Institutions facing significant increases in demand for computing courses have several options for how to proceed, as discussed in Chapter 2. In the near term, the options are to accommodate the demand, or to restrict access to majors or courses. Over time, these decisions may have an impact on the nature of the program and the institution’s character and diversity, even as demand continues to change. Thus, it is important for institutional leaders to take the long view, and to act strategically in ways consistent with the institution’s values and mission.
This chapter discusses institutional resources and constraints, ways in which institutional character and goals can inform decision making, and a range of options for dealing with the computer science (CS) enrollment increases and underlying drivers, as well as the risks and advantages of specific strategies.
For academic institutions, responding to substantial and rapid changes in student demand is challenging. This is not unique to CS—and CS is not the only academic field that has experienced rapid enrollment increases—but the rate and magnitude of the recent change in demand for CS courses, from both majors and non-majors, appear to be extreme. Furthermore, the relatively small number of CS Ph.D.s pursuing an academic career, coupled with the broad opportunities for CS faculty in the private sector, has made faculty hiring especially difficult, which limits what institutions can do and further exacerbates the challenge.
The key resources in support of a department’s teaching mission are faculty, teaching staff, teaching assistants (TAs), facilities, and support staff, all of which are dependent on a department’s budget. An institution’s capacity to respond to
increased student demand for computing courses will be limited by constraints on any one of these resources, as discussed here.
Faculty and Teaching Staff
The most significant resource constraint in CS departments is the faculty and professional teaching staff. This includes tenured and tenure-track faculty, teaching faculty, lecturers, instructors, professors of practice, and other titles for faculty with a primary teaching role. Tenured or tenure-track faculty have multiple additional responsibilities, such as research and institutional or other professional service requirements. Teaching faculty with an academic rank may have their own research program, often in the area of CS education. All teaching staff typically take on a variety of additional roles, including managing other teaching staff or TAs, or advising, tutoring, or mentoring students. Faculty workload can be increased to respond to greater demand, but only so far before additional instructors become necessary to accommodate increasing numbers of students.
Academic growth in computer science has historically aligned with times of industrial opportunity, as is the case presently. Thus, even if there is institutional commitment to expand CS faculty, it may be difficult to attract competitive candidates, as discussed in Chapter 3. This challenge may be exacerbated when faculty are recruited into a climate of extremely high teaching loads or high student stress.
The mix of faculty types—not just the number—is a significant factor that relates to institutional values and mission. CS departments at research universities have primarily tenured and tenure-track faculty. Adding large numbers of faculty not on the tenure track can create administrative and cultural issues, including the perception of “classes” of faculty, though it may help improve the diversity of instructors which can in turn have positive impacts on student diversity. While this is an especially acute issue at research universities, it is a potential problem at all institutions and underscores the importance of considering long-term consequences and institutional values.
TAs—drawn from either the graduate or undergraduate student bodies—are crucial for the successful delivery of education in many institutions. TAs lead small lab sections and discussion sessions, prepare projects, grade projects and assignments, and provide individual instruction.
Ph.D. students are an experienced and preferred source of TAs in Ph.D.granting departments, providing the primary source for help with higher enrollment courses. Ph.D. students in computing are typically supported through non-TA sources of funding such as fellowships and research assistantships. Departments have to balance the need for an experienced TA against the adviser and student
goal of making progress on research. As CS enrollment has surged, the Computing Research Association (CRA) Enrollment Survey suggests that graduate students have increasingly been asked to teach courses in departments with CS Ph.D. programs.1 However, the increase in enrollment has not been matched with an increase in doctoral students to serve as TAs.
As illustrated in Chapter 2, master’s degree production in CS is an order of magnitude greater than Ph.D. production, and has increased rapidly in recent years as students have sought skills in high demand in industry. Many institutions have relied upon these master’s programs as a source of revenue, and over 55 percent of all master’s degrees were awarded to foreign students in 2015. Although CS master’s students are commonly admitted with no promise of financial support and they have a higher concentration of coursework, drawing TAs from the master’s student body is crucial for many departments.
Undergraduate CS students are another source of TAs.2 As enrollment increases, the size of the undergraduate TA (UTA) pool scales linearly with the undergraduate enrollment. At many institutions that do not offer graduate programs, UTAs are already supporting instruction, but this is an increasingly common practice at research universities as well. Serving as a grader, lab assistant, peer instructor, mentor, or tutor can be a valuable part of an undergraduate’s educational experience. These opportunities provide motivated UTAs with a strong connection to the student body and can allow valued interaction with graduate TAs. Undergraduate TAs can also be inspirational to other undergraduates, who may become inspired to pursue academic or teaching careers. Teams of early-stage UTAs require careful management and oversight by the instructor, as well as institutional support and training, and may not be appropriate at all institutions.
Support staff play a crucial role in maintaining a healthy and well-managed learning and work environment, and can play a role in easing the pressure of growth in student demand. The responsibilities and level of support vary by institution, but can include undergraduate advisors and course or technical support staff, who may be tasked with addressing exceptions and points of potential confusion among students.
As the number of students increases, various factors may lead to increases in the demand for staff time and attention. For example, the number of exceptional situations (such as illness, accidents, time conflicts, or family issues) likely
2 See, for example, Figure F3 in CRA (Computing Research Association). 2017a. Generation CS: Computer Science Undergraduate Enrollments Surge Since 2006. Online. Available at http://cra.org/data/Generation-CS/.
increases with a larger student population, as does the potential to be lost in the crowd. In addition, the efforts required for routine course management tasks—preparing exams, supporting assignments and projects—increases with size. With larger class sizes, the demand on teaching and support staff to handle issues such as cheating and harassment likely also increases. Some aspects of staff work exhibit scale easily, while others do not, such as student advising, which requires one-on-one time, and depends on advisors’ knowledge of students.
