5

Increasing Recruitment, Acceptance, Retention, and Graduation of PhD Students in Computing

ADMISSION TO DOCTORAL PROGRAMS

After supporting undergraduate students through computing majors and research experiences, development of the PhD production pathway is dependent on active recruiting and preparation efforts by computing departments. Typically, students are recruited from outside institutions to pursue a doctoral degree. However, departments must prepare students from their home institution to successfully apply for and be accepted into PhD programs to reach the goal of encouraging more domestic students to pursue doctoral degrees in computing.

Active Recruiting of PhD Students

Active recruiting of PhD students includes all efforts pursued by a department or institution with the goal of reaching prospective students who will apply to, and hopefully attend, their graduate program. For this report, these efforts will be focused on recruiting students to computing PhD programs, but efforts may be notably different for computer science and engineering compared to that of information sciences departments and information schools.¹ Approaches and strategies currently employed will differ depending

___________________

¹ Information science departments and iSchools are generally younger programs, with graduate programs often created before the undergraduate counterparts. The academic background of PhD students is expectedly varied and often includes computing graduates. Further discussion on iSchools is included in Chapter 5.

on the target audience and prestige of the institution doing the recruiting. Active recruiting by the top 10–ranked computing departments will likely be different from lower-ranked institutions. For these top-ranked institutions, the number of applications from international and domestic students can be assumed to be significantly higher, and less than 10 percent of applications are likely to be accepted.

For most computing departments with active research faculty, recruiting domestic students with clear potential for success in a PhD program is a challenge. While such an institution may have a large undergraduate program in computing and graduate more than 500 students per year, the number of domestic students interested in pursuing a PhD is relatively small. The most talented students, even ones who have a successful research record, are often interested in pursuing industry positions due to the increased financial support, access to top computing resources, and potential for career advancement soon after graduation. Many departments have greater success in recruiting top international students by building relationships with well-known international institutions and having faculty with ties to these institutions, or the country, serve as primary contact points. Notably, excellent students may not have a traditional computing background, including those from interdisciplinary backgrounds, informatics, and computational biology, chemistry, or physics.

Many departments are working to actively recruit students to their PhD programs and using recruiting strategies that focus on individual students, but this approach may be unsustainable for long-term success. Rather than focusing efforts on specific students, supporting summer research programs that bring undergraduates to campus to interact with faculty and graduate students and provide information about the doctoral program during their research experience may be more effective. Additionally, recruiting that focuses on highly ranked undergraduate institutions can severely limit the pipeline and misses out on a large number of highly qualified candidates from other, less well-known undergraduate institutions. Instead, departments can create sustainable partnerships with teaching-focused institutions, such as primarily undergraduate institutions, master’s-only institutions, minority-serving institutions, and other 4-year colleges. Finally, computing departments should be pursuing some sort of active recruitment for success in reaching long-term goals of increased domestic doctorate production and broadening participation in computing. In some cases, this recruiting can begin at a department’s home institution with students who have taken introductory and upper-level computing courses or recruiting at current graduate students’ alma maters (see Recommendation 2). The University of Michigan “Recruit-at-Home” Program provides funds for current computer science and engineering graduate students to apply to travel to their alma maters to present a recruitment talk on behalf of the department. Further recruiting outreach can be done at virtual recruiting events or informational sessions at meetings or

conferences, including Grace Hopper, Tapia, and Association for Computing Machinery (ACM) Special Interest Group Conferences.

Preparing the PhD Application

After investing the time and effort into raising research awareness, expanding undergraduate research opportunities, and actively recruiting for doctoral programs, departments can play a key part in preparing their own students to pursue and successfully apply for a PhD. To improve the domestic PhD pipeline, departments, research advisors, and other mentors could offer critical support as their students prepare for a doctoral degree, as preparing a strong application and selecting programs that match a student’s interests is not a simple task.

Undergraduates often underestimate the effort required as well as the competitiveness of the PhD application process. Although acceptance rates for doctoral programs in computing are not widely available, it can be assumed that top programs will have acceptance rates in the single digits. For the majority of programs at R1 institutions, acceptance rates for domestic applicants can range from 15–30 percent, while rates for international applicants are often lower. To improve the chances of being accepted to a doctoral program, students should be encouraged by their departments to apply to a broad list of schools.

