SCIENCE & ENGINEERING INDICATORS

The U.S. higher education system consists of diverse academic institutions that vary in mission; public, private nonprofit, or private for-profit status; degrees offered; learning environment; selectivity level; religious affiliation; and cost (McFarland et al. 2019). This institutional diversity is often regarded as a strength of the U.S. higher education system (Harris 2013), allowing it to serve a range of students and meet many societal goals.

During the 2017–18 academic year, there were approximately 4,400 postsecondary degree-granting institutions in the United States, of which about 37% were public, 40% were private nonprofit, and 23% were private for-profit (Table 2-1).^{​Whether an institution is operated by publicly elected or appointed officials, or by privately elected or appointed officials and derives its major source of funds from private sources, is referred to as its control.,​Most data on institutions and some data on degrees used in this report come from the Department of Education’s Integrated Postsecondary Education Data System (IPEDS), which includes completions data for degrees (associate’s, bachelor’s, master’s, and doctoral), as well as certificates below (fewer than 1 academic year, at least 1 but fewer than 2 academic years, and at least 3 but fewer than 4 academic years) and above (postbaccalaureate and post-master’s) the bachelor’s level.} Public institutions awarded two-thirds of all degrees and nearly 70% of S&E degrees overall (Figure 2-1).

Institutions of higher education can be classified in ways beyond the split among public, private nonprofit, and private for-profit. The Carnegie Classification of Institutions of Higher Education is widely used to characterize differences in academic institutions.^{​The Carnegie Classification of Institutions of Higher Education is available at http://carnegieclassifications.iu.edu/. The Basic Carnegie classification categorizes academic institutions primarily based on highest degree conferred, level of degree production, and research activity. This report uses the 2015 Carnegie classification.} The 115 Carnegie-classified “highest research activity” doctoral universities, for example, play a key role in producing S&E doctoral degrees in the United States (Table S2-1).^{​Graduates from many different types of institutions go on to earn S&E doctoral degrees (Burrelli and Rapoport 2008; Fiegener and Proudfoot 2013).}

There are more than 700 federally designated minority-serving institutions (MSIs) of seven types (National Academies of Sciences, Engineering, and Medicine [NASEM] 2019).^{​For a list of all types and how they are designated, see NASEM 2019: Table 3-1 and 3-2.} MSIs may be defined by legislation (“historically defined”) or by the percentage of minority student enrollment and other characteristics of the student body (“enrollment defined”) (Li 2007; NASEM 2019). Historically black colleges or universities (HBCUs) and tribal colleges or universities are historically defined.^{​The Higher Education Act of 1965, as amended, defines an HBCU as “any historically black college or university that was established prior to 1964, whose principal mission was, and is, the education of black Americans.” In 2015–16, 102 HBCUs were in operation in 19 states, the District of Columbia, and the U.S. Virgin Islands. Half were public institutions, and half were private nonprofit.} High-Hispanic-enrollment institutions (HHEs), in contrast, are one example of enrollment-defined MSIs.^{​The number of HHEs more than doubled from 189 in 1994 to 432 in 2015. About 300 institutions enroll between 15% and 24% Hispanic students and are considered “emerging HHEs.” Many researchers use high-Hispanic enrollment and Hispanic-serving institution (HSI) interchangeably. HSIs meet a federally designated criterion (i.e., public and private nonprofit institutions whose undergraduate, full-time equivalent student enrollment is at least 25% Hispanic) and are eligible to apply for Hispanic-serving institution status. Because there is no information on whether institutions apply for the HSI designation, the National Center for Science and Engineering Statistics (NCSES) uses the 25% enrollment criterion to determine which institutions have high Hispanic enrollment. For additional information, see https://www2.ed.gov/about/offices/list/ope/idues/hsidivision.html.} Some institutions may qualify as more than one type of MSI, and there is substantial diversity in institutional characteristics between different MSIs (NASEM 2019). The number of institutions in historically defined MSI categories are more or less fixed, whereas the numbers of institutions included in enrollment-defined MSI categories may change. This has implications for interpreting trends over time.

MSIs enroll a substantial fraction of underrepresented minority undergraduates. The number of blacks earning S&E bachelor’s degrees from HBCUs has remained roughly constant, likely reflecting the capacity of these institutions. Overall, across all institutions, numbers of S&E bachelor’s degrees earned by blacks are increasing (NCSES WMPD 2019: Table 5-8). In 2016, S&E fields accounted for roughly 30% of the bachelor’s degrees blacks earned at HBCUs; the percentage of S&E bachelor’s degrees blacks earned across all institutions was about the same.