Facilities, such as classrooms, lecture halls, and computer labs, and associated equipment are also a necessary resource. In computing lectures and labs, computers and projectors are generally a key requirement; students need computers to complete programming and other assignments, whether their own or university-issued laptops or desktops.
Increasing student enrollments present a host of facilities challenges that can impact the character and attractiveness of the program. At some institutions there are simply no lecture halls big enough to accommodate all students, and, when such rooms do exist, they may be so large that they are an impediment to teaching and learning. For example, at the University of California, Berkeley, in the fall 2016 semester, more than 1,800 students initially enrolled in the introductory computer science course for majors, dropping to 1,567 students after the first several weeks. Early in the semester the course was taught in the Zellerbach Auditorium for Performing Arts, which has a seating capacity of 2,525. Even dividing such very large courses into multiple sections can challenge the capacity of larger lecture halls—and such large class sizes can have a negative impact on the students’ experience—or challenge the availability of a larger number of smaller rooms. These pressures are not necessarily limited to introductory courses or non-major offerings; major enrollment is up and non-majors are progressing more deeply into computer science programs at many institutions.
Video capture and replay, or even studio preproduction,3 of large lectures can help to mitigate facilities constraints and help to tailor large lecture presentations to a diversity of students and learning styles. The ability to go back and review particular aspects of a presentation, especially when learning computational techniques or problem solving is involved, allows students to absorb material at their own pace. It can also help to mitigate demands that might otherwise be placed on staff if a student misses class.
Computer science courses often have substantial laboratory components, which are not typically present in large social science and humanities courses, but not as specialized or complex as large biological or physical science courses. With growth in enrollment, more sessions in lab facilities (or more seats during lab ses-
3 This is distinct from online courses, though there may be overlap in methods.
sions, as many students increasingly use their own laptops) must be provided. At the same time, lab sessions may offer a more personalized small-course setting where individual hands-on learning takes place. In this case the quality of teaching assistants and support staff may become even more critical to the quality of the course, as does the lead instructor’s guidance and management of teaching support staff. In addition, content distribution and collection, grading, and approaches to resolving errors and disputes must scale. Of course, all these benefits of labs are contingent on the availability of lab facilities and equipment, which are subject to similar space constraints as are lectures. Expense also obviously goes up with the scale of lab facilities.
In general, tiered instruction—with large lectures, smaller discussion or lab sessions, small group or peer advising, and one-on-one tutoring—can assist with the efficient allocation of resources when dealing with large numbers of students.
One of the reasons that U.S. higher education is the envy of the world is its unique diversity. The nation’s more than 4,000 institutions of higher education range from technical institutes and community colleges that award certificates or associate’s degrees to Ph.D.-granting, research-intensive universities; and from small, private liberal arts colleges with a few hundred students to gigantic flagship public universities with tens of thousands of students. Each has important roles to play in supporting the nation’s need for computing professionals and innovation and CS’s need for the faculty and research that sustain the discipline.
Given this institutional diversity, it is likely that one institution’s response to the growth in CS will be inappropriate—if it is even feasible—for another. A reliance on doctoral student TAs is obviously infeasible at a liberal arts college or master’s institution, for example. Less obvious, but of potentially profound importance to the institutions, are actions that would push against long-standing institutional practices and core values. For example, for an institution that allows admitted students to major in any field they choose, it would be precedent setting, even mission altering, to impose limits on the number of CS majors.
Beyond institutional diversity, there are also important considerations related to student diversity. U.S. higher education institutions value contributions from a demographically diverse group of students and faculty with a range of experiences and perspectives. The offerings and atmosphere of any institution will attract different types of students, and institutional actions—including admissions and enrollment practices—play a major role in determining who is drawn to the institution and to specific programs, including the participation of underrepresented communities, and can impact the success and retention rates of different groups.
Every institution has its own distinct distribution of intellectual endeavors. In the face of increasing demand for a specific discipline, ensuring the vitality of less-populated disciplines might be a priority. Too much migration to popular
fields could decrease the level of interest in the smallest programs to a level below a critical mass. For institutions that prioritize a broad, liberal arts education for their students, requiring early declaration of major, in particular during admissions, could inhibit the tradition of discovery during the first years of a broad education. Institutions that prioritize deep technical training in science and engineering may find it detrimental to the institution’s character to limit student access to experience in computer science, which is increasingly important for a range of science and engineering disciplines. Alternatively, they may want to maintain or increase the academic profile or ranking of the institution by maintaining or introducing high entrance requirements.
It is crucial that institutions be thoughtful and intentional in choosing strategies to respond to growth in computing enrollment. This includes a consideration of mission and values, and the consequences of actions—or inaction.
While the challenges of increasing enrollments in undergraduate computing courses are very real and some institutions are in dire need of relief, it must be noted that resources at academic institutions may be limited. In particular, investments in one area typically mean that cuts must be made to another. Accordingly, strategic decisions about how to respond to the surge in demand for computing must be made based on an analysis of the overall costs and benefits for the institution of the various possible actions.
In this section the range of responses available to academic institutions are reviewed, including those listed in the CRA survey and discussed in the previous section. In general there are four categories of responses available to institutions: (1) limit participation, (2) leverage resources in new ways, (3) grow programs and the resources that feed them, and (4) restructure the nature of computing education within the institution. The categories range from relatively inexpensive and nondisruptive actions to major changes in organization and resource commitments.
