JOURNAL OF RESEARCH IN SCIENCE TEACHING VOL. 30, NO. 4, PP. 367380 (1993)
SelfGenerated Analogies as a Tool for Constructing and Evaluating
Explanations of Scientific Phenomena
E. David Wong
Department of Teacher Education, 206 Erickson Hall, Michigan State University,
East Lansing, Michigan 488241034
Abstract
How can students be taught to develop explanations for scientific phenomena on their own when their background knowledge is incomplete or poorly organized? Evidence from historical accounts of scientific discovery suggest that selfgenerated analogies—analogies produced by the learners themselves—are a tool by which individuals can generate, evaluate, and modify their own explanations. The central research questions for this study were: Can students use a series of selfgenerated analogies to bring about change in their understanding of a given scientific phenomenon, and what is the nature of the change in understanding? Participants were asked to create, apply, and modify their own analogies—as opposed to applying a specific analogy provided by an outsider—as a heuristic for constructing, evaluating, and modifying their own explanations for a given scientific phenomena. Nontrivial changes in explanation facilitated by the use of generative analogies were observed. Changes in understanding ranged from the emergence of new explanations to the raising of important questions about the nature of the phenomenon.
With dramatic flair, the ninthgrade science teacher flourishes the bathroom plunger in front of the class. Then, in one quick move, he tums and squishes it against the hard surface of the blackboard. The entire class gasps, then giggles as the plunger remains firmly stuck . . . horizontal.
"Why is this happening?," the teacher asks, motioning to the plunger.
"Suction!" "A vacuum!" are the common responses heard above the clamor of students simultaneously answering.
"Good thinking!" replies the teacher. "Now, how exactly does a vacuum or suction make the plunger stick to the surface?"
The class quiets noticeably. After a pause, one student says, "When you push the plunger, a vacuum is created inside and that makes it stick. You know what I mean?"
The teacher pushes the students a little harder. "Okay. But, why does a vacuum make the plunger stick?"
1993 by the National Association for Research in Science Teaching
Published by John Wiley & Sons, Inc. CCC 00224308/93/04036712
368 WONG
The initial enthusiasm subsides even more. Some students are thinking that the question had already been answered. Others see that the teacher is asking for a deeper explanation, but do not know how to proceed.
The teacher begins to ask individual students what they think. He is met with "I don't know's," shrugged shoulders, or silence.
"It's okay that you don't know," he says brightly to comfort them. "I want you to think of your own explanation. Push yourselves a little. Make a conjecture like a scientist might!"
No response at all. This is the quietest the class been since he announced the grading scale for the semester exam. After a long awkward silence, a student in the back, sitting low is his seat with his arms folded across his chest, speaks loudly—as if for the whole class.
"We said we don't know. Why don't you just tell Us the answer?!"
As this vignette derived from my own teaching experiences illustrates, science students will often find that moving beyond the boundaries of concrete memorized facts, into the conceptual gray area where understanding is tenuous and incomplete, is an unfamiliar and uncomfortable experience. Typically, students are neither expected nor prepared to construct accounts of scientific phenomena on their own. Instead, they see their role as students to be followers, led carefully by teachers to new and better understanding. Teachers often share the same view.
When individuals are out in the real world, trying to figure out whether hot or cold air is best to defog a car window, or striving to design a better space shuttle Oring, they will often be compelled to move into that gray area of understanding. To say "I don't know," or to defer to a more knowledgeable authority, is often not an option considered or available. An important part of a science curriculum, therefore, should be to develop students' ability to construct their own understanding for scientific phenomena.
The goal of this study was to develop a means by which students could develop, evaluate, and modify their own explanations for scientific phenomena with minimal subjectspecific guidance from the teacher. The potential of selfgenerated analogies —where learners use their own analogies as opposed to analogies provided by a teacher—was explored as a tool for advancing conceptual understanding. In brief, participants were presented with a scientific phenomenon to explain. When they encountered difficulties in their explanations, they were asked to construct their own analogies as a means of addressing these conceptual problems. Instead of receiving an ideal analogy from a more knowledgeable authority, participants were required not only to create and apply their own analogy, but to also evaluate and modify the analogy. The central research questions for this study were: (a) Can individuals use a series of selfgenerated analogies to bring about change in their understanding of a given scientific phenomenon? (b) What is the nature of the change in understanding?