Numbers of Hispanics earning S&E bachelor’s degrees from HHEs have increased, whereas the percentage has remained roughly the same (NCSES WMPD 2019: Table 5-9).^{​For information on S&E bachelor’s degrees awarded by tribal colleges and universities to American Indian or Alaska Native students, see NCSES WMPD 2019: Table 5-10.} This reflects increasing numbers of Hispanics earning S&E bachelor’s degrees at HHEs and at other types of institutions. In 2016, S&E fields accounted for roughly one-third of the bachelor’s degrees that Hispanics earned at HHEs; the percentage of S&E bachelor’s degrees Hispanics earned across all institutions was about the same.

MSIs also play an important role in training underrepresented minority students for doctoral-level study in S&E fields. A considerable share of black and Hispanic S&E doctoral recipients received their bachelor’s degree from an MSI. Around 25% of black S&E doctoral recipients between 2013 and 2017 earned a bachelor’s degree from an HBCU (NCSES 2019b). Likewise, around 37% of Hispanic S&E doctoral recipients between 2013 and 2017 earned a bachelor’s degree from an HHE (NCSES 2019b). The percentages of black and Hispanic S&E doctoral recipients whose baccalaureate origins were HBCUs and HHEs, respectively, have been relatively stable for several decades.^{​This analysis is based on a custom data tabulation from the Survey of Earned Doctorates.} More data on the importance of these institutions as baccalaureate origin institutions is available in Burrelli and Rapoport (2008) and Fiegener and Proudfoot (2013), as well as in Hrabowski and Henderson (2017, 2019), who emphasize the role that predominantly white institutions must also play in preparing underrepresented minority students for S&E doctoral training.

Community colleges (also known as public 2-year colleges or associate’s colleges) play a key role in providing broad access to higher education. Community colleges prepare students to enter the workforce with certificates or associate’s degrees, or to transition to 4-year institutions (frequently without earning associate’s degrees).^{​A recent trend among states is to allow more community colleges to offer bachelor’s degrees. See Fulton (2015) for more information.} Of students who earned bachelor’s degrees in any field between 2010 and 2017, more than half (53%) had done some coursework at a community college and about one-fourth (26%) earned associate’s degrees. The figures were 47% and 18%, respectively, among S&E degree recipients.^{​These figures come from the 2017 National Survey of College Graduates, accessed using the Scientists and Engineers Statistical Data System: https://www.nsf.gov/statistics/sestat.} Community colleges also provide a pathway that ultimately leads some students to earn doctoral degrees. About 20% of students receiving S&E doctoral degrees in 2017 reported having attended a community or 2-year college (NCSES SED 2017: Table 30), and about 6% had previously earned an associate’s degree (Table S2-2).​^{These figures are for U.S. citizens and permanent residents earning S&E doctoral degrees. The percentages for students on temporary visas are lower (NCSES SED 2017: Table 30), likely reflecting that many foreign students come to the United States specifically for graduate training.}

Community college attendance, however, varies across degree fields and between demographic groups. The percentage of 2017 doctorate recipients with associate’s degrees ranged from less than 4% among engineering doctorates to nearly 11% among medical and other health sciences doctorates; percentages having attended community college were higher (NCSES SED 2017: Table 30). Black S&E doctoral degree recipients were most likely to have earned an associate’s degree (11%) and Asians least likely (less than 3%). In addition, among the civilian college graduate population in the United States, military veterans were more likely to have attended community college, and to have earned associate’s degrees, than nonveterans (Milan 2018).

In 2017–18, there were about 1,000 degree-granting private for-profit higher education institutions in the United States. About half of these institutions award only associate’s degrees; the remainder award higher degrees, including 10% that award doctoral degrees (Table 2-1). Degrees awarded by for-profit institutions rose dramatically throughout the 2000s but declined in recent years.^{​Data are from IPEDS.}

For-profit academic institutions are not large producers of S&E degrees (Figure 2-1). The S&E degrees they award cluster in a few fields: 75% of S&E bachelor’s degrees awarded by for-profits in 2017 were in computer sciences, psychology, or political science and public administration;^{​Most of the remainder, 18%, were in “other social sciences.”} at the doctoral level, 88% of S&E degrees were in psychology, medical sciences, or political science and public administration.

