Discipline-based education and education research:
The case of physics -- Part 2

Edward F. Redish

Department of Physics, University of Maryland, College Park, MD 20742-4111

To return to part 1 of this paper, click here.

Models of the student

The theoretical model of learning that underlies the planning and interpretation of experimental results in physics education research is often tacit, especially for researchers within the discipline of physics itself. Nonetheless, the model is clearly present and draws broadly from research in cognitive science and education. I refer to this model of student learning as the cognitive model.

The cognitive model

This model may be summarized very briefly by a few key phrases:

Students build their knowledge by processing the information they receive. (constructivism)
What students construct depends on the context -- including the students' mental states.
Producing significant conceptual change is difficult and can be facilitated through a variety of known mechanisms.
Individuals show a significant variation in their style of learning along a number of dimensions.
For most individuals, learning is most effectively carried out via social interactions.

(For details, see the books of Arons[16], Byrnes[17] or my recent paper.[18]) Some implications of this model that are crucial for our current discussion are:

Individuals can support incomplete and even simultaneously contradictory ideas.
The use of knowledge is context dependent.
Spontaneous thought is rarely metacognitive.
For most students, scientific thought and mathematically logical thinking not spontaneously occur and must be explicitly taught.
The logical structure and epistemology of the content must be carefully analyzed and taught explicitly.

Since they haven't thought about it much, most physics faculty have a model of student learning that is less well-formed and consistent (and indeed, may contain incomplete and even simultaneously contradictory ideas). Nonetheless, by observation of practice, we may infer some implications of the traditional transmissionist or broadcast[19] model.

The broadcast model

Previous knowledge is not relevant. (Students are blank slates.)
Knowledge is binary. (You either know it or you don't.)
The student is idealized. (Students possess good motivation, independence, a knowledge of what to do, and a willingness to do it.) If the student differs from this ideal image, it's their fault.
The student is assumed to be metacognitive. (Students learn from their mistakes).
Scientific thought and rational thinking are taken to be natural -- even obvious.

The first element is hard to overcome, even for teachers whose consciousness has been raised on educational issues. I can provide a specific example from my own experience. In January of 1994, the Physics Education Group (PEG) at the University of Washington reported the results of a study of engineering students' responses to being taught about magnets.[20] Traditionally, teachers and textbook writers assume that students know little about the subject, so a good way to introduce it is by analogy with electric charge, the topic typically presented just before magnetism. The Washington PEG demonstrated that upon entry over 80% of the students confused electric charges and magnetic poles. After a traditional instruction, this number remained above 50%. I was both flabbergasted and distressed at hearing this. I have taught that subject off and on for over 25 years and was teaching it at the time of the presentation. I furthermore believed that I listen carefully to students, and am sensitive to the issue that students brought in previous knowledge. Yet I had never imagined such a confusion was possible, much less common.

Needless to say, I probed my class upon my return and found exactly the same results as the Washington group. The second element of the broadcast model has serious implications for testing. Ignoring the context dependence of student knowledge means that one is satisfied with a less than robust student response on exams. My colleagues are often happy if a student recognizes and can replay a principle when they are given a large number of contextual clues. They tend to consider my exam problems that probe whether students can recognize the context in which a principle is relevant to be too difficult. Once, during a colloquium I was presenting on physics education research at a physics department, a colleague complained that the questions I had shown from the FCI were "trick questions". When I asked him what he meant by that, he replied: "To answer them, you have to have a good understanding of the subject matter."(!)

Elements 3-5 of the broadcast model come naturally to highly trained scientists, especially if one is operating under the assumption that one should be acting as a filter to select those few students for whom these items are in fact true.

Why is it so hard to produce change?

The implicit presence of the broadcast model in colleges and universities throughout the country helps us explain at least some of the barriers to reform. Others arise through structures in the system.