Conceptual Background
The process of understanding andproblem solving. In schools, science instruction often follows a common path: The concepts are taught; the problems are provided; the concepts are used to solve the problems. The conceptual understanding necessary
SELFGENERATED ANALOGIES 369
for solving the given problems is assured by prior or concurrent instruction. Furthermore, problems and the requisite knowledge are presented in close conjunction with each other, making it easier for the individuals to determine which conceptual knowledge to apply. By contrast, in situations outside of school, the relationship between a problem and the knowledge appropriate to address it is not as obvious. In fact, the nature of the problems themselves is often not clear and is strongly influenced by the individual's prior knowledge and interpretation of the situation. Schon (1979) argues that in human learning and understanding, the process of probleminding is as central, if not more so, to understanding than the process of problem solving.
Problem finding is a direct function of conceptual understanding: What one knows affects the types of problems perceived and the manner in which these problems are construed (cf. Larkin, McDermott, Simon, & Simon, 1980). This contrasts to typical schoolwork, where identifying problems is either a nonissue because the problems are stated explicitly, or is a function of understanding task demands. For example, keeping in mind the most recent topic of instruction is usually a failsafe strategy for determining the nature of a problem on a test. Reliance on these types of contextual factors may be one reason why students often do so poorly on final exams where the nature of problems cannot be as easily determined by factors unrelated to conceptual understanding.
The potential of analogies. The rationale for the study is based on the following argument: Because of the nebulous nature of many problem situations outside of school contexts (especially in contrast to traditional instructional contexts), students need to develop the ability to identify and represent scientific phenomena in a manner that enables them to gain a greater understanding on their own. Science instruction, therefore, should enable students to identify and work on problems with a minimum of reliance on teacher guidance and instructional cues, and with a maximum emphasis on bringing to bear their own existing knowledge about the particular phenomenon. In this study, selfgenerated analogies were investigated as a tool by which individuals can construct—largely on their own—better accounts of scientific phenomena. The conceptual argument and design underlying the study was informed by an examination of analogy use in two different settings: the classroom community and the scientific community.
Analogies in the classroom. In instructional situations, the analogy is typically constructed by a knowledgeable individual (teacher) and is used as a means to represent a concept to those who are less knowledgeable (students). Studies demonstrating or exploring the power of instructional analogies have varied in their emphasis from using analogies to overcome student misconceptions (Clement, 1987; Spiro, Feltovich, Coulson, & Anderson, 1989; Stavy, 1991), to analogies as an aid to comprehension and memory when reading text (Halpem, 1990), to general discussions on the benefits of analogies in instruction (Duit, 1991; Simons, 1984).
In each of these studies, analogies are provided by an outside authority—the role of the individual is to make sense of the received analogy. Individuals have minimal responsibility for representing the scientific phenomenon in their own terms. Although these studies provide important evidence that analogies do facilitate leaming and understanding, they do not suggest that selfgenerated analogies might be of value.
Analogies in the practice of science. The nature of classroom communities— where experts teach novices—promotes the utility of teacherprovided analogies. Selfgenerated analogies might assume a more central role in a context where members all see themselves as learners. Such a context exists in the scientific community, and in d~e practice of science can be found naturally occurring selfgenerattd analogies.
370 WONG
Historical accounts of scientific discovery and insight describe compelling examples of the potential of analogies (cf. Dreistadt, 1968; Hesse, 1966; Leatherdale, 1974; Nagel, 1974; Oppenheimer, 1956; for accounts of discovery through analogy in mathematics, see Polva, 1954). The spontaneous and unpredictable use of analogies makes it difficult to examine the phenomenon as it occurs—most accounts are based on scientists' retrospective reconstructions of their insights. However, a rare example of a direct investigation of scientists' use of analogies as a natural part of their work was provided by Schon (1979). Schon observed researchers at a paintbrush factory as they tried to develop a better paintbrush. During their brainstorming sessions, significant insights were made through the discovery and application of an analogy that compared a paintbrush to a pump. Considering the brush as a pumping device, instead of simply a spreading device, led them to design brush bristles that more effectively hold and squeeze paint when used. Schon argued that the power of the analogy lies in its ability to lead to a reframing of the original problem. The altering of the problem space facilitates the emergence of new ideas or solutions previously unaccessible in the original problem space. Schon contends that it is impossible for new ideas to arise directly from existing conceptions; only through the altering of some initial assumptions can truly novel ideas emerge. (Note: Clement's 1988, study of problemsolving methods of physics experts, although conducted in an artificial setting, also provides examples of spontaneous use of analogies.)