Distance education is learning where the teacher and student are separated by time or space (Miller, Topper, and Richardson 2017). Distance education has been around for more than 100 years, whereas online education is a relatively new phenomenon, mirroring the growth of technologies such as the personal computer and the Internet (Miller et al. 2017; Perna et al. 2014).^{​Information on the history of distance and correspondence and online education may be found in Harasim (2000).} Distance education offerings can be delivered through a range of technologies and in a variety of modes: completely distance education, traditional or in-person instruction, and a combination often referred to as hybrid or blended instruction (RTI International 2017). Distance education may occur synchronously (live) or asynchronously (prerecorded) in time, and online courses may be offered to students living on campus, without any physical distance from the instructor (RTI International 2017).

Distance and online education enable higher education institutions to reach more students by providing greater flexibility and expanding access in remote locations. About 18% of the U.S. adult population live in “higher education deserts,” whereas only about 2% live in “online deserts,” according to a study by Rosenboom and Blagg (2018). The study defines a higher education desert as “either having no colleges and universities within 25 miles or having access to a single community college as the only broad-access public institution within 25 miles.” Online deserts are “areas where Internet speeds are below 25 megabits per second for downloads and 2 megabits per second for uploads.”

Nationally representative data on distance education are available through the Integrated Postsecondary Education Data System (IPEDS). These data include instances where instructional content is delivered exclusively through distance education (i.e., hybrid or blended courses are not included; see RTI International 2017).^{​No standard guideline exists that specifies how much education must be delivered via technology to qualify as online or distance education (Miller et al. 2017). IPEDS defines distance education as “education that uses one or more technologies to deliver instruction to students who are separated from the instructor and to support regular and substantive interaction between the students and the instructor synchronously or asynchronously.” Distance education courses are courses “in which the instructional content is delivered exclusively via distance education.” Distance education programs are those “for which all the required coursework for program completion is able to be completed via distance education courses.” For more detail, see https://surveys.nces.ed.gov/ipeds/VisGlossaryAll.aspx.} The number of distance education programs offered by degree-granting higher education institutions, as defined by IPEDS, stayed roughly the same between 2016 and 2017 (at about 21,000), after a period of rapid growth (from about 11,000 in 2012). This overall trend includes a decline in programs offered at the associate’s and bachelor’s levels and a continued increase in those at the master’s and doctoral levels. In contrast, enrollment of undergraduate and graduate students in distance education increased each year from 2012 to 2016, except for enrollment at for-profit institutions, which declined. Public, private nonprofit, and for-profit institutions all offer distance education programs; however, public universities account for most students taking these courses (Seaman, Allen, and Seaman 2018).^{​This analysis is based on 2016 IPEDS data.}

Most distance education students also take courses on campus; of those who take only distance education courses, 56% do so through institutions in their home state. In S&E fields, distance education programs are most likely to be offered in computer sciences, social sciences, and psychology at the associate’s and bachelor’s levels; in computer sciences and engineering at the master’s level; and in medical and health sciences and engineering at the doctoral level.^{​This is for doctoral degrees-research/scholarship only. Medical and health sciences distance education programs are prevalent at all degree levels but are only included in S&E at the doctoral level. This analysis is based on 2017 IPEDS data.}

Although no nationally representative data exist specifically for online education, a recent analysis of data from massive open online courses (MOOCs) taught on edX by its founding partner institutions found that these courses primarily draw students from affluent countries and continue to face low completion and retention rates (Reich and Ruipérez-Valiente 2019). According to the study, MOOCs have been most successful in helping universities outsource online master’s degrees for professionals. The study, however, noted that this finding might not be broadly generalizable and that information from other edX partners or MOOC providers might “reveal different dynamics.” Related to this, although the literature around online education is robust, it remains unclear how generalizable results from any one program or class are to others, given the complexity of the landscape (Miller et al. 2017). Studies have explored issues such as whether online options substitute for existing university options or expand enrollment (e.g., Goodman, Melkers, and Pallais 2018), how learning outcomes compare with those from traditional courses (e.g., Joyner 2018), and whether course completions lead to beneficial outcomes like higher earnings or job mobility (e.g., Hadavand, Gooding, and Leek 2018).