One barrier arises from the way colleges and universities currently evaluate their teaching. If a class contains a large fraction of students who are used to "learning without thinking" by rote learning of algorithms and pattern matching, faculty evaluation by "student happiness" leads to testing which encourages students to keep inappropriate learning strategies. Students tend to ignore the faculty's verbal admonitions if they are not confirmed by examinations. An instructor who holds the broadcast model of student learning may be satisfied with the results, despite the fact that little real learning has taken place. This pandering to the students produces a "slow accumulative deterioration" (SAD) of course quality. After a few years of this, expectations for learning can become disappointingly low.

A second barrier arises from the lack of communication between instructional faculty and learning specialists. If advances produced by researchers in the learning sciences are to produce change in faculties in the discipline of physics, information must flow between the groups. For example, very few physics faculty are aware of results from educational research on group learning or learning styles. On the other hand, educational specialists are often unaware of serious subtleties lying behind apparently "simple" physics. Communication tends to be difficult for a number of reasons.

Physicists and cognitive scientists speak different languages. Both sides often assume a large vocabulary and a number of technical concepts.
Reading the literature in the learning sciences is difficult for physicists. The fundamental technical terminology is complex and may be used by different workers in different ways.
Reading the literature in physics is difficult for educational specialists. The language tends to be highly technical and mathematical.
Much of the research done in the learning sciences is qualitative. Physicists are not used to evaluating or building inferences based on qualitative research.

A further difficulty is produced by extremist views in both fields.

Radical constructivists raise a barrier to communication with scientists. Many of their statements look to physicists as either ignorant of science or anti-science.
Some physicists, unaware of recent gains in cognitive science and education, reject research in education out of hand as unscientific and having little to offer.

Even for educational research done within the discipline of physics, communication is weak with the rest of the physics community. I have no explicit data, but given the lack of communication mechanisms, I would be surprised if more than 2% of tertiary physics teachers were aware of students' confusion between magnetic poles and electric charge two years after its discovery and documentation.[21]

Finally, change at the tertiary level is difficult because physics faculty lack knowledge about models for building non-traditional, more effective learning environments. Physicists (perhaps most academics) tend to be selected from that small group of students who learn effectively from lecture. Some physics faculty members who are seriously concerned about reaching their students have made a substantial effort to be more effective -- in lecture -- and found little no effect. This discourages them. They may then oppose change, because they infer that improvement is impossible and that the difficulty is the students' fault. (It's easier to blame the students than to blame themselves or to change educational structures they take for granted.)

WHAT DO WE NEED TO DO TO REFORM TERTIARY PHYSICS INSTRUCTION?

Some changes are indeed taking place in the teaching of physics. Pressure on physics departments to be more responsive and innovative is being brought to bear by university administrations and state legislatures (not to mention by parents) all over the country. Yet much of the work that is done as a result is done in ignorance of results from the learning sciences. This is highly inefficient. Many innovations fail, drop out of sight, and remain invisible, to be tried again and again by colleagues at other colleges and universities who are unaware of previous failures. If significant change is to occur in tertiary scientific instruction, the light of research -- carefully done, documented, and cumulative as a community activity -- must be used to illuminate this work. If progress in the learning sciences is to have an impact on physics instruction, a number of initiatives need to be undertaken and supported over the next 10-15 years.

We need more research on specific student difficulties. (Many subjects of importance have received little or no attention.)
We need new presentations of the results of the learning sciences in a format that can be understood by discipline specialists in a place where they will read it.
We need a strongly supported effort for faculty development at the tertiary level. Here, it is not enough to aim at new faculty. Established (tenured) faculty are the ones who make decisions and provide guidance.
The transformation of tertiary instruction in scientific disciplines to bring it into better accord with the results of the learning sciences will be difficult. There is no quick fix. A sustained and well-coordinated effort is required. This implies that two more elements must be added.
A strong cadre of specialists firmly grounded within the discipline are needed.

These discipline-based education researchers are essential for two reasons.

First, much of the information needed to reform instruction requires a sophisticated knowledge of physics. Even the transformation of traditional elements often requires disciplinary experts.