While the details of the PhD application may vary slightly between institutions, the major components remain the same, including academic transcripts, statement of purpose, and letters of recommendation. Notably, not all programs require Graduate Record Examination (GRE) scores. A successful application must (1) convince a reviewer that the applicant has the potential for success in their PhD program, (2) match the research interests of a faculty member as well as their willingness to supervise a new student, and (3) compete successfully with strong applications coming from all over the world. The majority of computer science PhD programs provide financial support for all PhD students (as teaching assistant, research assistant, or fellowship); however, the lack of sufficient financial support can also be a reason for rejection from a program. Faculty can provide valuable insight in identifying programs outside top-rated departments with highly recognized researchers in an area of interest.

Computing departments and academic institutions can provide support for their undergraduate students applying to PhD programs by proactively providing advice and guidance on how to obtain strong letters of recommendation, examples of good research statements, and feedback on prepared application materials. This could be as simple as having undergraduate academic advisors and computing departments provide basic information on the process, as well as dos and don’ts of preparing a PhD application. Also, departments should inform students preparing applications where writing

help is available, including university writing centers, workshops focused on preparing graduate school applications, or online resources with examples of good versus weak personal statements.

The letter of recommendation is an essential component of the application to doctoral programs and should adequately reflect the applicant’s potential for academic and intellectual success and future achievement in graduate school and beyond. Students should be encouraged to contact faculty with whom they have done research or taken advanced courses. For students who may not have close relationships with their research faculty, departments can provide additional support in choosing letter writers who will best reflect their potential.

For an institution to best ensure its students’ success when applying to doctoral programs, students should be encouraged to seek out non-local individuals who can provide help, advice, and feedback on their application. Many institutions also support graduate-student led initiatives where prospective applicants can receive feedback on their application before applying. For example, the Graduate Application Assistance Program at the Massachusetts Institute of Technology pairs eligible applicants with current doctoral students to receive feedback on personal and research statements. Similar programs exist at Carnegie Mellon University, Brown University, the University of Massachusetts Amherst, the University of Washington, The University of Texas at Austin, the University of California, Berkeley, and the University of Pennsylvania.

At non-research focused institutions, it may be difficult for students to find someone who is knowledgeable and available for advice on preparing strong admissions materials or selecting which schools to apply to. Often, informational materials, best practice documents, or examples of successful admissions packets may not be sufficient. Having a mentor throughout the application process can be critical for student success. This can include a knowledgeable faculty member at another institution, a faculty advisor from a previous research experience, a mentor made available through a national, often area specific program, or a graduate student-led program providing feedback and advice on pre-application materials.

Individuals in the workforce who completed undergraduate education some time ago face unique challenges throughout the PhD application process. Many no longer have ties to faculty they can ask for advice and they may struggle finding appropriate individuals to write strong letters of recommendation. At the same time, they are likely to have developed useful skills in the workforce that translate into a productive researcher. The National Science Foundation’s (NSF’s) Computer and Information Science and Engineering Graduate Fellowships (CSGrad4US) Fellowship Program (Box 5-1) is a 3-year fellowship and mentoring program targeted at individuals returning from the workforce and provides individual coaching to help recipients prepare a strong application.

who is admitted into their laboratories. Individual faculty in a research area in which a large number of applications are received may only selectively review applications, while faculty in less in-demand areas may also review applicants in adjacent areas. Faculty will agree to serve as a student’s advisor if funding is available and the candidate is a good match with respect to such factors as preparation and interests. One consequence of this distributed admissions model is that some applications may not receive the attention they deserve, especially for students with a non-traditional computing background.

Indeed, some departments operate a parallel process to review promising students who have not identified a research area, evaluate domestic applications, and review applications viewed as deserving additional consideration. In other cases, all the admission decisions are made centrally by departments, or departments select only a subset of candidates for evaluation by individual faculty.

Admission to top doctoral programs is highly competitive. For example, for the 2022–2023 academic year, the Department of Computer Science at the University of Washington received about 3,000 applications, admitted 150 of the applicants (5 percent acceptance), and expected 50–60 of them to enroll as PhD students (University of Washington 2024). The acceptance rates at many other institutions are higher but institutions are still selective, with average acceptance rates of 10–20 percent being typical (Samantaray 2022).

