Paul Penfield, Jr. and Jesús A. del Alamo, The MIT EECS Master of Engineering, Proceedings of the Engineering Foundation Conference, Realizing the New Paradigm for Engineering Education, Baltimore, MD, pp. 38 - 42; June 3-6, 1998. Abstract. Presentation. Text.

The MIT EECS Master of Engineering

Paul Penfield, Jr.
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology

Jesús A. del Alamo
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology

Abstract

In 1994 thirty-five students were in the first group to receive the Master of Engineering degree from the MIT Department of Electrical Engineering and Computer Science. Now, four years later, about 200 such degrees are awarded each year. About two-thirds of EECS undergraduates participate.

This new degree program was intended to prepare a person for a successful career as a practicing engineer, or at least a career that starts out that way. Its design was based on a simple theory of the various types of careers our graduates might aspire to. It was intended to be the department's flagship program.

The result is an integrated five-year program leading to the simultaneous award of a bachelor's and a master's degree. The structured style typical of undergraduate programs is seamlessly combined with the advanced specialization found in graduate programs, so that students can plan a five-year experience in a unified way. Students may specialize anywhere within the broad discipline comprising the union of electrical engineering and computer science.

Because the program combines undergraduate and graduate elements in novel ways, securing university approvals was not straightforward. A business plan was necessary to justify the additional resources needed to staff the program. Care had to be taken to ensure that our bachelor's and doctor's programs were not adversely affected. Concerns were raised about possible narrowing of student experience, premature specialization, and general increase in student pace and pressure. Some wanted the new program to have an increased liberal or general education component.

While designing this new program, we took the opportunity to rethink our undergraduate programs. We made the electrical science and engineering program and the computer science and engineering program identical in structure, and started a third program with a broader foundation in both EE and CS. All three programs are accredited by ABET and two of them also by CSAB.

The program is considered to be a success, yet its essential features have not yet been adopted by other departments at MIT, nor, to our knowledge, at other universities.


I. What Is the New MIT EECS Master of Engineering Program?

Since 1993 the Department of Electrical Engineering and Computer Science at MIT has offered a program leading to the Master of Engineering (M.Eng.) degree. The Department is the largest at MIT, typically awarding over a third of MIT's 1000 bachelor's degrees per year. The M.Eng. program prepares over 200 students per year for careers that start out with the practice of engineering.

This five-year program combines the classroom structure typical of undergraduate study with the advanced material and specialization needed by today's engineering graduates. Students receive simultaneous S.B. and M.Eng. degrees. It is currently described as the Department's flagship degree program, aspired to by the vast majority of undergraduates.

II. Why Was This Program Needed?

The educational programs of the Department are based on a simple model.

Bachelor's degrees are intended for general education of students, for preparation for life, and for foundation for a variety of careers, including especially electrical engineering and computer science. The Department's S.B. programs are designed for people seeking a general education with strength in science and engineering. These programs may be considered the modern form of a liberal education -- modern in the sense that they are centered on science and technology, and liberal in the sense that they provide broad intellectual development and can lead to many different types of careers.

The new M.Eng. program includes the liberal education of the bachelor's programs but then goes further. It leads students toward a particular career that, at least initially, includes the practice of engineering. This program requires a combination of technical breadth and specialization, along with a thesis. These prepare students for the design, analysis, and synthesis tasks of engineering. Experience has shown that for the greatest success in engineering, education at the master's level is needed. MIT EECS graduates have demonstrated this last point by seeking, in very large numbers, professional education beyond the bachelor's level.

Doctoral study, with its deep research experience, is for those who seek careers involving the discovery, codification, and transmission of the knowledge on which EECS is based. These careers can include not only traditional teaching and research, but any activities that benefit from knowing what it is like on the frontiers of knowledge.

In some fields prospective researchers often continue their education past the doctorate in a post-doc position. While these programs are useful in further developing research skills, they too often have the reputation of serving as an employment buffer when supply of research personnel exceeds demand.

At many universities EE and CS (and computer engineering) are considered separate fields of study, but at MIT the prevailing attitude is that they constitute an inseparable single discipline. Graduates will find themselves in situations where the rapidly changing boundaries between hardware and software, between algorithms and chips, and between communications and computation defy engineering solutions that are narrowly based. Therefore at the undergraduate level the curriculum includes material from all these areas, and students are required to have broad exposure.