Design of the Task
Although selfgenerated analogies can occur naturally (as evidenced by scientific discovery), their use is typically spontaneous and unpredictable. One of the goals for this study was to develop a task that explicitly encouraged the systematic creation and application of selfgenerated analogies as a heuristic for conceptual growth. Theoretical work from psychology and philosophy on analogical reasoning and problem solving was reviewed to identify processes that would seem likely to promote conceptual growth (Clement, 1988; Petrie, 1979; Rumelhart & Norman, 1985; Schon, 1979; Vosniadou, 1989). These processes, upon which the task for this study was based, include
1. Adopt a generative learning perspective: Understanding is viewed as an iterative,constructive process where the purpose of the task is to develop a plausible explanation rather than tofind the single correct answer (Vosniadou & Brewer, 1987).
2. Work on a meaningful problem: The starting point is a problem formulated from each individual's unique understanding of the phenomenon rather than a problem provided and defined entirely by an outsider (Petrie, 1979; Schon, 1979).
3. Engage in concrete activity with the phenomenon: Sensory activity may stimulate associations that are not yet possible through abstract, verbal means (Petrie, 1979; Schon, 1979).
4. Generate analogies: Representing the phenomena by relating it to a familiar domain of knowledge facilitates understanding (Duit, 1990; Simons, 1984).
5. Apply analogies to the phenomenon: Specifying exactly how "A is like B" provides different perspectives from which one's understanding can be examined and may yield relationships not recognized when the analogy was first generated (Clement, 1988; Gick & Holyoak, 1983; Petrie, 1979).
SELFGENERATED ANALOGIES 371
6. Evaluate the analogies: Any analogy, by definition, suggests inappropriate similarities that need to be identified (Spiro et al., 1989).
7. Modify existing analogies or generate new analogies: A single analogy is not commensurate with understanding or the comct explanation. Therefore, the use of selfgenerated analogies is a cyclical process of geneMting, evaluating, and modifying.
Methods/Task
Participants. The influence of selfgenerated analogies on conceptual understanding should, ideally, be explored for all students of science. Although such a population includes students of all ages, the present study focused on university students as a starting point for investigation. With the thinkaloud methodology used, these individuals were likely to have a better developed ability to express their thought processes. Future studies will expand the focus of investigation in two directions: toward younger, schoolaged students of science, and toward wholeclass rather than individual activity.
E]even individuals (4 men, 7 women) from the teacher education program at a major Califomia university were recruited to participate in the study. Participants did not receive academic credit or monetary reimbursement for their involvement. They were recruited from a variety of different subjectmatter areas (physics, chemistry, geology, biology, foreign language, English) to provide a range of prior knowledge for the task phenomena.
Procedure. Participants were presented with a piston/cylinder device (a large 50cc syringe; see Figure 1) that demonstrated the following three air pressure phenomena:
• Compression: As the plunger is pushed into the syringe with the nozzle covered by a finger, the amount of force required to move the plunger increases.
• Decompression: As the plunger is pulled out of the syringe with the nozzle still covered, the amount of force required to move the plunger increases.
• Equilibrium: When the plunger is released after being pushed or pulled, it returns toward its initial position.
A set of heuristic procedures (see the foregoing) intended to foster understanding through selfgenerated analogies was designed. During the task, participants worked alone in the presence of the facilitator (myself). They were told that the purpose of the study was to examine how they came to produce and modify explanations about the air pressure phenomena. As a facilitator, I emphasized that I was not concemed about whether their explanations were correct or not; instead, it was the process by which they came to understand that was of interest. In spite of these assurances, participants would often look to me for feedback on their explanations throughout the