Traditionally, admissions decisions have focused on GRE scores, grades in required undergraduate computing courses, possession of a computing undergraduate degree, academic pedigree, academic achievements in advanced courses, and prior research experience. Prior research experience should be considered beyond publications, which are not always an indicator for potential success. Setting aside the details of selection processes, evidence shows strict admissions thresholds are poor selection methods for the most capable students and show weak correlation between the test and student success in science, technology, engineering, and mathematics (STEM) fields (Miller and Stassun 2014).

Similarly, selection based on prior research experience, by definition, negatively impacts students without undergraduate research experience. Notably such selection disproportionately affects students in computing programs that do not traditionally send graduates to PhD programs and first-generation students.

At the same time, student decisions about where to apply are often made difficult because expectations for successful applications are unclear and the admissions decision-making process is opaque. Knowledge gaps include the attributes of a strong applicant within the context of the institution, list of faculty seeking doctoral students, and the overall competitiveness of a department.

One consequence is that a small set of schools receive the bulk of applications, threatening to overload their decision process, while many other schools receive too few applications or have application pools that are even more highly skewed sociodemographically. Another is that many applicants are summarily rejected, with decreased likelihood of applying to PhD programs again, either because their application did not highlight their strengths or because they overrated the strength of their application because they were unaware of what realistically was required to be a strong candidate for a given program.

Students who are not accepted generally do not receive useful feedback on why their applications were not successful. For example, they may not learn whether there was some weakness in their application or simply that there were simply too many qualified applications for a small number of slots. The lack of specific feedback or transparency about the admissions process deters candidates from reapplying. This is particularly true for domestic students who have many better paid options available to them in industry.

Some computing departments have taken steps to address these challenges and increase the number of domestic students whose applications are successful. These include the following:

• Creating paths from master’s to doctoral programs, which provides opportunities, especially for domestic students, to explore whether they wish to pursue graduate school and research. Offering teaching assistant positions to such students eases financial considerations and helps broaden participation.
• Creating multi-stage review processes in which candidates who may have had limited access to computer science research opportunities and/or atypical trajectories, but have otherwise excelled, are given a second review (Box 5-2).
• Broadening recruiting of students with strong academic backgrounds—including interdisciplinary programs that include computing, adjacent fields (such as information science for an application to a computer science program), or other STEM degrees.
• Establishing graduate-level bridge courses to allow students with gaps in their education to build up the needed skills and knowledge. This is often preferable to assigning them to take undergraduate courses.
• Allowing otherwise promising students who are not accepted in a doctoral program to start a master’s program with the option to transfer upon successful completion of coursework.
• Providing constructive feedback to those with unsuccessful applications and encouraging promising candidates, especially from groups for which departments seek to increase enrollment, to reapply.

• Provide financial support during the first year to attend a conference in any area.
• Offer research area rotations for students unsure about their research topic or advisor.
• Allow students to defer admission for a year.
• Build community among admitted applicants through cohort activities.

Recommendation 4: Universities and computing departments should adopt a more holistic approach to engaging, encouraging, and evaluating applicants to doctoral computing programs, especially for domestic students with limited research experiences, holders of degrees in non-computing-related disciplines, and those who have not identified a preferred research advisor.

Recommendation 4-1: Doctoral programs in computing should strive to make the admission process transparent. This includes providing the following to prospective applicants: (1) basic data on the number of applications and admission rate, (2) information on what strong applicants look like, and (3) a clear explanation of program logistics, including but not limited to funding opportunities and deferral possibilities.

Recommendation 4-2: Admission decisions should be the result of a hybrid evaluation approach that includes individual faculty selecting students matching their research interests as well as programs reviewing applications from promising individuals who have not identified a research area or who have non-traditional career pathways.

Recommendation 4-3: Doctoral programs in computing should inform all applicants of the admissions decision. When feasible, non-admitted but promising applicants, especially domestic students, should be given constructive feedback on their application and be encouraged to reapply.

Recommendation 4-4: Computing departments should consider developing transitional programs that allow additional time to complete course prerequisites for otherwise promising students who do not meet traditional admissions criteria, especially domestic students. These programs, which might take the form of a transitional master’s program,

should offer faculty advising and financial resources, such as teaching assistantships, to help manage the transition phase.

An increasing number of universities have created pathways into computing master’s programs for students who did not study computing as undergraduates, which also creates an opportunity for students to take advanced courses or gain research experience to prepare them for admission to a doctoral program.