Before 1990, the EECS programs at MIT did not adhere to these principles. There were separate EE and CS bachelor's programs. It was difficult to fit in the breadth needed by modern engineers. At the same time, the master's program served mainly as a proving ground for would-be Ph.D. students, and over the years the S.M. theses had become extremely long and deep. Admission to the S.M. program was based on potential to perform doctoral research, rather than on ability to be an engineer. Finally, the prevailing pattern of four years for the bachelor's program followed by one or two years for the master's introduced an inherent inefficiency, although it did facilitate shuffling of students among different universities at the end of four years.

III. How Was This Program Developed?

The need for reform was recognized at MIT, principally by Prof. William M. Siebert, during the 1980s. He discussed the notion of a new approach to "the First Professional Degree" in several venues, and gradually persuaded the faculty in the Department of its merit. In 1989, this project was made a priority within the Department. A small committee was set up to oversee its design, and the department faculty were given the opportunity, in several forms, to let their views be known, and to participate in the details. The key people were Profs. Siebert, Campbell L. Searle, John V. Guttag, and the department head, Paul Penfield, Jr.

In 1992 the plans were formulated more completely. Informal presentations were made to many MIT committees. One of the problems encountered was that the traditional governance of the university was (and still is) split along the lines of undergraduate/graduate programs. The proposed new program was not based on this paradigm and therefore committees were unsure how to treat it. To this day the administration of the program is made more difficult by the fact that it does not respect this split.

It was decided that for the program to succeed there had to be a consensus among department faculty. Several informational meetings were held, and in the end every faculty member, without exception, was expected to write a letter to the department head stating whether he or she was in favor, and if so, why, and if not, why not. This was not a secret ballot -- everybody had to stand up and be counted. The result was that there was overwhelming support along with recognition of some concerns outlined below.

A business plan, made to justify the additional resources needed to operate the program, was approved by the MIT Provost in 1992. A description of the planned program and its need was provided to the MIT faculty [1]. The necessary faculty votes were made at the MIT Faculty Meeting on November 18, 1992. The first graduates of the program received their degrees in 1994. In 1998 an internal review, mandated by the 1992 faculty vote, considered numerous concerns that were raised when the program was being designed.

Implementation was done under policy guidelines developed by the Department's Professional Education Policy Committee, chaired by Prof. Jesús A. del Alamo. While the program was being developed, the engineering education community was kept informed by a series of conference presentations [2], [3], [4]. Now that the program has been in operation long enough for an assessment, it is appropriate to present an evaluation of its success.

IV. Concerns

There were many concerns expressed by various people about the new program. Would the highly successful VI-A Internship Program be affected? (This program already offered a more or less seamless five-year program leading to simultaneous S.M. and S.B. degrees with industrial experience.) Would the department's doctoral program be adversely affected? Would the program prove to be so attractive that an unhealthy fraction of MIT undergraduates selected EECS as a major? Would there be enough short-term thesis topics available? How would the M.Eng. thesis be kept from becoming as long as most S.M. theses of that day? How would the new S.M. thesis be kept similarly short? Would students be able to afford the cost of the fifth year of education? Would the notoriously intense pace and pressure of the undergraduate experience be increased, thereby undermining the liberal goal of the program? Would students be driven to try to finish the program in four years rather than five, perhaps to save money? Would there be negative effects on those not admitted? There was concern about the program's effect on gender and racial diversity of the MIT EECS student body -- would minority groups be selectively favored or disfavored?

Some people would have liked the additional year to be spent in part on topics to broaden the students perspective on society and the context in which engineering is done. As it was, one additional free elective was added but all the rest of the fifth year was to be devoted to technical material. Some considered this to be an opportunity lost.

V. The Program as It Turned Out

While the M.Eng. program was being designed, the opportunity was taken to reform the S.B. programs also. The structure of the "VI-1" Electrical Science and Engineering program and the "VI-3" Computer Science and Engineering program were aligned. A common core of material at the sophomore level was retained. The junior-level advanced undergraduate courses were examined and a new one dealing with control, communications, and signal processing was introduced. To serve the needs of students who wished to retain breadth past the sophomore year and not necessarily specialize in either EE or CS, a new "VI-2" program was designed, named Electrical Engineering and Computer Science. This had the same structure as the other two programs, but (after 1996) students were required to have junior-level exposure in both EE and CS. This new program is the most constrained and difficult of the three, and it is also the most popular. The M.Eng. program, by encouraging breadth within EECS, is seamless with respect to the distinction between EE and CS.