This section provides information on undergraduate and graduate students’ cost and debt. In terms of cost, published tuition and fees have increased far more than net price. Although the level of undergraduate debt varies by type of institution, the frequency and amount of borrowing have increased little over the past 5 years. Among graduate students, master’s students are largely self-supporting, whereas doctoral students rely on multiple funding sources and mechanisms to support their education. The percentage of doctorate recipients holding debt related to their graduate education has changed little over the last 10 years.

Earning a college degree commands a substantial wage premium (Carnevale, Cheah, and Hanson 2015) and provides other benefits to individuals and society (College Board 2016). For these and other reasons, many students and their families invest in higher education. Increases in published prices over time have far exceeded inflation or increases in average family income,^{​Published price is “the price institutions charge for tuition and fees, as well as room and board, in the case of students residing on campus” (College Board 2018a).} contributing to concerns about affordability of higher education (Archibald and Feldman 2012; U.S. Congress Joint Economic Committee 2017).

College pricing is complex and often poorly understood. Institutions operate at different prices. For example, during the 2018–19 academic year, average tuition and fees were $3,660 at public 2-year institutions, $10,500 at public 4-year institutions, and $36,980 at private nonprofit 4-year institutions.^{​Tuition and fee figures represent charges to full-time first-year undergraduate students over the course of a 9-month academic year of 30 semester hours or 45 quarter hours. Average tuition and fee prices reflect in-district charges for public 2-year institutions and in-state charges for public 4-year institutions. In addition to tuition and fees, room and board constitutes another expense for students.} Published tuition and fees have greatly increased over the last 30 years (College Board Trends in College Pricing 2018: Figure 3).

Net price, defined by the College Board as “what the student and/or family must cover after grant aid and savings from tax credits and deductions are subtracted,” is more relevant for students than published price. Across institutional types, net prices rose much more slowly than published prices (College Board Trends in College Pricing 2018: Table 7). For 2018–19, average net tuition and fees were $3,740 at public 4-year institutions and $14,610 at private nonprofit 4-year institutions. At public 2-year institutions, students on average received enough funding through grant aid and federal education tax credits and deductions to more than cover tuition and fees. Net price, however, varies based on family income (College Board Trends in College Pricing 2018: Figure 2017_11). In addition, the cost of attending public 4-year universities varies by state of residence and whether students attend an institution in their own state or in another state.^{​Limited data on the residence and migration of first-time freshmen are available from IPEDS.} The affordability of college also depends on time to degree and the amount of money spent on other living expenses, like room and board.

Level of undergraduate debt varies by type of institution, but the frequency and amount of borrowing among those who graduate have increased little over the past 5 years. At public 4-year institutions, 58% of 2016–17 graduates borrowed, holding an average of $26,900 in debt. At private nonprofit 4-year institutions, 61% of 2016–17 graduates borrowed, holding an average of $32,600 in debt (College Board Trends in Student Aid 2018: Figure 15). Since the 2011–12 academic year, the percentage of bachelor’s degree recipients holding debt has remained essentially the same (about 60%), while the average level of debt has risen by 3% in inflation-adjusted dollars (College Board Trends in Student Aid 2018: Figure 15).^{​Note that these debt figures are for degree recipients only. Students who have taken on debt but left without earning a degree are not included.}

Students who attend private for-profit institutions are more likely to borrow (87% of bachelor’s degree recipients) and to borrow larger amounts (32% borrowed more than $50,000, vs. 7% of those earning bachelor’s degrees from public universities) (College Board Trends in Student Aid 2018: Figure 16).

Debt level also varies by state. Average debt for 2016–17 graduates of public 4-year colleges and universities ranged from $19,800 in Utah to $36,700 in Pennsylvania. Average debt for graduates of private nonprofit colleges and universities ranged from $16,800 in Utah to $38,800 in Connecticut (The Institute for College Access & Success 2018). Cost of living may account for some of the differences among states.^{​For public institutions, data were not available for Connecticut, the District of Columbia, or North Dakota. For private nonprofit institutions, data were not available for Alaska, Arizona, Delaware, Idaho, Nevada, North Dakota, Rhode Island, or Wyoming.}

Graduate students’ sources of financial support depend on their level of study.^{​This report discusses sources and mechanisms of graduate student funding. Funding sources include federal, institutional, and personal or self-support, among others. Funding mechanisms include assistantships, fellowships, and traineeships, among others.} Master’s students are largely self-supporting, whereas only a small minority of doctoral students self-finance.^{​Starting with 2017, the Survey of Graduate Students and Postdoctorates in Science and Engineering (GSS) collects data separately for master’s and doctoral students. Analyses using GSS data in this report include health fields.} In 2017, two-thirds of S&E master’s students paid for their graduate program using personal sources (Arbeit, Davies, and Yamaner 2019; NCSES GSS 2017: Table 3-1); by contrast, only around 10% of doctoral students did so.^{​Personal sources include loans (including federal loans) or personal or family financial contributions. Note that there may be some differences between numbers reported in Arbeit and colleagues (2019) and those reported here because different fields are included in some analyses.} These differences generally hold across all S&E fields.