INCREASING FINANCIAL SUPPORT

A major factor with respect to both recruitment and retention of graduate students is financial support. Unlike other STEM fields, such as chemistry, biology, and physics, where a PhD is necessary for advancement in a research career, lucrative positions are often available to recipients of undergraduate degrees in computing. The majority of computing departments provide financial support for their PhD students in the form of teaching assistantships, research assistantships, and fellowships.²

Fellowships provide doctoral students with the most flexible kinds of support for their studies. The length of fellowships ranges from 1 year, typical of the fellowships provided by departments and institutions to recruit top candidates, to 3 years, as in the case of NSF and some Department of Energy (DOE) fellowships. Most departments and institutions have internal fellowships available to attract top students as well as to strong students during their studies. Major computing companies offer fellowships aimed at broadening participation; some are offered globally, and some are targeted at certain research areas. The total number of awards is small, and institutions are often limited to making a small number of nominations.

The NSF Graduate Research Fellowship Program (GRFP) makes about 2,500 awards per year supporting outstanding graduate students across all STEM fields. Since 2019, approximately 8 percent (150–180 fellowships annually) of all NSF GRFP fellowships have been consistently awarded to computer and information science and engineering (up from 5 percent in 2016). Despite a dramatic scale-back of the program in 2025, 207 of the 1,500 fellowships (13.8 percent) awarded went to computing disciplines. Notably, due to funding uncertainty, NSF has reduced the number of fellowships by a quarter compared to previous years. Despite cuts to this program overall, the number of students receiving support in computing has increased in 2025.

___________________

² In a compilation of 50 R1 institutions, 49 of them provide financial support for all PhD students. See E. Berger, “CSStipendRankings: PhD Stipend Rankings,” https://csstipendrankings.org, accessed May 17, 2024.

The Graduate Assistance in Areas of National Need model supports the development of a strong domestic talent pool in computing and related disciplines by providing institutional grants that fund fellowships for qualified graduate students. Focused on U.S. citizens and permanent residents with financial need and academic promise, the model enables universities to recruit and train future researchers and educators in high-demand areas such as computer science, data science, cybersecurity, and artificial intelligence (AI).

In addition, CSGrad4US, managed by the Computing Research Association for NSF, aims to increase the number of diverse domestic graduate students pursuing graduate study in computing by helping bachelor’s degree holders return to academia (see Box 5-1).

Most students are supported by research assistantships (RAs) and fellowships, and faculty research grants frequently include support for graduate students. The support is tied to the research funded; deliverables and flexibility in the research topic depend on the expectations of the funding agency. For DOE and Department of Defense funding, the expectations and research details can be quite specific and include visits to the agency, giving demonstrations, and preparation of quarterly reports. Research of RAs supported by NSF grants is generally more flexible. RA support is less stable than a fellowship as the support is tied to the length of a grant and is generally tied to a particular advisor.

Finally, a smaller share of students are supported by teaching assistantships, especially those admitted without a specific research area. Some schools require students to serve as a teaching assistant for at least one semester to meet graduation requirements.

To be competitive, admission letters often include offers for multiple years of support (3 or 4 years is common) subject to good progress in the program. Departments that admit PhD students without financial support often direct such students to other employment opportunities on campus such as teaching assistantships in computationally oriented departments or in general computing services units.

There have long been concerns as to whether graduate student stipends are adequate. A variety of factors determine the stipend a student receives. However, many of these factors often mean that increases in stipends do not track with inflation or other cost of living factors. Externally funded fellowship stipends are set by the funder and are uniform across institutions. Stipends for institutionally funded fellowships, RAs, and teaching assistantships are determined by a combination of federal and institutional policies. Some institutions establish uniform stipend levels while others leave this to individual departments. Some institutions maintain parity between teaching and research assistantships while others do not. At some institutions, stipends are subject to collective bargaining agreements.

Even the highest stipends just provide a reasonable standard of living for a single student with no dependents, and many institutions fall well short of this goal.³ Federal research funders and institutions have in some cases taken steps to help close the gap. For example, in its fiscal year (FY) 2023 funding opportunity, the DOE Office of Science announced a significantly higher new minimum annual stipend of $45,000, excluding benefits. It is intended to provide a “fair and equitable wage sufficient to allow a reasonable standard of living” (DOE 2023). Other funding agencies have made efforts to increase the minimum annual stipend for fellowship opportunities, including NSF’s GRFP, which also increased the stipend to $37,000 for FY 2023. A number of universities, often as part of collective bargaining negotiations, have also increased graduate stipends in recent years.

