Paul Penfield, Jr.
William M. Siebert
John V. Guttag
Campbell L. SearleMassachusetts Institute of Technology
Two years ago at this conference (1) we reported our plans for a five-year program in electrical engineering and computer science, leading to the simultaneous award of bachelor's and master's degrees. An update on our plans (2) was presented one year ago. The program has now started and we can describe the implementation and some preliminary results.
The new program was motivated by three distinct changes that have occurred since the last major changes in engineering education:
The technology we teach has changed in at least two major ways in recent decades. First, there is a higher digital, as opposed to analog, content in most systems that electrical engineers deal with. Coverage of analog circuits, electromagnetism, and classical control theory must be accompanied by more exposure to digital systems and computation. Second, advances in microelectronics have made it possible to design and fabricate very complex systems, with superior performance, very cheaply. Today the disciplines of computer science and electrical engineering, once considered separate or diverging, are closer than ever.
Today's engineers, more than ever before, need an appreciation of the societal, business, technical, and human context in which the process or product being designed will work. Gone are the days (if indeed they were ever here) when an engineer designed, to a specification generated by someone else, a product to be manufactured by others. Engineers are now expected to participate in marketing, product definition, manufacturing, cost control, and many nontechnological aspects of the job.
Finally, society has benefited greatly from new electrical and computer technology and from such products as global wideband communications, personal computers, and other products with embedded computation. We are just beginning to feel the effects of this information revolution. Yet society is still led, by and large, by people without a deep understanding of science and technology, who cannot appreciate either the technical advances nor their significance and potential. Society needs leaders who are technology-literate. General universities are not filling the need. Maybe it is up to us engineers. If so, we should provide science- and engineering-based general education.
These three external changes call for changes in engineering education. Our new Master of Engineering (M.Eng.) program is designed to meet this need. It has more technical material and a seamless merging of electrical engineering and computer science. It has room for contextual material. The bachelor's degree portion provides an excellent general education grounded in science and technology.
In our zeal to make our graduates fit the needs of society, we must not forget that our primary mission is to educate young people for a successful life, not merely a successful career. Another way of saying this is to note that our customers are not society or prospective employers, but rather the students themselves. They have all the usual problems of young people growing up. They cannot be successful in either career or life without understanding themselves, understanding society, and appreciating the diversity of thought, method, and style they will encounter.
However, engineering education is more than general education. Our students need enough preparation for immediate and, if they wish, lifelong employment as an engineer. They need the necessary mathematical tools, the scientific basics, and a knowledge of their particular engineering discipline, and they also should know (or quickly learn on the job) the way engineers work.
Beyond this, we need to ensure that our education prepares students to do things besides act as engineers. Many if not most of our students will experience career changes more than once after they graduate. We want their engineering education to be an enabling foundation, not a confining one.
With all this in mind, we believe that a modern engineering graduate requires at least the following:
The department may not be responsible for satisfying all these needs, but it certainly is for the first six or seven. There are specific features in our new curricula that address these, and there are other programs at our university for the remainder.
We will describe our new program from three different points of view. First, we describe the structure we have selected. Next we discuss the content of the curriculum. Finally, we report how the resources needed for the program have been estimated and secured.
We concluded that it is no longer possible to cover the needs outlined above in four years, without seriously compromising the technical content. The fact that four years is not enough has not escaped the attention of our graduates, since most of them continue for a master's degree, at MIT or elsewhere, either immediately or after some work experience. Their employers agree, and usually support them in their further schooling.
Our new M.Eng. program enables our students to have this experience here at MIT, with a minimum of fuss and a great deal of flexibility. Previously most of our own students could not pursue a master's degree here, because we thought of the master's program as a prelude to the doctoral program, and only admitted those few whom we deemed capable of writing a doctoral thesis. This high standard, not relevant for a program like the M.Eng. which is intended to prepare people for an engineering career, is now only used for those seeking the Ph.D. Admission to the M.Eng. program is based on whether a student is capable of taking graduate-level courses and handling a short thesis project.