Other main sources of support for graduate students include academic institutions (where a student is enrolled) and the federal government. Academic institutions were the primary source of support for 24% of master’s students and 57% of doctoral students. Federal support is discussed in the section “Federal Support for S&E Graduate Students.” Financial support may be delivered through various mechanisms, including research assistantships (RAs), teaching assistantships (TAs), and fellowships.^{​In 2017, these were the most common funding mechanisms for S&E doctoral students (37% received RAs, 27% TAs, and 15% fellowships) (NCSES GSS 2017: Table 3-5).} TAs and fellowships are mainly institutionally funded, whereas nearly half of RAs are funded through federal academic research grants.^{​Tabulations from the 2016 GSS were created using the NCSES interactive data tool.}

Most doctoral students are supported by multiple sources or mechanisms during graduate school, even in a single academic year. Patterns of support vary by field (Figure 2-2; GSS 2017: Table 3-1) and the type of institution attended (Table 2-2). For example, in 2017, doctorate recipients from the “highest research activity” doctoral universities (based on Carnegie classification) were less likely to self-support (7%) than those who attended other types of institutions. Even among this group of institutions, however, public and private universities used funding mechanisms at different rates (Table 2-2).

Funding mechanisms also vary by demographic groups (Table S2-3) for U.S. citizens and permanent residents who earned S&E doctorates between 2015 and 2017. Overall, men (31%) earning S&E doctorates were more likely to be supported by RAs than women (22%), whereas women (18%) were more likely to self-support than men (10%). Women (16%) and men (17%) were around equally likely to be supported by TAs. Asians (32%) and whites (29%) were most likely to have primary RA support. To some extent, demographic differences in support mechanisms relate to field-of-study differences. However, certain patterns hold across fields. For example, black doctorate recipients were more likely to use personal sources of funding in every S&E field for which data were available. Data on primary source of support for S&E doctorate recipients by race, ethnicity, sex, and field are available (NCSES WMPD 2019: Table 7-26). These differences may also reflect variation among groups in the types of institutions attended.

Debt levels are an indicator of external financial support for doctoral training. The percentage of doctorate recipients (around 35%) holding any debt related to their graduate education has not changed much in the last 10 years (NCSES SED 2017: Table 39).^{​At the time of doctoral degree conferral, 44% of 2017 doctorate recipients held debt related to their undergraduate or graduate education, or both (NCSES SED 2017: Table 38).}

Debt levels vary by field of study. A greater percentage of doctorate recipients in non-S&E fields (50%) than in S&E fields (29%) reported graduate debt (NCSES SED 2017: Table 38). Levels of debt also vary among S&E fields. For instance, doctorate recipients in psychology and social sciences are more likely to hold graduate debt (49% hold debt, with an average of nearly $25,000) than those in physical and earth sciences (19% hold debt, with an average of about $5,000) (NCSES SED 2017: Table 38).

Debt levels also vary across demographic groups. Women are more likely than men to accumulate higher graduate debt (NCSES SED 2017: Table 40). Across fields, black doctorate recipients are most likely to hold more than $30,000 in graduate school debt, and Asians are least likely (NCSES SED 2017: Table 41). Other factors influencing debt include the type of institution attended, time to degree, marital and dependent status, and highest level of parental education (Schacht, Hoffler, and Fiegener 2018 in preparation).

Many factors leading to greater indebtedness are related. For instance, black doctorate recipients are more likely to possess other characteristics related to higher indebtedness: for example, many are female, attend for-profit institutions, and earn degrees in non-S&E fields or S&E fields like psychology and social sciences, which have lower levels of external financial support (see the later section “Demographic Attributes of S&E Degree Recipients” and Scott-Clayton and Li 2016).^{​A recent analysis, however, concluded that underrepresented minority status mattered to debt levels even after accounting for educational and other factors (Schacht et al. in preparation).}

This section discusses sources of financial aid for undergraduate and graduate students, as well as federal support for S&E graduate students. The federal government is the largest provider of financial aid to undergraduate and graduate students. It also supports a large number of S&E doctoral students through RAs (through research grants to universities) and other funding mechanisms.