However they are set, and despite efforts to reach a reasonable standard of living, graduate stipends fall far short of salaries offered to bachelor’s graduates in computing fields. For example, ZipRecruiter reports an average starting salary for new bachelor’s graduates in computer science of $107,500, nearly double the highest reported doctoral student stipends (ZipRecruiter 2024). Domestic bachelor’s degree holders in computer science opting for doctoral studies over employment thus will experience a significantly lower standard of living while in school and at best a long payback period for their investment. The payback opportunity is further diminished because graduate students do not receive retirement benefits.

In addition to low stipends, many doctoral programs do not provide many other benefits, such as parental leave, that would otherwise be available if they were to pursue full-time employment in industry. Increased financial support would be required to bridge the gap for students with dependents or who otherwise financially support family members.

If federal research budgets remain flat, a decision to increase stipends would mean supporting fewer graduate students, suggesting that new sources of support and new partnerships will be needed. Given the computing industry and government laboratories’ collective interest in increasing doctorates in computing, they are encouraged to invest in supplemental stipends or support for students looking to go back to complete a doctoral degree.

Recommendation 5: All institutions supporting fellowships and assistantships to doctoral students in computing should explore available options to increase the amount of financial support provided to each student in order to allow a reasonable standard of living, make graduate study in computing

___________________

³ E. Berger, “CSStipendRankings: PhD Stipend Rankings,” https://csstipendrankings.org, accessed May 17, 2024.

more attractive relative to employment, and make graduate study more attractive to domestic students.

Recommendation 5-1: Academic institutions and funding agencies should increase the minimum stipend for graduate students in computing and index it to the inflation rate.

Recommendation 5-2: Funding agencies should consider establishing fellowships to provide supplemental funding for benefits such as housing allowances, health insurance for dependents, parental leave, and childcare.

Recommendation 5-3: Academic institutions and funding agencies should pursue the development of partnerships with computing industry leaders to provide supplements to department stipends or augment fellowships.

Most doctoral students in computing rely on federally supported RAs. It is thus impossible to both sustain or grow the number of graduate students (Recommendation 1) and provide more resources per student (Recommendation 5) without either growth in federal research investment or new partnerships that expand or augment RAs.

Industry partnerships are likely an essential ingredient for implementing Recommendation 5. The combination of federal and institutional resources will probably be insufficient to increase stipends to the needed level and because university rules, practices, and labor agreements may dictate uniform stipends or otherwise limit variation across an institution. Federal research budgets have not seen significant growth in recent years and future growth is far from certain. A partial solution in such an environment would be for industry (both the computing sector as well as other sectors that rely on computing talent) to step up and provide more support to help meet their immediate (i.e., new PhDs) and long-term (i.e., faculty to teach future employees) needs.

RETENTION PRACTICES

Another way to maximize the number of doctorates earned in computing is to increase completion rates. Earning a PhD in computing is a challenging endeavor and takes a median of 7 years from the start of graduate school. Though not a new concern, student attrition has been a long-standing issue in computing education and an opportunity for

growth. For example, across all University of California system campuses, the 4-year retention rate for students pursuing doctoral degrees in engineering and computer science is 75 percent, while the 8-year completion rates drop to 67 percent for cohorts entering between 2009 and 2014 (University of California 2025).

Many factors may lead PhD students to leave before completing their degrees. Some academic challenges include wildly different breadth of requirements and qualifying structures across programs, with some requirements having the end result of weeding out students who want to complete the PhD. Additionally, students with non-traditional preparation for graduate school may have difficulty completing requirements to comply with timelines established by programs for students with traditional preparation. Limited research funding availability may result in students being required to serve as teaching assistants, making it difficult for them to make progress on completing their PhD quickly and causing them to leave the program. Finally, insufficient mentoring or lack of departmental community support and culture can often be a cause for students, especially domestic students, to leave a program early. Improving mentoring and community culture in computing departments are key interventions to improve domestic participation.

Mentoring

Because the process of earning a PhD involves a multi-year apprenticeship working with a faculty research advisor, the relationship between the student and advisor has an outsized impact on a student’s progress, well-being, and likelihood of completion. Therefore, effective and supportive mentoring and advising plays a significant role in a student’s success (NASEM 2019).