The M.Eng. program involves admission with a minimum of fuss and bother after the third year. Those admitted can be assured of their ability to stay for the fifth year and seek the M.Eng. and S.B. degrees simultaneously, and can therefore plan their final two years with that objective in mind. They can delay until the fifth year some S.B. requirements in order to fit into their senior year a graduate course that may be offered alternate years. Or they can take, as seniors, graduate courses that lead to an M.Eng. thesis. In general they can plan a single program of study rather than two separate 4-year and 1-year programs. Also, the classroom orientation of the undergraduate years is carried through to the fifth year. These are the two ways in which this program is seamless across the undergraduate-graduate boundary.

Despite its seamless character, the program still retains some important distinctions between the undergraduate and graduate years. Financial aid is not as readily available for the fifth year, and goes from being need-based to being merit-based, although funds were raised to pay the interest on loans taken out by M.Eng. students. The thesis, normally done during the fifth year, requires a graduate level of motivation and independence on the part of the student. In general, fifth-year students are expected to display the level of maturity and responsibility normally associated with graduate study. They are also expected to maintain B or better grades.

At the same time this program was put into place, the S.M. program (which continues to be available to graduate students from outside but not to MIT EECS students) was changed so that the thesis expectations match those of the M.Eng. That is, the S.M. thesis no longer can be a multi-year mini-doctoral thesis.

To enforce the shorter theses, the department adopted a policy that financial aid administered by the department would stop after four semesters for S.M. students or three semesters (beyond the eight for the undergraduate program) for M.Eng. students.

The need for M.Eng. students to be exposed to contextual and professional material has been partly addressed. Some small-scale experiments have involved new courses with names like "Ethics and the Law on the Electronic Frontier," "The Structure of Engineering Revolutions," and "Structure, Practice and Innovation in EECS." An annual department-wide event named Master Works features talks by master's candidates about their theses; so many students participate that six parallel sessions are needed.

VI. Results

The M.Eng. program has proven to be very popular. Typically there are between 300 and 350 students per year majoring in EECS (out of about 1000 undergraduates at MIT per year), and two-thirds of those stay for the M.Eng. degree. It was a surprise that only 65% of the M.Eng. degrees are actually given simultaneously with the S.B. degree; the other 35% of students manage to earn their S.B. in four years and decide to march at commencement with their classmates.

Accreditation was not sought for the M.Eng. program, but rather for the three S.B. programs. The VI-1 and VI-3 programs, being continuations of prior accredited programs, were accredited in 1995-96, the only difference from prior years being that VI-3, because of its name "Computer Science and Engineering" required accreditation by both ABET and CSAB. The VI-2 program, because it was innovative in not requiring specialization in either EE or CS, was more of a challenge to the accreditation process. The accreditation team visited in October, 1995. At that time the program was in the process of being strengthened to require breadth at the junior level. This change convinced the visitors that all graduates had sufficient advanced technical material in both EE and CS. The VI-2 program was accredited by both ABET and CSAB. The program was also back-accredited for two years so as to cover all its graduates.

There continues to be great interest among MIT undergraduates in EECS. Student enrollment has increased since 1994, and in particular the increase in computer science has been extraordinary. Prior to 1990 the number of students in EE was generally twice the number in CS. During the 1990s there has been a shift of interest away from EE into CS, and in 1998 there are more than twice as many VI-3 as VI-1 students (this trend is being felt at almost all universities). However, the most interesting development is that over half of the EECS students now select VI-2, the most demanding of the three programs, and the one that does not force specialization in either EE or CS. The students apparently believe that their options will remain the greatest if they are familiar with both EE and CS technical material.

Enrollment in the VI-A Internship program is about 30% lower now than it was before the M.Eng. program. Perhaps this is a result of changes in industry (more emphasis on short-term results). Or perhaps it is because some earlier VI-A students, not really interested in the industrial experience, joined VI-A only because it was a fast route to a master's degree. Without such students, whose needs are now served by the new M.Eng. program, the VI-A program is stronger, even if smaller.