In four years it is still possible to provide an excellent general education based on science and technology, even if not one that will be a suitable preparation for the practice of engineering. We have retained four-year S.B. degrees for those who want to do things other than engineering, or who want to attend graduate school elsewhere, or who may want an entry-level engineering position. These are honorable degrees that serve a valid purpose.
The M.Eng. degree is awarded after five years of study. Its requirements include the requirements of our bachelor's degree as a subset, and normally the two degrees are awarded simultaneously. The program is designed to be seamless with respect to the traditional boundary between undergraduate and graduate education. That is, continuation to the fifth year resembles the transition between the third and fourth years, more than the traditional steps of graduation followed by entrance to graduate school. Students know at the end of their third year if they have this opportunity, and can plan accordingly. They can optimize their schedule, for example by postponing some of the undergraduate requirements until the fifth year, or by taking early a specialized graduate course that is not offered every year.
The new curriculum is also seamless in another dimension. For twenty years we have had two S.B. degree programs, in EE and in CS, with different requirements and structures. For the S.B. part of the new program we designed a single structure into which both curricula fit naturally, with overlapping courses. This had a major benefit. A student can now follow a personalized curriculum that is neither EE nor CS but is sort of in-between, yet just as rigorous. We are calling this new, more flexible, curriculum EECS (electrical engineering and computer science). For the benefit of students who (we presume) will want either EE or CS to appear on their diplomas, we have designed the ESE (electrical science and engineering) and CSE (computer science and engineering) curricula. It was easy -- we simply replaced a few restricted electives with required courses to force some specialization. A student does not need to declare which of the three S.B. degrees is expected until the last semester. The ESE and CSE curricula are accredited, and we expect the new, more flexible EECS curriculum to be accredited soon.
We listed above ten things that our graduates need. We now explain the features of the new M.Eng. and S.B. curricula that help provide each.
Let us start at the bottom of the list. The last four items are important for all students, not only those in engineering. They are part of any good general education, and courses to provide them are prescribed by the university and taught outside the department. There is an extensive humanities requirement (twice as much as ABET requires) and a writing requirement. Because these needs apply to all our students, they are part of both the M.Eng. and the S.B. curricula.
Now at the top of the list, the cornerstone of any engineering education is the technical content. Some have advocated that departments reduce the engineering-science content of curricula to make room for also-needed material devoted to context, practice, ethics, and other nontechnical topics. On the other hand, the ability of engineers to keep up with rapid advances depends on their understanding of fundamental technical material of permanent value and relevance. For the past forty years engineering education has been served well by its emphasis on engineering science. Our new curricula have a stronger, not weaker, technical content. The freshman core (physics, math, chemistry, and now also biology) is common for students in all departments. The curricula continue with laboratory experience and technical courses in EECS. All department students take four courses in the basics of electrical engineering and computer science, plus some advanced mathematics.
Facility with mathematics is essential for engineers. Our curricula include two courses in calculus, one in differential equations, and for most students one in probability and one in discrete mathematics.
The need for breadth and depth is satisfied by a requirement that students select nine EECS courses (5 for the S.B.) from seven lists grouped by topic:
Each list contains a "header" course that is a prerequisite for most of the rest of the list.
Some of these courses are in the EE side of the department and some are from CS. Depth is assured by the requirement that three of the nine courses come from any one of the seven lists; breadth is assured by a requirement that four of the courses come two each from two other lists. The final two courses may come from any list. The student has a great deal of flexibility in the choice of what to specialize in, but must specialize in something. There is similar flexibility in the choice of breadth. Students who wish their S.B. degree to be in either ESE or CSE simply make their selections accordingly.
Design experience is distributed throughout the department courses. Each course carries with it a certain number of "design points" and students must accumulate a substantial number of points. Again there is great flexibility in the way individual students can satisfy this requirement.