Unfortunately, there are many issues that can arise that cause conflict between an advisor and a student. Lack of communication on expectations and limitations, lack of perceived research progress, miscommunication on next steps and goals, working style conflicts (including too little or too much interaction), changes in the research focus of the advisor or student, insufficient support through the failures or challenges of research, and insufficient professional support and networking can cause strife between a student and an advisor.

Mentoring in computing departments can be improved by providing training for faculty members on effective mentorship, including expectations for faculty as research advisors, teachers, and mentors. For example, training materials could be adapted from work by the Center for Improvement of Mentored Experiences in Research or the graduate student mentor training program developed by CRA’s Undergraduate Research to PhD program. Larger research departments that hire multiple faculty members each year could create cohorts of junior faculty members who go through this training and do peer-mentoring among themselves over their first 3 years as faculty members.

Of particular importance is ensuring the success of admitted students who have not yet identified a research area. Experienced faculty mentors can play a pivotal role in helping students select appropriate courses and identify potential advisors who align with their interests. Mentors can also introduce students to potential allies, such as peers and faculty members, who can answer questions and assist them in navigating academic milestones. Additionally, facilitating opportunities for students to engage with research laboratories, industry partners, or nonprofit organizations can be instrumental in helping them identify a suitable research area.

Fostering Supportive Research Community

A collaborative and supportive research group and departmental environment can also contribute to student retention. Such efforts may be especially useful options for retaining domestic students, who have a wider range of options available should they leave before program completion. For example, interventions can include dedicated student spaces (such as libraries or lounges), student involvement in department governance, and computing graduate student associations.

Organizing a departmental faculty research day, or aligning with a campus-wide research event, allows both undergraduate and graduate students to learn about the various research activities within the department and associated research centers. Seminars presented by senior doctoral students can provide new students with the chance to learn about the breadth of research being conducted across the department. Moreover, providing travel funds enables students to connect with a national network of researchers, further aiding in their professional development and exploration of research interests. Such strategies help departments create a supportive environment that fosters the success and growth of students who are still in the process of defining their research focus.

Recommendation 6: Computing departments, in concert with professional organizations, should improve mentoring and advising for graduate students, especially for domestic students, to improve student retention and success.

Recommendation 6-1: Professional organizations and the Computing Research Association should create and disseminate materials on mentoring best practices for adoption within computing departments. Computing departments should implement and adopt these mentoring practices department wide.

Recommendation 6-2: Computing departments should provide training for all faculty members on how to mentor students effectively, including providing structured guidance for all new junior faculty members during their first 3 years. This should include emphasizing the role of junior faculty members as research advisors, teachers, and mentors, providing concrete advice on how to be successful in those roles.

Recommendation 6-3: For students who enter a doctoral program without a matched research advisor, computing departments should initially assign a department advisor to provide support and guidance and assist students in finding an advisor.

Recommendation 6-4: Computing departments should periodically assess the state of graduate student support, mentoring, department and research group climate, student experience, and the impact of these factors on graduate student retention.

BUILDING THE FACULTY NEEDED TO EDUCATE THE NEXT GENERATION OF COMPUTING DOCTORAL RECIPIENTS

Industry and Government Career Opportunities

Students who successfully complete their doctorate in computing have a wide range of occupations open to them, particularly those who focus on high demand specializations, such as AI and cybersecurity. As discussed in Chapter 2, currently 56 percent of graduates pursue positions in industry, 3 percent opt to continue their careers in government laboratories and nonprofit organizations, and 3 percent choose other career paths. Of the remaining graduates, 38 percent choose to pursue academia positions immediately after graduation, with 11 percent choosing tenure and 6 percent choosing non-tenure-track faculty positions and 21 percent becoming postdoctoral researchers.

Doctoral recipients in computing have a wide array of employment opportunities in industry, including at start-ups, large technology firms, the financial sector, health care, and so on. The attractions to working in industry include lucrative salaries and perks; access to emerging technologies, large computing resources and real-world, large data sets; the opportunity to lead and participate in exciting foundational or applied research; and the opportunity to invent new products and applications impacting the lives of millions. Many of these positions are located in major technology hubs that afford a high quality of life as well as professional advancement and employment opportunities.

