What Can a Physics Teacher Do with a Computer? (Part 1)

Edward F. Redish

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

Invited talk presented at Robert Resnick Symposium RPI, Troy NY, May, 1993

Go to part 2. (References are at the end of part 2.)


When Bob Resnick[1] and David Halliday undertook to change the way introductory physics was taught in the late 1950s, the computer was largely restricted to research labs. As an upperclassman in 1960, I was delighted to be allowed to use a Marchant mechanical calculator. It was driven by electricity so you didn't even have to turn a crank! Each multiplication was performed by decimal wheels spinning in repeated additions. To multiply two six-digit numbers took about three seconds.

Today's computers are not much bigger than that Marchant calculator and more than a million times faster. But more than that, the computer can do a lot more than just manipulate numbers. Modern personal multi-media computers seem closer to the television that my students watch these days instead of reading books. This changes a lot. Can these new computers be used effectively in teaching introductory physics? Teaching and learning physics has always been difficult and using innovative technology, especially at the introductory level, has not been particularly successful. Moreover, research in physics education has shown that at the introductory level, traditional methods do not help the majority of our students achieve the most basic and elementary goals we have for them.[2]

It is in this context that we must assess the application of modern computer technology to physics education. Can the computer help us overcome the difficulties that have become apparent through research in physics education? Or will it simply take us one more step down the path of slow, accumulating deterioration (SAD) that can be seen in the way that many introductory physics classes have changed over the years?

To judge the value of any innovation in physics education, we have to consider three important questions:

  1. What are our detailed goals for our students?
  2. What is the state of our students' knowledge and expectations of learning when they begin?
  3. What can we do to help them change the state of their knowledge?
  4. Only in the context of these answers can we effectively ask the question: Can the computer help?
I first focus on articulating the issues of teaching and learning that might be relevant for including computers, a step often ignored or suppressed. Then, I review some uses of the computer that have either been demonstrated to facilitate learning or seem to have good prospects for doing so.

Physics education research

Where are our students when they come to us? Listening

Over the past decade and a half, a number of physicists and educators have begun to treat the problem of student learning as a scientific one. To do this, one must

  1. shift focus in the course from the content to the student;
  2. accept the idea that we want to model our students' knowledge, both statically and dynamically;
  3. listen carefully to our students and find out what they think and know both before and after instruction.

Adopting this viewpoint requires us to make a significant change in our attitude towards teaching. In physics education research (PER), extensive observations (including long interviews with dozens of individual students) are made to find out what students really think. It has been compellingly demonstrated that our traditional assessments (which are used mostly to certify and hence filter out some groups of students) do not often yield this information (see for example ref. 2 and references therein).

What PER studies have taught us is that most of our students, especially at the introductory level, learn far less in our classes than we had thought. They come to us with strongly held ideas about what they are supposed to do and how the world works, and many of these ideas conflict seriously with what we want them to learn.[3]

From this research, and from research in cognitive science, we learn a number of important and relevant facts.[4] Perhaps the most important is that all students "construct" their ideas and observations -- pulling together what they see and hear into a (not necessarily coherent or consistent) mental model.[5] Mental models have some surprising characteristics.[6] They may contain contradictory elements. Different elements of a mental model may become confused because of superficial similarities. Access to elements of the mental model may not be automatic and appropriate. The access links are a part of the model that must also be built.

These results remind us of the fact (well known to good teachers for decades![7]) that we have to worry not only whether students "got" the content but whether they can also get to it when they need it and use it appropriately.

Another fact that has become clear from the result of much detailed research is that standard problem solving does not always help students build the appropriate mental models.[8] Part of the difficulty is that the mapping from a mental model to a problem output is not one-to-one.

Understanding the basic concepts of physics

Students’ experience with the world has led them to construct a set of beliefs about how the world works. Students misunderstand basic ideas of physics, in part because of their previous experience and in part because of the way we use language. In addition, “common sense” speech and reasoning are "fuzzy". Similar concepts are not differentiated, resulting in confusions that seem bizarre to a physicist who is accustomed to precise reasoning and operational definitions.