One of the goals listed above is the desire for continued learning throughout life. There may not be a single best way to accomplish this goal, but several things help. First, if the quality of instruction is high, then classroom learning can be fun. Second, learning done under strong immediate motivation is effective, and hands-on projects can provide such a setting. Finally, learning done with minimal detailed guidance is usually ultimately satisfying, and the required thesis includes such an experience.
Lastly, one of the items listed above remains to be discussed, the need of engineers to appreciate the context of their work. Many students already have the right attitude, but we have not figured out yet how to best help those who do not. This area is perhaps the weakest part of the new curricula.
We expect to offer 80% of our undergraduates the opportunity to continue through the fifth year to the M.Eng. degree, and we expect about 80% of those to accept our offer. During the authorization process for the new program we wrote a business plan. We estimated the additional number of course takings per year, and the increases in thesis-supervision, classroom teaching, and advising loads. We estimated classroom teaching to rise by 8%, thesis supervision by 3%, and advising by 3%. We examined in detail the particular courses that would be taken and judged whether additional faculty or teaching assistants would be needed.
We then estimated the increased resources needed. This amounted to a 3% increase in faculty, a 10% increase in the number of TAs, and one support staff, for a total budget increase of 5%. We then estimated the tuition from the additional students, and demonstrated that the added revenue represented about 10% of our budget. Thus the program pays its portion of central costs. It has a 50% "gross margin," or, looked at another way, contributes at an effective overhead rate of 100%.
We actually did this exercise not only for the eventual steady state, but for each of three transition years, and the Provost is approving the increases on a year-by-year basis.
An important question is how the students will pay for the fifth year of study. They are only eligible for university-administered financial aid for their first eight semesters on campus. We have estimated that a combination of additional TA openings, some external fellowships (the students will qualify for most fellowship competitions), and an increase in our industrially supported internship program will cover over half the need. The rest would be covered by a combination of family funds and loans. Other professional education, e.g. legal and medical, is routinely financed by loans which are then paid back with the added earning power conferred by the advanced degree. The same idea should work for engineering education. To encourage this approach, the department has established a program under which it will pay the interest on loans taken out by fifth-year students, thereby making the loans interest-free until the student is done with the program. At the time of writing, this program has just started and it is too early to tell whether it will be effective.
The Master of Engineering degree was approved by the MIT faculty in December, 1992. About 25 seniors were immediately admitted to the program. They were selected from a larger number who applied. Also, some of the students in our five-year internship program elected to follow the new curriculum. As a result, 35 students were in the first wave of M.Eng. graduates in 1994.
About 75 students from the class of '94 were admitted to the program in June 1993. Members of this class were permitted to follow the new S.B. curriculum. As of this writing we do not know how many of these will register as graduate students in Fall 1994, but we have in our business plan a target of 62 over and above historical levels.
In the summer of 1994, about 130 juniors from the class of '95 were admitted. Our target first-year graduate population of these people, in Fall 1996, is 100. This is the planned steady-state population in later years. Most of the people in this class are following the new S.B. curriculum, even if they are not planning to stay for the M.Eng. Students in later classes are all expected to follow the new curriculum.
Two new courses are currently under development because the act of defining the new curricula exposed a need for them. One is a course in discrete mathematics at the sophomore level, and the other a junior course in signals, control, and communications. These courses were offered during the past year to small groups of students, and from now on will be taught to a much larger group.
To encourage M.Eng. students to develop oral-presentation skills, we held a mini-conference called "EECS Master Works," where students gave talks on their theses. The submissions were refereed, and prizes were awarded for the best presentations.
During the past year or so, as the program has been implemented, many decisions had to be made. The department administration paid great attention to details, under the theory that if we did not, then the details would somehow pay attention to us and we might not be pleased. We have attempted to make the program consistent with all other programs of the department.
Although it is still too early for many lessons to have been learned from the program, there are some things we can report.