The responses are briefly evaluated in terms of their advantages and risks. Because the needs and priorities of U.S. institutions of higher education are diverse, even among members of any given category of institution, the committee does not believe it would be appropriate or effective to prescribe any single strategy or combination of strategies to any given institution or institution type. Instead, institutions must identify how the advantages and risks will translate to their individual circumstances, and decide how to proceed. What is best for any particular institution depends on many factors, including and especially the institution’s mission and vision.
Limit the Number of Majors
Institutions can take actions to limit the number of majors in their computer science program. These limits could be imposed at the point at which students are formally accepted into a major program, which occurs either before (if students must apply directly to a degree program at the time of application, as at Carnegie Mellon University, for example) or after they enter the college or university (whenever the institution requires major declaration). Enhanced restrictions to limit enrollment could be made by imposing a set of threshold performance requirements for entry into the program based, for example, on high school or college grades, grades in particular courses, performance on exams in certain courses, entrance or qualifying exams, or other factors. Alternatively, caps could be imposed with acceptance into the major based upon lottery, first-come-first-served, or some other non-performance-based prioritization.
Advantages: Imposing limits would prevent potentially unmanageable growth in computer science degree programs, and alleviate the associated pressure on departmental resources. Depending on how the restrictions are implemented, limiting enrollment should be administratively expedient. There is the additional benefit of near certainty about student numbers, making it much easier for units to plan.
Risks: Limiting the number of students who may declare as CS majors at any point in the student’s undergraduate experience may cut them off from their true passion, or prevent others from discovering theirs. Depending upon where in the process the limit is imposed, it may also introduce stress or an environment of real or perceived competition for students who desire to enter a CS program, which could discourage participation among underrepresented groups. A ranking-based approach for determining eligibility would promote head-to-head competition based on the chosen requirements, which could cause students to focus on these requirements rather than other educational objectives. This strategy could also affect the climate for faculty, instructors, TAs, and support staff, by requiring them to respond to enhanced student stress, angst, or competitive attitudes, which risks diminishing the talent within the university.
However, some of these challenges could be overcome by introducing the limits at the time of acceptance of prospective students into the college or university, so students would have advance knowledge of whether a given institution will permit them to major in CS or another computing field. At the same time, in the extreme, limiting CS enrollments at all institutions would not be in the national interest, as it would restrict the total production of CS degrees feeding into the economy at a time when demand for computing professions is expected to grow.
Measures to limit major admission or declaration would likely result in reduced demand for more advanced courses or major-only introductory courses,
but would not necessarily limit the number of non-majors who want to enroll in courses. On the other hand such limits could affect the overall climate in such courses, and thus the level of interest in the course.
Limit Course Enrollment
Another action, which is not mutually exclusive with limits on majors, is to simply limit course enrollments. Decisions on whom to allow into courses could be made based upon a student’s major (or intended major), their class (e.g., seniors before underclass students), via performance requirements, by lottery, or on a first-come-first-served basis.
Advantages: This approach will remove the pressure on departmental resources associated with growth in student course enrollments, and would enable more individual attention to be paid to students than would be possible in larger classes or cohorts.
Risks: Again, this approach risks barring students who may have a sincere passion for the field, or who may need the course or associated skills for their major area or intended career path. It also limits the ability of a student population to be exposed to computing. These limitations restrict the potential of a student’s experience, and the measures imposed could make students feel their future is determined by factors beyond their control, especially if students are performing or achieving at the required level but are still unable to enroll in desired courses, which can affect a student’s ability to complete his or her program in four years. Students may try to influence enrollment decisions by pleading with or pressuring faculty responsible for enforcing limits, which can lead to a negative environment for the department and the institution.
This option carries many of the risks associated with limiting majors. In particular, capping course enrollment based on past performance or experience could disproportionately affect women and underrepresented minorities. It could also create a competitive environment among non-majors, which discourages participation and exposure to CS in ways that are becoming increasingly important for a range of disciplines, thus limiting students’ educational experiences.
Redirect Demand for Courses and Majors
Restrictions on course enrollment or declaration of major are often associated with efforts to redirect students into alternative courses or majors. This can be done by communicating (e.g., via advising or in presentations to potential future students) how difficult it is to get into courses or majors and other associated challenges, and encouraging them to consider alternatives.
Advantages: This approach has the benefit of clearly presenting the challenges of the program to the students before they decide to pursue them. It may also prompt students to look more closely at what is best for them.
Risks: This approach could also serve to discourage students who might otherwise enjoy and succeed in computing courses or the major, and close what might have been successful pathways. In addition, telling students who are still deciding which institution of higher education to enroll in about the challenges or limits associated with CS at a given institution may cause them to lose interest in the institution. In particular there is a risk (as discussed in previous chapters) that limiting access to courses could lead to declines in participation of women and underrepresented minorities.
Growth is an obvious response to increased demand, but growth may have its own disadvantages, including potentially large opportunity costs for other university programs and priorities.
Increase Class Sizes
One response to increased course demand is to allow class sizes to grow to accommodate the demand.
Advantages: This has the benefit of allowing all interested students access to the courses that they want to take, and would likely help to avoid the sense of scarcity and competition that comes with the imposition of limits.
Risks: Larger class or lab sizes may or may not align with pedagogical needs or goals, depending on the nature of the course, and could negatively impact learning outcomes. Larger classes will be less agile in meeting individual student needs. They could also affect the student experience further by limiting individualized interactions with faculty and teaching staff, creating the sense of being lost in the crowd, and heightening competitive pressures, which could also have a negative impact on student diversity. A larger number of students per class or lab will increase the workload on faculty or teaching staff and the required management skills may limit who can teach these effectively. Even if additional hires are made, junior faculty with less experience are generally not effective instructors for large classes. In addition, physical resources such as classroom space may be stretched beyond capacity.