There are, however, downsides to working in industry. Even the most prosperous businesses may reduce staffing and rebalance their product portfolio, leading to unexpected job loss and corresponding psychological uncertainty. Additionally, computing resources may be abundantly available to those who work on only the highest prioritized projects but severely limited to others. The opportunity to publish research results may be restricted, especially for those working closely with products and services. Historically, only researchers working at academically oriented laboratories such as Microsoft Research could freely publish results. However, these norms are shifting, and more researchers in industry now regularly share their work with the academic community, particularly for basic research in fields like AI, and particularly those researchers at the largest technology companies. One general pattern is that researchers working more on basic research topics are more likely to be permitted to publish their findings, while those working on topics closely related to specific products and services are less likely.

Positions in national security at national laboratories, federally funded research and development centers, university research organizations, defense contractors, and in the federal government provide access to substantial resources, including funding, advanced technology, and specialized tools. Additionally, these positions provide opportunities to work on projects that directly contribute to national security, public safety, and the well-being of citizens, and also provide a strong sense of purpose and fulfillment. Compared to industry positions, government positions offer greater job security and stability. However, government positions generally tend to have salary constraints, yielding lower salaries than industry, and lower even than many academic institutions. Also, many positions often require security clearances, which can be a lengthy and rigorous process. Maintaining security clearances may also limit career mobility outside of government or national security work.

Recommendation 7: To encourage their employees to pursue doctoral degrees in computing, industry and government laboratories should provide fellowships for graduate study in exchange for a commitment to return for a period following completion.

Owing to the large portion of undergraduates with degrees in computing that pursue positions in industry following graduation and perceived lack of doctoral-level talent available to pursue positions in industry and government, especially in high-demand areas such as AI and national security, these institutions should invest in their own employees to help meet their internal demand for doctoral recipients. This support could take the form of fellowships to cover graduate student stipends, release from work to pursue doctoral degrees, or resources to supplement and encourage research toward a doctoral degree.

Academic Career Opportunities

As compared to other career paths in industry and government laboratories, academia has a variety of benefits, particularly job security in the form of tenure or long-term teaching contracts. For faculty members with research expectations, academia can offer freedom in what research and projects they work on and have substantial impact on the research community. These positions can give individuals the opportunity to mentor and serve as role models for the next generation of individuals. Logistically, another benefit is that employment opportunities exist all over the country in different types of cities and towns.

Disadvantages of academia that may dissuade new graduates include earning lower salaries than offered in industry and having decreased ability to frequently change employers. Over the past 15 years or so, increasing undergraduate enrollments in computing fields has placed increasing demands on faculty time, resulting in a shift in the research versus teaching equation that may discourage exceedingly research-focused individuals from entering academia. Another consideration impacting career choices is that faculty may have limited access to compute resources and large data sets at some institutions, restricting what sorts of research projects they can explore and the research impact they can have. The need to constantly obtain research funding and publish results in selective conferences and journals may also discourage some graduates from exploring academic careers and drive even senior faculty to industry positions.

Academia also offers an array of different academic institutions and types of faculty positions to match the different career and personal goals of graduates. Some motivations for selecting between the different opportunities may include a person’s desire to conduct research and have research impact, a desire to teach and mentor undergraduate or graduate students, or support the development of students with a specific background (e.g., first-generation college students, students with interdisciplinary or non-computing academic experience). Additionally, geographic or personal constraints may also impact what type of school and position graduates embrace after graduation. Two important constraints on the availability of academic positions are (1) the smaller number of institutions in higher education compared to companies in industry and (2) their broad geographic distribution.

Research-focused faculty positions at R1 universities, characterized by very high research activity, and non-R1 research universities differ mainly in research expectations. R1 universities offer tenure-track positions that attract candidates due to the focus on research impact, autonomy in research agenda setting, high research productivity and external grant acquisition rates, and lower teaching loads. R1 institutions often have higher salaries and the most selective student bodies. Non-R1 institutions appeal to graduates seeking slightly lower research expectations regarding publication rates and funding.

These institutions might also attract individuals based on geographic location or specific student demographics.

Teaching-focused faculty positions are available across various academic institutions, including research universities and non-doctoral-granting 2- and 4-year colleges. Graduates interested in these roles are drawn by the focus on teaching and pedagogy, and mentoring. Expectations for research and scholarly productivity vary widely.

