We can all agree that departments of computer science can benefit from external reviews on a regular basis. Even the best departments can get stuck in a rut and fail to see slowly growing problems until they have turned into a crisis. Other departments find themselves with problems that they cannot solve without help from outside the department (e.g., shortages of vital resources—human, budgetary, computational, and space).
Sometimes these problems are visible to larger academic units, and outsiders need to be brought in to help the department see and overcome the existing problems. More often, the department sees the need for change and needs administrative support to obtain the resources needed to effect change.
How can we get the best feedback and advice to departments in computer science? Professional degrees like engineering typically have licensing requirements. On the other hand, computer science—like math and physics—does not have a professional license. CSAB (Computer Science Accreditation Board) was set up starting in 1984 to provide accreditation for computer science in the same way that ABET provides accreditation for engineering degrees. CSAB, which later merged with ABET, is designed to set and ensure minimum standards for degree programs in computer science.
Looking at the accreditation guidelines, most seem quite reasonable for professionally-oriented degree programs. The general ABET guidelines have some critical reequirements.
- Student performance be evaluated.
- Programs have published program educational objectives.
- There be publicly stated student outcomes.
- The program must assess and evaluate the extent to which student goals are being met.
- The program curriculum requirements are consistent with the educational objectives.
- The faculty teaching in the program must have appropriate education and expertise.
- Facilities must be adequate to support attainment of student objectives.
- There be sufficient institutional support to ensure the quality of the program.
The computer science requirements for ABET accreditation add extra detail including a specification of a larger number of courses than the general guidelines, as well as a detailed specification of the areas that must be covered in CS and math.
These are all the right things to be reviewing to ensure a program's quality.
What are the problems with accreditation?
Unfortunately, many accreditation regimes have a rigidity that reflects a one size fits all approach, resulting in a focus on criteria that are not central to the success of a specific unit. This often results in a narrow focus on specific criteria rather than a thoughtful consideration of the issues most relevant to a specific situation. The accreditation process can end up occupying faculty time and effort on less relevant criteria, rather than taking the time to brainstorm creative solutions to real problems.
It is the issue of how to focus faculty time and creativity in a department that gives us most pause. In our experience, even schools with similar goals often have very different issues that require examining different parts of the curriculum. One school may have a problem with articulation of the parts of the intro sequence; another may have issues with retention; while another may be suffering because of lack of space and equipment. When we start looking at schools with different goals (e.g., liberal arts colleges versus engineering schools), the differences become enormous, and the solutions require very different approaches.
The costs for accreditation are not negligible. Fees for the initial visit are nearly $11,000, with annual maintenance fees of $1,380. Accreditation is for a maximum period of six years, so these costs will recur. The only program with a review like accreditation at most liberal arts colleges is chemistry (called certification rather than accreditation) and it has no cost (though there is no site visit).
Computer science shares its orientation with other sciences, like physics, and mathematics, which do not have accreditation agencies. In these fields, students graduate with deep inquiry skills and the ability to fully develop their professional skills in real-world contexts.
What are the alternatives?
Liberal arts colleges and other institutions typically conduct external reviews according to outlines set up at the institutional level. Our institutions set up external review committees of two or three members, including peer and aspirant institutions. Departments prepare a self-study consisting of, at least, the following information:
- a history of the department/program, including changes in faculty, facilities, etc.;
- a statement of the department/program learning objectives;
- an assessment of student learning in the major based on actual student work, not student opinion, in light of the department/program learning objectives—this could be a review of the department's recent annual assessments of student learning in the major;
- data on enrollments and majors, disaggregated by gender and ethnicity (as provided by one's institutional research office);
- feedback from and profiles of current students and alumni obtained from surveys taken early in the self study process;
- comparative data from departments or programs at similar colleges;
- faculty profiles, such as CVs or biographies, including teaching, research, and service;
- the department's tenure and promotion standards;
- catalog copy and course syllabi; and
- a history of recent financial support for faculty and students, including grants received for research, travel, and senior projects.
In short—this information is like that which would be prepared for an ABET accreditation visit; however, with less specification of details and more room for focus on the areas the department considers most important.
With this background, the department highlights issues or areas for the external review committee to focus upon. The committee is also charged with discovering issues of which the department may not be aware.
The external review committee reads the report in advance and then arrives for a two-day visit with the department. The department visit begins with a discussion with the Dean and then proceeds with meetings with all departmental faculty both as a group and individually, with students taking courses and majoring in the department, with support staff and IT, and with faculty from related departments.