Strategies can be deployed to mitigate the negative aspects of large lecture classes, such as recording lecture material to permit student replay at a later time, expanding tutoring or office hours, and optional special sessions to provide further challenge or enrichment. Additional actions may be necessary to manage the burden of larger classes on academic resources. For example, course management staff could help to reduce a primary instructor’s management load. Technology can be leveraged to support grading and course communications, including online forums, and is already quite common at many institutions. Faculty teams, rather than individual faculty, may offer courses collaboratively, enabling greater
scheduling flexibility or specialization. In addition, new pedagogical strategies such as collaborative projects and peer-to-peer instruction (discussed earlier in this chapter) could also ease some of the workload associated with teaching larger classes. New teaching arrangements may require initial investment of time and resources, especially at the onset.
At the same time some aspects of teaching do not scale easily, such as accommodating exceptional situations that arise in students’ lives. The ability to handle exceptions effectively or not can contribute significantly to the climate of a course and the program, either positively or negatively.
Increase the Number of Sections or Courses
Departments may offer additional lecture sections for a given course, or offer similar courses that address the same student needs, rather than simply increasing the class size of one offering.
Advantages: This approach meets student demand without the downsides associated with increased class sizes. If the same instructor is simply teaching additional sections of the same course, the additional time required for preparation of materials is minimal.
Other benefits of scale may derive from this approach. Multiple offerings provide an opportunity to specialize individual sections toward different student backgrounds and the freedom to respond to student interests in a given course. For example, an introductory course with two sections could target students with prior programming experience in one section, and those without in the other. Similarly, the need to offer more sections could provide an opportunity to offer more distinct courses to better explore the specific interests of a diverse group. It could also augment the course schedule by making sections available at alternative times, such as evenings, weekends, or summers, which would make better use of existing facilities, provide benefits to students with nontraditional schedules, perhaps due to commuting distance and job schedules, and additionally help students to avoid conflicts with other courses. In particular, summer offerings not only expand capacity, they can also alleviate difficulties in satisfying prerequisites in the presence of over-enrollment.
Risks: In the absence of additional faculty or teaching staff, an increased number of course sections serving a larger number of students will increase workload in the form of lecture hours, advising time, grading responsibilities, and other individual interactions with students. The question of how teaching load is counted for an instructor teaching multiple instances of the same course needs to be addressed. Proper synchronization between sections of the same course taught by the same or a different instructor is often needed. This approach is also constrained by the number (rather than size) of available classrooms to accommodate additional sections.
Hire More Faculty
Advantages: Increasing the number of faculty or teaching staff makes it easier to offer more courses or course sections, and provides more opportunity for students to interact one-on-one with instructors. Adding faculty also leads to an expansion of research and related activities which benefits both the institutions and the regions that they serve.
Risks: Other than the issues associated with adding faculty in any field—such as institutional budget constraints and the costs associated with a faculty search—computer science faces an acute shortage of Ph.D.s pursuing an academic career, as already discussed in previous sections. In addition, the number of qualified individuals interested in a teaching faculty or other instruction position may be similarly low. The availability of short-term or contract instructors varies considerably, often driven by the geographical location and other employment opportunities around an institution. Furthermore, the same pressures make current faculty more difficult to retain, as they are recruited by both industry and other institutions.
Leverage Resources Creatively
Rather than maintaining the status quo and continuing to conduct computing instructional programs in the manner of past decades, programs have the opportunity to embrace change and make use of existing resources in new and more flexible ways. Several examples follow.
Use Technology for Teaching
Technology for education is rapidly growing, affording many potential pathways for responding to the demand for CS coursework and degrees. In addition, computer science material is some of the most amenable to technological enhancement, and CS instructors are likely to have the knowledge to wield and even develop such technology. This includes online course materials such as lecture videos and automated administration and grading of homework, exercises, and tests.
Advantages: Technology-based education offers many advantages, and when blended with more traditional teaching methods, may improve the quality of the educational experience while reducing the workload of instructors. Online, automated processes can provide instant validation of correctness, and repetition can solidify mastery, enabling faculty-student and TA-student interactions to focus on strategic approaches to problem solving. Online material frees students in both time and place, allowing them to learn material at their own pace and in their own space. Students with CS experience may move through course material more rapidly, and those with less precollege experience or exposure to CS may work more slowly. The freedom to complete coursework in a self-paced manner
can be especially powerful for those constrained by job, family, and financial obligations, and to those who may need to spend more time with the material to achieve mastery due to lack of prior exposure to the subject; these students are more likely to be underrepresented minorities.
Beyond online material for a lecture or a course, a number of high-quality online degree programs in CS exist. A master’s in CS is offered by Georgia Tech wholly online. While the program is still relatively young, the response to it has been huge, reaffirming the large demand for CS degrees and the appetite for online programs.
Risks: Technology-based learning is important and growing, but it is not a panacea, at least not yet. Experience with massive open online courses (MOOCs) and other online material shows that great attention must be paid to the learning experience and to the variation in the individual needs of students, and that their effectiveness varies; some analyses suggest that they may further exacerbate educational inequalities rather than mitigating them.4 While one can be confident that technology in the hands of a successful instructor will enhance student learning, this cannot be assumed for online material unsupported by a learning infrastructure. As the reliance on technology for education increases, overall issues related to the sociology of learning must be addressed: academic integrity, the nature of online interaction and collaboration and the role of social media, the appropriate mix of virtual and physical, and no doubt more, currently unanticipated issues. Finally, it may be more difficult to automate elements of advanced courses, which are also seeing high demand, in particular due to the need for deep mentoring and supervised research.