Teaching track faculty at research institutions will likely be attracted by the opportunity to teach at scale, with low or no research expectations, although scholarly contributions are frequently still expected. They may also appreciate the use of long-term, repeatable contracts instead of the research-centric tenure system and the fact they are granted academic rank and a variety of voting rights in the department and institution, although their voice and impact may be reduced compared to research-focused faculty. Teaching-focused faculty at small colleges, alternatively, may appreciate the opportunity to teach courses with a low student-to-faculty ratio, equally balance their effort on teaching and research, and to engage undergraduate students in research. Graduates drawn to master’s-only institutions may enjoy the ability to work on research with both undergraduate and graduate students without the high research demands of research-track positions and the ability to focus on teaching, including at scale.

As discussed in Chapter 3, increasingly, some faculty members have a foothold in both academia and industry. These hybrid positions can vary with 20–80 percent of their time split between industry and academia, depending on individual negotiations and institutional policies. These dual appointments offer both advantages and disadvantages for individuals and the organizations involved. Faculty in industry positions tend to have greater access to computing and data resources, cutting-edge technologies, and real-world, at-scale research challenges. At least those not in managerial roles also have fewer non-research institutional demands on their time than full-time academics. They are also paid more for the industry portion of their time.

Although departments and students may benefit from the faculty member’s access to cutting-edge industry research, there are also drawbacks. Faculty members with dual appointments have less time to devote to nurturing graduate students and other departmental responsibilities. They also are often perceived as not carrying their full weight by fellow faculty members.

Addressing Faculty Hiring Demands

With doctoral degree recipients in computing increasingly taking more lucrative industry positions rather than tenure-track academic positions, and with academics increasingly pursuing joint industry-academic appointments, it is becoming harder to sustain

the robust faculty ranks needed to produce the next generations of PhDs. Moreover, to the extent that faculty in “hot areas” (AI is a current example) are more likely to move to industry, it becomes that much harder to produce PhDs in the very areas for which demand is highest. Additionally, some roles cannot be filled by international students who have earned their PhDs, restricting the ability of new PhD graduates to engage in some of these positions, limiting the supply of new hires to these areas, especially at the national laboratories and in other national security roles.

Although faculty members often can move from academia to industrial positions, it is much harder for individuals to transition from industrial positions back to academia, often because they are unable to publish extensively when in industry and do not have access to opportunities to gain the teaching experience expected when hiring senior candidates. Consequently, doctoral graduates opting for industry positions will generally find it difficult to move back into academia.

Finally, the variation of compensation between the different types of positions can constrain which route individuals may take, particularly for new graduates who have student loans or who have family obligations. These individuals may feel compelled to choose career paths based on finances, limiting the number of individuals that are entering from lower-compensated fields (i.e., academia and government laboratories), resulting in a myriad of downstream effects.

Recommendation 8: The National Science Foundation and other federal agencies that support computing research, as well as industry, potentially through government–industry partnerships, should create new fellowships and assistantships in computing that include requirements to serve in faculty positions for a specified period.

For industry partners, the benefits include sustaining and growing the faculty that will educate future employees, growing the supply of talent in high-demand areas. Such fellowships and assistantships could be targeted in various ways including for faculty positions at non-doctoral-granting institutions, and for subfields that are in especially high demand. There are several examples of federal computing fellowships with requirements for service, including the CyberCorps® Scholarships for Service and the Artificial Intelligence Scholarships for Service Initiative.

Ensuring adequate production of PhD recipients requires a sufficient academic faculty. Faculty hiring and faculty composition, including in areas that are currently highly sought after, directly influence PhD production.

Recommendation 9: Computing departments and academic institutions should take a broader view of what merit is considered in faculty hiring and promotion, not just the number of publications or PhD pedigree, but other attributes including but not limited to commitment to best practices in mentorship, preparing students for their future careers, and research that builds bridges with other fields.

* * *

Industry and academic hiring practices have distinct impacts on the supply of doctoral recipients in computing, akin to the metaphor of eating seed corn versus planting more crops. Industry hiring focuses on more immediate needs and tends to emphasize the application of research to real-world problems. This fosters innovation in the short term but depletes the pool of academic talent, leading to a shortage of educators and a potential decline in the quality of future computing professionals. Over time, the emphasis on short-term industry demands may also detract from long-term foundational research critical for future innovations.

In contrast, academic hiring prioritizes sustainable growth by recruiting doctoral graduates to mentor and train future generations. However, lower compensation and funding constraints make academic positions less attractive. To balance these approaches, collaborative programs, increased funding for academic careers, and industry–academia partnerships are essential. Fostering a symbiotic relationship between industry and academia can ensure that the field of computing continues to thrive with a sustained supply of highly qualified doctoral recipients.