After a two-year transition period when students and supervisors alike were testing the system, the shorter master's thesis is working very well. There are a sufficient number of interesting, appropriate topics, and each year there are only one or two students who lose financial aid because their theses are not finished in time. About 70% of the M.Eng. students receive some sort of financial support, typically as a Research Assistant or a Teaching Assistant. The rest pay their way through a combination of outside fellowships, loans, and personal funds.

The M.Eng. students are selected on the basis of their ability to perform academically at the first-year graduate level, whereas S.M. students are selected (from outside applicants) on the basis of their potential for doctoral work. Thus one might expect the S.M. students to do significantly better academically. This is not the case; the GPA of M.Eng. students after their first four years is 0.08 point below that of S.M. students. The acceptance rate of M.Eng. students into the doctoral program is high for those who apply. The five-year schedule is a realistic one -- the average time from entry to MIT to the M.Eng. degree is 5.1 years. M.Eng. students participate in undergraduate research at a high level, and the number of double degree programs and external minors is higher for M.Eng. students than the general population. Thus the M.Eng. program does not seem to narrow the outlook and perspective of the students.

It is interesting to ask whether M.Eng. students recover the cost of the program. Of course students should seek higher education because of the increased opportunities it gives them throughout their life, and the greater control over their own destiny. However, a case can be made that the M.Eng. is a sound financial investment. Starting salaries for MIT EECS M.Eng. and S.M. graduates are each about $8400 per year more than the average MIT EECS S.B. graduate. The financial investment of the M.Eng. degree, without any financial aid such as fellowships or RA or TA positions, is the tuition (about $24,000) plus the lost earnings for one year (about $52,000), or a total of $76,000. The payback period is thus less than ten years, even assuming that the gap between the master's and bachelor's pay does not increase with time.

In general students are pleased with the program. In the 1997 exit survey, 97% of the students said it was worth the extra effort to complete it. Aspects that were particularly appreciated included the convenience and efficiency of the seamless program, and the thesis experience. About half the respondents said they were headed directly for a technical job; the other half planned such things as working toward a Ph.D., a job on Wall Street, or medical school.

The recent search for an EECS Department Head gave an opportunity for assessing faculty opinions about the M.Eng. program. The search committee let the faculty who met with them raise issues on their own, and the M.Eng. program was a frequent topic. Many faculty feel that the idea is sound but the implementation needs significant adjustment. A suggestion often made was that admission be more selective. Some faculty blame the M.Eng. program for the lack of resources, including both faculty time and TAs, to devote to advanced graduate courses. Some faculty believe that M.Eng. theses do not advance their own research operations as much as the normal Ph.D. theses or the traditional S.M. theses and therefore do not wish to participate as fully as others. Faculty who teach first-year graduate courses report that the M.Eng. students are not as well prepared as the S.M. students (this is not surprising since they are selected using different criteria) and it is necessary to reduce the amount of material taught as a result. They also report that the style of these courses changes to serve the larger audience, often in a negative way.

Another issue of great importance is the increased enrollment in EECS. MIT allows undergraduates a free choice of major starting at the end of the freshman year, and the fraction that select EECS has now risen to over 35%. Most observers believe that the popularity of the M.Eng. program is a significant cause, but other factors include the increased pervasiveness of electronics, communication, and computation in society, the attractive career opportunities in these fields, and in general their "high tech" image. The department's ability to respond to the perceived weaknesses in the M.Eng. program will depend on getting the resources needed to cope with the large student demand.

In the administration of the program, there are many ways in which the seamless character of the program is at odds with the typical university pattern. For example, the MIT Registrar feels a mandate to compute separate GPA and grade reports for the undergraduate and graduate portions of an M.Eng. student's overall program. Hence there is a flurry of activity every year to move courses around between the graduate and undergraduate buckets merely to satisfy the Registrar.

The M.Eng. program requirements are complicated, perhaps more complicated than necessary, so student advising is a challenge.

VII. Next Steps

Like all programs, the M.Eng. program must evolve. Continued attention must be paid to implementation details, under the principle that if you don't take care of the details, then the details will take care of you.

The M.Eng. program is being reviewed in 1998, in accordance with the original 1992 faculty vote. The present paper is based in part on that review, which identified some weaknesses that threaten the program, and offered several steps that should be taken to strengthen it.