Leverage Undergraduates as Nontraditional Teachers and Mentors
In addition to the traditional departmental teaching resources, undergraduates may provide valuable support in formal and informal capacities, as teaching assistants, discussion leaders, peer graders, project collaborators, and online forum discussants (though this approach is more commonly recognized as a strategy for MOOCs). As noted earlier the use of undergraduate students as TAs is a long-standing practice at many institutions, but it is a relatively recent or an underused resource at others. One prominent example is the Megas and Gigas Educate (MaGE) program at Mount Holyoke College, a women’s liberal arts college, which was launched in 2015 with support from Google’s CS Capacity Program to help meet demand for CS courses during the current enrollment surge in the face of limited resources. This program trains undergraduate students to teach and mentor each other in supportive and inclusive ways in order to increase CS enrollment capacity and diversity in the program.5
Advantages: Undergraduates have the potential to provide significant help and reduce the workload for graduate TAs and for faculty and other instructors. Undergraduates are typically paid per hour of effort, which can cost significantly less than graduate TAs. In addition, the shared experience of undergraduates may make them particularly attuned to understanding the problems and the challenges facing their peers around specific content. Furthermore, from a pedagogical perspective, peer teaching and evaluation can be valuable learning experiences in and of themselves, and can help empower students and build their confidence with the material. Finally, the undergraduate pool is larger than those of graduate students or faculty, and typically more diverse, presenting an opportunity for a more diverse set of instructors, which could contribute to a more inclusive culture.
Risks: Undergraduates who are unclear on the material may cause confusion among their peers. In addition, not all undergraduates have the knowledge or maturity to successfully teach, assess, or mentor their peers, or understand conflict-of-interest situations. If poorly implemented or not properly supervised, this approach can place additional strain on course instructors.
Retrain Faculty and Graduate Students from Other Fields as Computing Instructors
As computing becomes increasingly important in a range of academic disciplines, many Ph.D.s in non-CS fields are emerging with knowledge and experience in computer science. At the same time, Ph.D. production in a number of fields exceeds the number of academic opportunities, and many Ph.D.s seeking an academic position are passionate about teaching. One approach to expanding the instructor or teaching faculty pipeline in CS is to create rigorous training programs for such non-CS Ph.D.s with the goal of preparing them to teach undergraduate CS courses, possibly aimed at students in domains close to that of their Ph.D. research. Stanford University recently started an M.S. program for this purpose,6 and Ph.D.s in math, physics, and engineering are known to have pursued CS teaching positions successfully in the past.
Advantages: Training for non-CS Ph.D.s to become computer science instructors could attract individuals with research experience who are passionate about teaching, and increase the pool of CS instructors. Such individuals could provide especially valuable experience and insights in the context of a relevant X+CS program. Teaching computing outside of CS would also seem to offer institutions more flexibility if there are significant fluctuations in student demand in the future.
Risks: Creating successful M.S. programs targeted at individuals with Ph.D.s in areas outside CS comes with a number of challenges, given the fact that such
individuals likely have varied backgrounds and experiences with CS. Entry into such programs might need to be contingent both on demonstrated success in teaching and some threshold background or performance in CS. The curricula would need to be flexible to enable tailoring to the range of backgrounds and thinking that come with primary training in other fields. Furthermore, professional development of such instructors would also require special attention and mentoring, and consume already limited faculty time.
Leverage External Teaching Resources
There may be additional resources outside the institution (experts, teaching spaces, etc.) that can be leveraged with minimal overhead. For example, regionally co-located institutions could pool teaching resources; local industry or government experts could teach or support college/university classes; or public or private facilities could be found to serve as additional, low-overhead course meeting sites.
Advantages: Collaboration with private industry could provide additional, nonacademic perspectives on computing and insights into the workings of industry to students, and help them form connections with potential future employers. Similarly, those in industry and government may find value in contributing to undergraduate education, and see it as an opportunity to recruit new talent. Leveraging local experts and facilities could be a low-cost tactic, if of mutual interest to the parties.
Risks: Using outside resources could limit the institution’s control and ownership over its own program. Quality control is important and can be especially challenging with external teaching resources. Furthermore, the availability or quality of such external resources varies by region, and may not be feasible for institutions in rural areas. Finally, it is possible that reliance on external resources could deplete the pool of resources available for meeting K-12 CS needs, to the extent which they may overlap. Finally, contract instructors of any sort are difficult to hire in CS, given the current landscape of opportunities in industry, as discussed in Chapters 3 and 4.
Build Mechanisms for Continuously Aligning Resources with Workload and Demand
More responsive resource allocation strategies, such as adoption of Responsibility Center Management (RCM) models (Carlson, 2015), enhanced use of more flexible resource pools, or institution of differential fees that scale with program or course enrollment, can encourage and enable programs to accommodate increasing enrollments.
Advantages: Many of the challenges associated with increasing enrollments relate to budget constraints or a lack of flexibility of resource allocations. Faculty
and teaching and support staff are not easily transferred between units, teaching budgets often do not track enrollments, and staff to expand the capacity of such faculty when needed are difficult to hire. Physical, administrative, financial, and human resources are often bound to decanal or departmental units for decades with no processes for reallocating or sharing them, or for clearly understanding how they are utilized.
