any indication that the growth will soon abate. Given a finite population, it must level off eventually, or even peak, but there is no strong basis for predicting when that will occur.

Comprehensive national data are not available on growth of course enrollments, which reflect the increase in demand for courses both by majors and by the general student population, as well as the emergence of various blended majors.

This topic is explored further in Chapter 3. Nonetheless, these data provide illustrative analyses of historical CIS degree production trends—in particular, in the context of the recent boom.

OTHER LEVELS OF DEGREE

Surges also occurred in the number of CIS associate’s degrees (peaking in 1984 with 12,913 degrees and in 2003 with 46,400 degrees). CIS certificates also exhibited a peak in 2003 (25,646 certificates), as did CE bachelor’s and associate’s and CIS master’s. In this recent growth wave, associate’s degrees in CIS from for-profit institutions have fallen substantially.

From Figure 2.7 it is apparent that a large number of CIS associate’s degrees and CIS certificates have been awarded, similar in magnitude to the number of CIS bachelor’s and master’s degrees, respectively, since the mid-1990s. It is notable that CIS, as classified in IPEDS, produces relatively more associate’s degrees and certificates than other science, technology, engineering, and mathematics fields.⁶ Several factors may contribute to this pattern. First, the large number of CIS certificates may reflect interest among non-majors in obtaining computing skills. The types of degrees that are categorized as CIS in IPEDS may also be at play—the corresponding CIP is inclusive of disciplines such as “Information Technology,” “Word Processing,” and “Computer Support Specialist,” which may be more vocational and thus more likely to offer these degrees. Finally, the nature of the market for computing skills may stimulate interest in these degrees.⁷

RESPONDING TO ENROLLMENT SURGES

In general, several categories of responses are available to institutions in the face of booming enrollments: (1) limit participation, (2) grow programs and the resources that feed them, (3) leverage resources in new ways, (4) restructure the nature of computing education within the institution, and (5) make no active changes, allowing higher burdens to be placed on existing resources. The categories range from maintaining current practices, to relatively inexpensive and nondisruptive actions, to significant changes in organization or resource commitments. They are not exclusive and typically are deployed in combination. A range of influences, including specific institutional actions and other influences, may apply over the course of a student’s academic career, and may affect the character and culture of the program and institution.

In the rest of this report dimensions of the current CS enrollment surge are explored, including potential impacts of related institutional actions, which are discussed in more depth in Chapter 6. A discussion of anecdotes from past CS enrollments surge is provided in Box 2.1.

___________________

⁶ This is discussed further in Chapter 4.

⁷ See Chapter 4 for further discussion of the labor market.