Interactive engagement instruction in physics often relies on the use of discovery laboratories using real-world hands-on experiments. Real world systems are significantly more complex than the traditional idealizations present in textbooks. No real world system is ever truly "simple".[22]
Rethinking instruction often requires a deep rethinking of the physics. Traditional textbook presentations often ignore deep issues which physicists have become used to but which students find confusing.[23]

As new technology and new tools become common among professionals, new curricular elements will have to be developed, combining the insights of disciplinary specialists with those of education specialists. Two obvious questions that will need to be answered in the next few years are:

Symbolic manipulators are becoming widespread. Many colleges and universities are even beginning to introduce them at the introductory level. What is their implication for student understanding of the role of mathematics in science?
Quantum devices such as lasers, LEDs, transistors, charge-coupled devices, etc. are becoming important parts of every technologists daily toolkit. What problems will there be in trying to bring an understanding of quantum theory to a wider group of students?

Second, educational specialists within physics can work to help link the discipline of physics to the disciplines of education and cognitive science in a way that cannot be done by extra-disciplinary education specialists. Intra-disciplinary educational specialists are needed to provide a transduction[24] between their discipline and the learning sciences. Through prolonged contacts with their own colleagues inside physics departments and through frequent colloquia and seminars in other physics departments, they can carry the message to their colleagues in physics in a way that is not possible for educational specialists.

Groups of physics-based educational specialists are springing up spontaneously within physics departments throughout the country,[25] but the growth is slow and often meets considerable resistance from traditional disciplinarians.

Finally, discipline-based educational researchers not only need support, they also need strong links to similar researchers in other scientific disciplines, and to researchers in cognitive science. This suggests my final proposed initiative.

5. Mechanisms are needed to increase the interactions between discipline-based educational researchers and researchers who study the fundamental principles of learning science. The rapid changes in the workplace being produced by technology is not a short-lived phenomenon. The pace of change over the past few decades has been very rapid and can be expected to continue or even to accelerate. If tertiary education is to respond effectively to these changes, we need to build new systemic structures -- structures that will permit a continuous review of our educational content and our educational effectiveness. Building a system that will encourage the continuous quality improvement of our technical education system is going to require a long term concerted effort, both from disciplinary educational specialists and from specialists in education and cognitive science.

[16] Arnold Arons, A Guide to Introductory Physics Teaching (John Wiley and Sons, NY, 1990).
[17] James P. Byrnes, Cognitive Development and Learning in Instructional Contexts (Allyn and Bacon, Boston MA, 1996).
[18] E. F. Redish, "Implications of Cognitive Studies for Teaching Physics", American Journal of Physics 62 (1994) 796-803.
[19] Jill Larkin, this volume.
[20] P. A. Krause, P. S. Shaffer, and L. C. McDermott, "Using research on student understanding to guide curriculum development: An example from electricity and magnetism", AAPT Announcer 25 (Dec., 1995) 77.
[21] Fewer than 200 teachers attended the session where this result was presented, and no report on it has been published as of this writing. In part the lack of publication arises from a lack of a place to publish such a result that would be read by the appropriate audience of physics teachers.
[22] For example, traditional texts treat motion as if it were described by a smooth, differentiable mathematical curve. In fact, a close look at any motion will show noise at a small enough scale. The actual path is closer to a fractal than to a smooth curve.
[23] An important example of this is Hestenes's observation that Newton's Laws contain a usually unstated principle which I refer to as "Newton's Zeroth Law": Objects respond only to forces acting on them directly and only at the instant considered. (See David Hestenes, "Modeling games in the Newtonian World," Am. J. Phys. 60 (1992) 732-748.) [24] From the Latin "to lead across". In physics this term "transducer" is used for a device which changes energy from one form to another.
[25] Such groups now exist in physics departments at the University of Washington, the University of Maryland, the University of Massachusetts at Amherst, Kansas State University, the University of Nebraska, Ohio State University, and Arizona State University, among others.