An RCM model is one approach to increasing the flexibility and alignment of resources, giving computing units more control over the resources they have available, and better enabling them to respond to the demands of increasing enrollment. Using more flexible pools of teaching resources, such as undergraduate TAs, rather than relying on a more limited number of graduate students, is another strategy. In some situations differential fees could both moderate some of the demand and provide additional resources to expand capacity. These measures can regularize administrative decision making, allowing resource availability to be anticipated well in advance and avoiding the stress and inefficiency of ad hoc responses.
Risks: RCM and other approaches to tying departmental resources directly to teaching activities create barriers to interdepartmental cooperation and collaboration. If a department receives funds for each student it teaches and must pay for each course that its majors take in other departments, it is inevitable that barriers to student enrollment will be erected. It must also be recognized that changing budget models and processes reallocates resources, it does not increase them—that is, some units will lose resources, while CS is gaining them, which could result in the weakening of units with fewer students that are nonetheless highly valued or central to an institution’s mission or character. Furthermore, it is unlikely that the specific issue of responding to demand in CS would be sufficient to stimulate a shift in an institution’s budget model if it was not already the best choice for the institution.
In general the use of market mechanisms as a way to allocate resources or to manage demand must be pursued with caution and with an appreciation for its impacts. As with any market-based approach, which by definition puts an emphasis on willingness to pay, there are impacts on those who may be willing but unable to pay. This is an especially important issue for underrepresented minority students.
Given the rapid evolution of computing disciplines, and the increase in non-CS disciplines as a source of student enrollment in CS courses and research opportunities, institutions should look outside of CS as they contemplate responses to enrollment growth. There are several possibilities that range from relatively modest changes in courses to the creation of new interdisciplinary programs that could alter in profound ways the relationships among units. Such
strategies go beyond administrative reordering, requiring institutions to engage in important intellectual questions about the role of computing in the future of other disciplines and the role and use of computing in the careers that all graduates will pursue.
Much of the increase in enrollments in computer science courses has come from non-majors, and we can expect the demand from these sources will increase in the future as computing becomes an integral part of more disciplines and is expected by employers from a wide range of more industrial sectors. Here, we identify several specific approaches for the non-major.
Increase the Presence of Computer Science in Non–Computer Science Courses
Incorporating computational and computer science knowledge, skills, and experience into non-CS courses could provide non-majors who are primarily interested in obtaining computing skills with sufficient exposure to computing to meets their needs, or enough of an introduction to computational methods relevant to their coursework or career. This approach would leverage the computational knowledge of instructors or faculty in other fields if and when it is available. It would create more pathways for non-CS students to incorporate computing into their education and future careers, as well as providing appealing hooks for CS students to learn more about other disciplines. Shared courses with other domains can enrich the experience of CS faculty and TAs, especially if approached as a collaborative effort, rather than a service relationship. The non-CS faculty and TAs may welcome the opportunity to incorporate new methods and create new partnerships with CS faculty and TAs.
Tailor Introductory Computer Science Courses to the Needs of Non-Majors
Incorporating more of what non-majors need into introductory or specialized CS courses could reduce the need to take additional upper-level courses. For example, after learning basic programming skills such as iteration, abstraction, flow of execution, data structures, and functions in an introductory course, non-majors can focus on additional concepts and projects related to applications in their major. The use of systems, packages, tools, and environments may also be targeted to relevance and use in their major. Successful tailoring of offerings to non-majors in this manner is an open and important area for further research and exploration.
All of these approaches with the non-major in mind offer similar advantages and risks.
Advantages: Shifting CS content to courses in other departments, which obviously provides direct relief to CS, and tailoring CS coursework to the non-major both have the merit of providing a targeted and more efficient educational
experience for the student and the institution. If done well and timed effectively, such approaches can also provide a more effective way for non-majors to learn computing concepts, as well as an environment enriched by the issues and examples from the student’s major. Rigorous instruction in computing outside CS courses would also seem to address some of the underlying demand for computing skills throughout the workforce while offering institutions more flexibility in the event that there are significant fluctuations in student demand for CS programs at some point in the future.
Risks: It is possible that these approaches merely shift the burden from one department to another, although that can be viewed as beneficial by the institution if it serves to distribute the burden more evenly. It is also likely to be more difficult to ensure the quality of CS courses for non-majors, whether taught inside or outside CS departments or related units. Such courses must balance the needs of non-majors with instruction of the relevant CS principles. This can be achieved through interdisciplinary collaboration on teaching strategies, and possible co-teaching. At the same time, strengthening CS offerings for non-majors could deter some students from selecting CS as a major, thus depriving the field of their contributions. Finally, it is possible that these measures could backfire by stimulating even greater demand for CS courses.
Develop a Wider Range of Computing-Related Programs
Computing is important in all aspects of the economy and plays a crucial role in science, technology, engineering, and mathematics fields. Many institutions have included a computing requirement (typically a programming course) for all or a selected subset of majors. However, to be prepared for the workforce or future research, one programming course is generally not enough. Today’s students understand the job market and where the opportunities are. Students majoring in CS or pursuing a double major with CS may actually be more interested in another field but are pursuing CS because they see it as the best way to achieve the opportunities they seek. Creating new programs of study targeted at students interested in CS for its applications to another domain could help to reduce the number of majors and enrich offerings available to students.
Data science, naturally spread over computer science, statistics, applied mathematics, and various application domains, provides a compelling, focused opportunity in which to develop such institutional capacities. Other established majors can be approached computationally as well, such as cognitive science, operations research, econometrics, computational biology, digital media, and others. Several institutions have established or are exploring new CS+X and X+CS blends with synergistic fields, such as computational anthropology, computational linguistics, computational advertising, and so on, as discussed in Chapter 3. Importantly, computational coursework in other domains can provide significant value to other fields and enable a more customized undergraduate experience for
students while reducing some of the burden on CS programs. Such blends could also be effective at improving the diversity of computing-related programs, even if they are not housed in CS units.