Among the important issues discussed in that review was the impact of the less well prepared M.Eng. students on first-year graduate courses. Recently the department's doctoral program was reviewed from a "clean slate" point of view, and one of the results from that study will be the development or refinement of first-year graduate courses that will serve both the potential doctoral students and the terminal M.Eng. students.

VIII. Other Similar Programs

This M.Eng. program seems to be based on a sound model of education at the various levels (bachelor's, master's, and doctoral). Overall, it is operating very successfully.

Despite the success of this program, other MIT M.Eng. programs have not been based on the same seamless five-year model. Instead, they have been one-year professional programs, rather like a typical MBA, sometimes with a requirement of industrial experience.

As far as is known, other universities have not adopted this approach either.


References

1. Penfield, Jr., P.L., J.V. Guttag, C.L. Searle, and W.M. Siebert., "EECS Plans Major Curriculum Changes," The MIT Faculty Newsletter, vol. V, no. 2, 1992, pp. 1, 16-17.

2. Penfield, Jr., P.L., 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, IEEE/ASEE, 1992, pp. 645-649.

3. Penfield, Jr., P.L., J.V. Guttag, C.L. Searle, and W.M. Siebert, "Master of Engineering: A New MIT Degree," Proceedings, 1993 ASEE Annual Conference, ASEE, 1993, pp. 58-61.

4. Penfield, Jr., P.L., W.M. Siebert, J.V. Guttag, and C.L. Searle, "MIT EECS Master of Engineering: a Status Report," Proceedings, 1994 Frontiers in Education Conference, IEEE/ASEE, 1994, pp. 228-232.


Authors

Paul Penfield, Jr.
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology
Cambridge, MA 02139-4307
Phone: (617) 253-2506
Fax: (617) 258-7354
penfield@mit.edu

Professor Penfield was born May 28, 1933 in Detroit, Michigan. He received the B.A. degree (cum laude) in physics from Amherst College, Amherst, Massachusetts in 1955, and the Sc.D. degree in electrical engineering from MIT in 1960.

He has been on the MIT faculty, in the Department of Electrical Engineering and Computer Science (EECS), since 1960. He served as Associate Head of the Department from 1974 to 1978, and as Director of the Microsystems Research Center from 1985 to 1989. Since 1989 he has been Head of the Department.

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, integrated-circuit design automation, and computer-aided fabrication of integrated circuits.

Professor Penfield is an IEEE Fellow, and former Chairman of the Boston Section. He received from IEEE the Darlington Award in 1983 and the Centennial Medal in 1984. He is a member of ACM, AES, APS, NAE, and Sigma Xi. He is the author of five books and dozens of articles in his various fields of interest. During 1996-97 he served as President of the National Electrical Engineering Department Heads Association (NEEDHA).

Jesús A. del Alamo
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology
Cambridge, MA 02139-4307
Phone: (617) 253-4764
Fax: (617) 258-7354
alamo@mit.edu

Jesús A. del Alamo received the degree of Telecommunications Engineer from the Polytechnic University of Madrid in 1980, and the M.S. and Ph.D. degrees in Electrical Engineering from Stanford University in 1983 and 1985, respectively.

From 1977 to 1981 he was with the Institute of Solar Energy of the Polytechnic University of Madrid, working on silicon solar cells. At Stanford University he carried out his Ph.D. dissertation on transport in heavily doped silicon. From 1985 to 1988 he was research engineer with NTT LSI Laboratories in Atsugi (Japan) where he conducted research on heterostructure field-effect transistors based on InP, InAlAs, and InGaAs. Since 1988 he has been with the Department of Electrical Engineering and Computer Science of Massachusetts Institute of Technology where he currently holds the title of Professor. His research interests include high-power high-frequency High-Electron Mobility Transistors based on compound semiconductors, Si Bipolar Transistors and Si Metal-Oxide-Semiconductor Field-Effect Transistors.

From 1991 to 1996, del Alamo was an NSF Presidential Young Investigator. In 1992 he was awarded the Baker Memorial Award for Excellence in Undergraduate Teaching at MIT. In 1993 he received the H.E. Edgerton Junior Faculty Achievement Award at MIT.


URL of this page: http://www-mtl.mit.edu/users/penfield/pubs/meng-results.html
Created: May 16, 1998  |  Modified: Feb 25, 1999
Related pages: Abstract  |  Presentation  |  Penfield publication list
Site map  |  To Paul Penfield's home page  |  Your comments are welcome.