In 2017–18, undergraduate students received $184 billion in federal, state, institutional, and other aid (excluding nonfederal loans) (College Board Trends in Student Aid 2018: Figure 3).^{​Sources used by the College Board are available at https://trends.collegeboard.org/student-aid/notes-sources. Note that this section reflects information for all students, not just those studying S&E fields.} Over the last 10 years, federal financial aid has constituted most undergraduate student aid, although the federal percentage of total aid declined from 74% ($153 billion) in 2010–11 to 61% ($112 billion) in 2017–18. In 2017–18, loans constituted half of the federal investment in higher education ($56 billion) and Pell Grants constituted 25% ($28 billion).^{​The remainder includes veterans’ benefits, education tax benefits, and work-study programs and Federal Supplemental Educational Opportunity Grants.}

Among nonfederal sources, grants from institutions themselves grew steadily throughout the last decade, increasing from 21% ($26 billion) to 26% ($49 billion) of total aid. State grants declined slightly, and private and employer grants remained steady (each constituting 6% to 7%).

Graduate students received $57 billion in federal, state, institutional, and other aid (excluding nonfederal loans) in 2017–18 (College Board Trends in Student Aid 2018: Figure 4). As with undergraduates, federal financial aid constituted the majority of graduate student aid over the past 10 years. Federal aid was around 72% ($41 billion) of total aid in 2017–18. Loans were the main component of federal aid: 91% ($38 billion in 2017–18); the remainder consisted of veterans’ benefits, education tax benefits, and work-study programs.

Federal support for S&E graduate students reflects a continuation of the historic partnership between the federal government and the nation’s research universities to integrate the performance of basic scientific research and the education and training of the next generation of scientists and engineers (National Research Council 2012). It is an indicator of the strength of the university-government partnership.

The federal government supported 15% of full-time S&E graduate students (just under 70,000) in 2017 (NCSES GSS 2017: Table 1-6), down from nearly 21% (84,000) in 2004. This overall figure masks differences in federal support for master’s (about 12,000 students, or 5%) and doctoral (about 57,000 students, or 24%) students. The federal government supports a higher percentage of doctoral students than master’s students in all S&E fields.

The largest numbers of federally supported graduate students are in engineering (22,000), biological and biomedical sciences (18,000), and physical sciences (10,000). Together, these three fields contain around half of total graduate students but more than 70% of federally supported students (NCSES GSS 2017: Table 3-1).

The National Science Foundation (NSF) and the National Institutes of Health (NIH) supported the most graduate students in 2017: 21,000 and 20,000, respectively. Together, these two agencies supported nearly 60% of federally supported graduate students (NCSES GSS 2017: Table 1-7). However, in recent years, the numbers of students supported have declined (Figure 2-3). Other agencies supporting substantial numbers of S&E graduate students in 2017 were the U.S. Department of Defense (DOD: 8,300) and the U.S. Department of Energy (DOE: 4,500), followed by the Department of Agriculture (USDA), the National Aeronautics and Space Administration, and other agencies in the U.S. Department of Health and Human Services.

In their support patterns across fields, agencies take on portfolios consistent with their missions (Figure 2-4, NCSES GSS 2017: Table 3-3). NSF supports substantial numbers of students across a range of fields, whereas about 70% of those supported by NIH are in biological and biomedical sciences and health fields. More than half of the students funded by DOD study engineering, and nearly 90% funded by DOE are in physical sciences and engineering.

RAs are the primary mechanism the federal government uses to fund graduate students. Among full-time S&E graduate students primarily funded by the federal government in 2016, 71% received RAs, followed by fellowships (13%) and traineeships (9%).

Reflective of overall enrollment patterns, about 70% of graduate students supported by the federal government in 2016 were enrolled in public institutions.^{​This ranges from less than 60% for NIH to more than 97% for USDA.} Seventy-eight percent were enrolled in Carnegie-classified “highest research activity” doctoral universities, and another 13% were enrolled in “higher research activity” doctoral universities. Only 2% of federally supported graduate students were enrolled in “moderate research activity” doctoral universities, along with about 4% at medical schools and 2% at master’s colleges and universities.^{​Tabulations of the 2016 GSS were created using the NCSES interactive data tool.}