Assessing the advantages and risks of these approaches is not straightforward. Unlike the efforts focused on non-majors, creating new computing and interdisciplinary programs goes beyond merely dealing with the excess demand for CS instruction. But, of course, that is the point. Such new programs bring intellectual progress and excitement that go well beyond the administrative benefits of shifting teaching loads.
New Organizational Structures for Computer Science
Historically, computer science programs were formed as departments in colleges of engineering, science, or arts and sciences. Increasingly, colleges of computing or schools of information and computer science or similar organizations, with CS being one of the units, are being formed. The emphases and character of these programs vary widely; Table 6.1 provides a sampling of institutions whose computing organizations are representative of a range of such models.
Concomitant with the increase in the number of students seeking to take CS courses, the reasons for taking these courses have become more diverse, new computing-related programs and majors are being created, and other disciplines are incorporating computation into their own courses. In any such environment placing a CS department into a more autonomous organizational structure that can respond effectively to emerging needs becomes increasingly important. Given the vast diversity of programs, this larger step is certainly not merited for the majority of institutions, but for the 10 percent of the institutions that produce 50 percent of the degrees it ought to be considered, and for the 3 percent of the institutions that produce 25 percent it cannot be ignored.
The organizational placement of a CS department can have significant impact on the undergraduate students choosing the program and courses most appropriate to their interests and abilities, and can affect the diversity of students choosing to enroll in the program. The organizational structure impacts approval and reporting channels, and can have significant impact on how quickly and effectively a program can respond to changes in course demand and reallocation of resources. The appropriate structure could also reduce duplication in courses and programs, and coordinated programs can lead to better graduation rates. The organizational structure typically does not directly affect the research conducted or courses taught by individual faculty.
Computer science departments in a college of science find that their programs differ from many science programs in a number of significant ways: (1) A bachelor’s degree in CS has a wide range of high-paying job opportunities immediately after graduation and students have well-paid internship opportunities. (2) CS Ph.D.s and CS faculty have a wide range of high-paying opportunities outside of academia;
|Arizona State University||School of Computing, Informatics, and Decision Systems Engineering||http://cidse.engineering.asu.edu|
|Carnegie Mellon University||School of Computer Science||https://www.cs.cmu.edu|
|Clemson University||School of Computing||http://www.clemson.edu/cecas|
|Cornell University||Computing and Information Science||http://www.cis.cornell.edu/|
|DePaul University||College of Computing and Digital Media||https://www.cdm.depaul.edu|
|Drexel University||College of Computing and Informatics||http://drexel.edu/cci|
|Georgia Institute of Technology||College of Computing||http://www.cc.gatech.edu/|
|Indiana University||School of Informatics and Computing||http://www.soic.indiana.edu|
|Long Island University||College of Information and Computer Science||http://www2.liu.edu/CWIS/cwp/cics/cics2.html|
|Montana State University||Gianforte School of Computing||https://www.cs.montana.edu/|
|New Jersey Institute of Technology||College of Computing||http://ccs.njit.edu|
|Northeastern University||College of Computer and Information Sciences||http://www.ccis.northeastern.edu|
|Pace University||Seidenberg School of Computer Science and Information Systems||http://www.pace.edu/seidenberg/|
|Rochester Institute of Technology||College of Computing and Information Sciences||https://www.rit.edu/gccis|
|State University of New York, Albany||College of Computing and Information||http://www.albany.edu/ceas/|
|University of California, Irvine||School of Information and Computer Sciences||http://www.ics.uci.edu/|
|University of Massachusetts||College of Information and Computer Sciences||https://www.cics.umass.edu/|
|University of Nebraska, Omaha||College of Information Science and Technology||http://www.unomaha.edu/college-of-information-science-and-technology/|
|University of North Carolina, Charlotte||College of Computing and Informatics||http://cci.uncc.edu|
|University of Pittsburgh||School of Computing and Information||https://sci.pitt.edu/|
NOTE: A number of institutions have information schools that currently do not include the CS department. These include the University of Washington; the University of Maryland; the University of California, Berkeley; Pennsylvania State University; the University of Michigan; the University of Illinois, Urbana-Champaign; Rutgers University; and the University of Texas, Austin.
in certain areas faculty are actively recruited by industry. (3) Increasingly, professional M.S. degrees in specialized computing fields (e.g., security and data science) are being created and have significant enrollment providing a source of income to the department. (4) CS faculty tend to be more entrepreneurial than science faculty.
Computer science departments in schools or colleges of engineering also find their discipline to be increasingly different from the other engineering fields: (1) CS departments offer courses taken by undergraduates across all majors; increasingly, the courses taken by non-majors include higher level ones. (2) Computing is immediately relevant to almost every area and domain outside engineering. (3) CS departments increasingly offer undergraduate and graduate degrees jointly with non-science and non-engineering departments. Undergraduate curricula increasingly allow more flexibility and choices depending on the students’ interest.
Advantages: There are many additional advantages. In a new college or school of computing, a dean represents departments with commonalities not found in other arrangements which allow for better-focused arguments to be made to the higher administration. The departments included in addition to CS can vary considerably and are institution dependent.7 New undergraduate and graduate programs are created in an environment bringing relevant groups together earlier in the needed approval process, allowing for better collaboration and communication. A college offering different undergraduate programs can create a curriculum with a common first year, allowing students to choose among different computing majors after the first year. Programs and majors in one organizational structure have the potential to lead to better course enrollment management and better understanding of objectives students have.