The visit ends with an oral report to the President (or Provost) and Dean focusing on strengths and weaknesses of the department and recommendations for the department. A more detailed written report is submitted to the college and shared with the department typically within four to six weeks.
After the report is received, the department meets with the Dean to discuss it. A year later the department chair submits a brief follow-up report to the Dean, outlining what changes have been made in response to recommendations in the report.
What are the differences then?
The structure of the external review, while involving substantial work for the department, provides flexibility for a careful examination of the state of a department in the context of the larger institution. Sometimes these visits are straightforward, with few surprises. However, issues can arise that were not anticipated. Generally these reflect problems that have been simmering and need to be resolved. In most cases, a careful preliminary self-study provides the committee with sufficient information to make useful suggestions. Sometimes there are unexpected criticisms of the department, but most departments respond well to them after the initial surprise.
Both authors have served on multiple review committees—Bruce for liberal art colleges and Martin for public institutions. The processes are similar to what was outlined above. Feedback from most schools has been quite positive, with both the departments and administrations feeling that they have a better understanding of the issues they are confronting.
The structure described above is comprehensive, yet much more flexible than that found in the accreditation guidelines. There is no arbitrary number of credit hours in various subjects, nor an excessive focus on internal processes. Instead, there is a careful discussion by departments of what is required and why, and what makes their program special. Arguments can be made for more resources (faculty, space, equipment) without rigid guidelines. Comparative data from similar schools is often more persuasive than specifications for values that may not be relevant to the program.
Finally, external review committees typically look at more than just the major program. There is concern about how well the department is serving those students wishing to take just one or two courses, those wishing a more substantive background to go with other majors, or those participating in interdisciplinary programs. Thus it can serve as a reminder that a department has broader responsibilities in the college or university.
Is accreditation necessary?
One could argue that even if one believes the ABET accreditation is too rigid, it is necessary to subscribe for competitive reasons. In our experience, this is not the case. Our students, their parents, and their employers are not concerned with whether our program was accredited.
In short, accreditation is simply not required in the computer science field. Neither graduate programs nor professional practice requires students to have attended an accredited computer science program.
If one is operating an engineering, business, or nursing program, there is no choice. Based on professional norms in the respective fields, these programs must undergo their respective accreditation programs.
We in the computer science community may make our own choice.
Broadly, liberal arts colleges—which are renowned for producing students with strong and complete backgrounds, ready for industry and graduate school—are not accredited. In our experience, the top liberal arts colleges have felt no need to pursue accreditation in CS.
We would note that the following research universities do not have ABET accreditation in computer science: Harvard, Cornell, Stanford, NYU, Yale, Columbia, Penn State, University of Wisconsin-Madison, University of Washington-Seattle, and Carnegie-Mellon. We assume that they found their own processes sufficient to ensure that they have high quality programs in computing.
Why undergo these reviews?
If a program requires accreditation to be accepted by its constituents, then an institution must ensure that it maintains the standards for accreditation. Then, if the accrediting agency demands that certain resources be provided to the program (e.g., personnel or equipment), that shall be done, or accreditation will be lost.
Accreditation recommendations thus have "teeth" in a way that the more self-managed processes that we describe do not. Accreditation processes may result in programs being provided with essential resources. For most programs, such threats are not necessary in order to accomplish improvements.
In schools with a healthy relationship between the department and its administration, these self-managed processes can provide excellent feedback on the quality of our departments as well as useful recommendations for improvement. Moreover, these processes are ones that their own administrations have confidence in and have bought into—making the recommendations more likely to get serious consideration.
Thus with strong department leadership and a supportive administration, self-managed reviews with external input can provide the necessary guidance for department improvement.
Just as departments are encouraged to adapt the ACM/IEEE curriculum recommendations to best fit their own goals, CS departments should be able to adapt accreditation criteria and apply them in the ways that make the most sense for their goals.
Accreditation is not the only way.
The authors would like to thank Henry Walker for his encouragement and constructive comments.
Kim B. Bruce
Department of Computer Science
Claremont, CA 91711
Fred G. Martin
Department of Computer Science
University of Massachusetts Lowell
Lowell, MA 01854
Copyright held by author.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2019 ACM, Inc.Contents available in PDF
View Full Citation and Bibliometrics in the ACM DL.
To comment you must create or log in with your ACM account.