It has been demonstrated by extensive PER experiments that a majority of students entering (and leaving!) a typical calculus-based physics class

Furthermore, the student's success on typical "end-of-the-chapter" problems has little correlation with developing an understanding of the deeper concepts.[13]

Knowing what science is and how it works

In addition to problems with the basic concepts of science, students in introductory physics courses often have deep misunderstandings of the nature of science. They often have little idea about how science works or how one does it. (Often, a physics major does not learn this last until graduate school!) Not only do many students get the basic concepts of physics wrong; their very idea of how science works and how knowledge is obtained is often wrong. , ,

The activities of a typical introductory class do little to disabuse our students of these deep structural misconceptions.

Re-forming an existing pattern may be hard
Once we've realized what difficulties our students have, why don't we just fix them? Unfortunately, it's not as easy as that -- a result which we might well have guessed by the repeated failure of innovative curricula designed without detailed testing and modeling of students’ responses.

If you're on the wrong track, it can be very hard to readjust. Getting students to make a "gestalt shift" in the way they see the world can be difficult. A classic example in cognitive psychology is the ambiguous picture of a woman in Fig. 2.

This picture can be viewed in two ways, as a young woman or as an old woman.[17] If students are presented with this picture in a less ambiguous form (certain features are enhanced to make it clearly the young or old woman) and then shown the ambiguous one, they find it exceedingly difficult to make the transition to the other, despite detailed explanations from members of the other group.

There is extensive documentation in the cognitive literature about the difficulty of transforming a well-established mental model.[18] (See ref. 1 and refs. therein.) By now there is also explicit documentation of how persistent are the mental models that students bring into a physics course and how these mental models make it difficult for students to understand what we are trying to teach them.

What do we want our students to learn? Deciding on our explicit goals

As a physics teacher I am not satisfied to have my students memorize a few equations and algorithms and be able to apply them in limited examples. I refer to this collection of ideals as a robust knowledge structure.

It is clear to most physics teachers that very few students achieve such a structure by the end of their introductory physics course. Indeed, many graduate students do not possess one. Is this goal ridiculously out of reach? I do not believe so. Educational efforts with non-traditional approaches have demonstrated that a very large fraction of our students can begin to build real understandings of physics and how it is done even at the introductory level.[19] I operate under the hypothesis that most of our students can succeed at developing a robust knowledge structure of physics, even at the introductory level, given appropriate time, materials, and environments.

How can we help them get there? Guidelines for effective instruction from PER

Of course the problem with the sentence at the end of the last section is that we may not be able to give the students the appropriate time, materials, and environments they need to develop this knowledge structure. Furthermore, some of the necessary elements may be beyond our control, such as whether students have the time, motivation, and background to study adequately independent of coterminous jobs taken to support their education or the presence of tempting and diverting social activities.

Given these limitations, it is essential that we do our best to identify those elements which hamper students and about which we may have some control: the cognitive ones. Even students who come to class, who listen to correct and well-delivered lectures, who take notes, who attend and perform laboratories, often do not learn from those activities what we expect them to.

Two of the most important observations are:

  1. If students have mental models that differ from the professor's, they may not interpret the presented material in the desired way.

  2. If the activity in question can be performed without engaging the student's mental model, the learning that takes place will be disconnected and superficial.

How can we get students to both hear what we are trying to say and change their deeply held ideas? A number of principles have been developed as a result of research in physics education, both as to the kind of activities that tend to be successful and the kind of subtle mistakes and misunderstandings that we have to watch out for. These are documented extensively in the books and papers by Arnold Arons and Lillian McDermott. See their talks in this volume and Arons' book (ref. 3). I will only mention a few that I think are relevant for the topic of this meeting:

  1. Go from the concrete to the abstract. (Make the link to the real world. Concept first, name second)

  2. Put whatever is new into a known and understood context.

  3. Make students articulate what they have seen, done, and understood in their own words.

  4. To change people’s ideas, you must first get them to understand the situation, then make a prediction, and finally, to see the conflict between their prediction and their observation.

  5. Telling someone something often has little effect in developing their thinking or understanding. You have to get them to do something, but "hands-on" activity does not suffice. It must be "brains-on".

  6. "Constructive" activities in which students feel they are in control are much more effective than activities in which the students are being shown results, no matter how eloquently or lucidly the results are presented.

These observations suggest that modern computer technology might help, but they also suggest a division between two kinds of computer activities:

Can the computer help? Some constructivist applications

Let us now combine:

in order to identify computer applications which can provide effective elements of an instructional program. We’ll consider three types of uses.

Go to part 2.

Edward F. Redish
University of Maryland