Risks: The risks of creating a new college would seem to be institutionally specific and derive mainly from distributional issues within the university. If a new college gives CS a more direct line to university leadership, for example, that presumably means that the leadership has another direct report, which could dilute the attention it can provide to any single college. There are unavoidably some costs of change and these must be weighed against the benefits. Furthermore, while deans of colleges of computing may be on equal footing with deans of other colleges, this may not change the potential for allocating resources to the new college.
FINDING 8: Departments facing sharp increases in demand for computing courses have experienced significant strain on a wide range of resources. Failure to respond thoughtfully to the demand and the resource deficits will result in adverse conditions for students, faculty, the programs, and the
7 They may include statistics, information sciences, library sciences, computational biology, information technology programs, management information systems, communication, policy, media, and computational social sciences. Alternatively, areas sometimes considered computer science—such as human-computer interaction, machine learning, and robotics—may have their own departments within colleges of computing.
institution as a whole in the near or long term. Conditions such as an unwelcoming academic climate and loss of faculty members can be especially harmful in the long term.
There is no single approach—a “silver bullet”—for responding to enrollment growth that will be optimal at all institutions. Every tactic has benefits and costs. Leaders will need to select strategies and make trade-offs that are appropriate to their circumstances and to their institution’s mission and values.
FINDING 9: U.S. institutions of higher education have differing missions, priorities, and business models, and serve different populations with different needs. There is no one-size-fits-all solution for responding to enrollment increases. However, all institutions need to assess the role of computer science and related fields and make strategic plans to address realistically and effectively the high demand for courses, student interests and needs, faculty and staff workloads, research and teaching allocations, and physical resources. At all institutions there is an opportunity to reassess the role of CS and computing and to consider changes that go beyond the current challenges and position the institution for future success.
Yet, this is a time of opportunity. It is a time for institutions to consider their missions and the constituencies they serve, and to determine what role computing should play in the experience, knowledge, and skills of its graduates of 2025 and beyond. Institutions should take action to meet these goals, and metrics should be defined and monitored to determine progress and unintended consequences.
Institutions may start by performing a self-assessment to establish current facts, feasible actions, and strategic goals. This knowledge will help to elucidate a path forward that is consistent with the institution’s goals. The key questions to consider are as follows:
- Situation and trends
- How has student demand for and access to computing changed in the past 5 years?
- How have the demographics—including motivation for enrolling—of computing majors and course enrollees changed?
- How have the number and workloads of faculty and teaching and support staff changed?
- Impacts of past actions
- How has the institution dealt with enrollment growth in the past? What can be learned from these experiences?
- What were the effects of past changes on program character and climate, and student and faculty demographics?
- Role of computing and computer science
- At what points throughout the process of student application, acceptance, declaration of major, enrollment in courses, is access to computing limited? How is this done? What are the effects of these access controls on students and staff, including on the participation of women and underrepresented minorities?
- What are student and faculty perceptions about the program? How does this vary based on student and faculty backgrounds and interests? How is the program viewed outside the institution?
- What is the perceived relationship of programs to job prospects?
- Are class sizes affecting student stress factors and climate? Do these effects vary by student background and demographic?
- Why are students enrolling in computing courses at your institution?
- What is the educational mission of your computing program for majors and for non-majors? Is this consistent with the interest of these students?
- What are the institution’s target goals for longer-term program growth? How many major, mixed-major, and non-major students should be accommodated in 5 years? 10 years?
- How will these goals impact the preparation of students for their professional lives, students’ intellectual growth within the college/university experience, and the culture and climate for students and faculty in computing programs and throughout the institution?
- Non-major interest and engagement with computing
- What are academic fields and career goals of the non-majors enrolling in your computing courses?
- Are these courses required for their degree program?
- Do computing courses provide a necessary or competitive skill in the student’s chosen field(s)?
- What other departments or programs on campus offer courses with some computing content?
- Are there opportunities for collaborative blending of computation with other fields that might help to meet student interest?
- Faculty, staff, and other resources
- How do the growth of undergraduate CS enrollment, TAs, faculty, instructors, and academics compare since 2006?
- How many new faculty and teaching positions were authorized since CS enrollments have begun increasing? How many of the positions were filled?
- If the department faced faculty retention issues, how were they handled? Did they have an impact on the morale and climate in the department?
- How does the institution provide resources and training for faculty not experienced in teaching large lecture classes? Is class size considered when determining departmental teaching loads?
- Institution-specific mechanisms for growth or restriction
- What measures are available for restriction, growth, or restructuring of computing programs in response to enrollment pressures?
- What measures are available for managing course enrollment, declaration of major, or admission into the program?
- What options exist for building resources and capacity?
- What are the inherent limits or drawbacks of these measures? How will they affect quality of education, student satisfaction, the participation of students from underrepresented groups, and institutional balance?
These assessments will require the examination of institutional data. This could include application and admission rates and thresholds for entry into CS programs and courses, as well as student enrollment, performance, and attrition rates. Some of the necessary information should be available via the institution’s administrative data systems; new data collection such as student surveys or discussions with faculty and staff may also be needed.
While there is no single, optimal strategy that is right for all, the committee believes that all institutions should be proactive and creative in responding to the challenge of large and increasing CS enrollments. Taking incremental actions to get through the next year or semester are unlikely to produce the best outcomes for the institution, and have in the past been associated with negative outcomes such as decreased participation of women undergraduates in computing. Accepting the overwhelmingly important role that computing will be playing in the future of society and the university, institutional leaders should view this as an important opportunity to be seized.
This page intentionally left blank.