Among the faculty and the students in the program there is high morale and enthusiasm. Most of them sense that they are helping establish a new mode of engineering education that eventually will, for good reasons, be adopted widely. We hope this degree of excitement will continue in future years.
One result of interest involves the percentage of students who declare an undergraduate major in electrical engineering vs. computer science. Traditionally students have preferred ESE to CSE by a ratio of almost two to one. Members of the class of '97, who selected a major in the summer of 1994 before their sophomore year, had the opportunity to choose ESE, CSE, or the new, more flexible EECS undergraduate major. The students apparently want the flexibility of the new degree. Registration in the three majors was almost equally split, with each drawing between 30% and 35% of the total.
We were initially concerned about the impact of the M.Eng. program on our highly popular internship program. In many ways the structure of the new program is based on that program. In particular, the internship program is a five-year one, with three summers and one fall spent at an industrial plant, and with relatively easy admission to the fifth year. The concern was that students might have been interested in the internship program not because of its industrial experience, but because it led to a master's degree. Now that the M.Eng. program has the same easy admission, the fear was that a much smaller number of students would apply to the internship program. We were pleased to find in Spring 1994 that there was as much interest in the internship program as ever, so apparently our students have been applying to it for the right reason.
Another concern had been that there would not be enough master's-level thesis topics available, or that faculty members would only supervise doctor's theses, because of their higher likelihood of leading to published papers. So far, M.Eng. students have not had difficulty finding thesis topics and supervisors. The real test, however, will come in future years when there will be more M.Eng. students requiring theses.
Our new program violates the customary paradigm that the break between classroom-intensive, structured education and research-oriented, apprenticeship education occurs between undergraduate and graduate years. This paradigm was deeply imbedded in administrative procedures throughout the department and the university, and our new programs required a variety of changes. We have been pleased at how cooperative all parts of the university have been, and how eager people have been to help make the new program a success.
We wish to gratefully acknowledge the faculty members of the department for the enthusiasm with which these curricular changes have been embraced. We also wish to acknowledge the contributions of countless members of the MIT community, both faculty and administration for their help in securing the necessary approvals and for thoughtful committee deliberations.
1 P. Penfield, Jr., J. V. Guttag, C. L. Searle, and W. M. Siebert, "Shifting the Boundary: A Professional Master's Program for 2000 and Beyond," Proceedings, 1992 Frontiers in Education Conference, Nashville, TN, pp. 645-649; November 11-14, 1992.
2 P. Penfield, Jr., J. V. Guttag, C. L. Searle, and W. M. Siebert, "Master of Engineering: A New MIT Degree," Proceedings, 1993 ASEE Annual Conference, Urbana-Champaign, IL, pp. 58-61; June 20-23, 1993.
Since 1989 Professor Penfield has been Head of the Department of Electrical Engineering and Computer Science at the Massachusetts Institute of Technology. His technical interests have included solid-state microwave devices and circuits, noise and thermodynamics, electrodynamics of moving media, circuit theory, computer-aided design, APL language extensions, and computer-aided fabrication of integrated circuits. Correspondence about this paper may be directed to Prof. Penfield, Room 38-401, MIT, Cambridge, MA 02139; (617) 253-4601; penfield@mit.edu.
Since 1993 Professor Guttag has been Associate Head of the Department of Electrical Engineering and Computer Science at MIT. He served on the committee that designed the EECS Master of Engineering program. His technical interests include software engineering and computer systems, particularly programming methodology, formal specifications, theorem proving, and programming languages.
Professor Searle retired in 1993 from MIT. Before then he chaired the committee which designed the Master of Engineering program. His technical interests have included solid-state circuits, semiconductor devices, auditory perception, and modeling of the human auditory system.
Professor Siebert joined the MIT EECS faculty in 1952. Recently he served on the committee that designed the EECS Master of Engineering program. His technical interests include biomedical engineering, random-process theory, and the application of communication and systems theory to the understanding of physiological systems.