"Think aloud interview protocols from three high school post-physics students who interacted with a relative motion computer simulation presented in a predict observe explain format are analysed. Evidence is presented for: qualitative and quantitative difficulties with apparently simple one-dimen-sional relative motion problems; students spontaneous visualization of relative motion problems; the visualizations facilitating solution of these problems; and students memory of the on-line simulation used as a framework for visualization of post-test problems solved off-line. Instances of successful and unsuccessful mapping of remembered simulation features onto target problems are presented. Evidence from hand motions and other indicators suggesting that the subjects were using dynamic imagery in mental simulations during the treatment and post-test is presented. On the basis of these observations, it is hypothesized that for successful students, dissonance between their incorrect predictions and simulations displayed by the computer initiated the construction of new ways of thinking about relative motion, and that the memory of certain simulations acted as an analogue framework for visualization of target problems solved off-line after the intervention."
"Because evidence concerning students learning processes and reasoning with mental simulations requires a deep level of analysis, this study examines a few students in depth, rather than a more cursory examination of a larger number of students."
"A remaining weakness in the literature is the difficulty of studying the learning process in enough detail to determine where elements of teaching strategies are succeeding or failing to contribute, and to uncover new factors in learning. Also, the social learning processes occurring in large and small group discussions seem integral to classroom teaching in this area, but the video tapes in this mode in our pilot studies did not yield enough information on the learning processes occurring in each student at the grain size addressed in this paper. Therefore, we decided to interview students learning on their own in this particular study in order to increase our ability to track each student's learning process."
Aardig om te lezen hoe protocol-analyse verweven is in het artikel en gebruikt wordt om veronderstellingen te ondersteunen en nieuwe ideeën genereerd. Onbevredigend is het volgende. Leerlingen gebruiken ervaringen met de simulatie om nieuwe problemen (off-line) op te lossen. Maar de vraag is in hoeverre ze ook werkelijk hun eigen oplossingen begrijpen. Ze zijn in ieder geval niet geworteld in voorkennis en/of kennis van de wereld om ons heen:
"If this view is correct, then we can say that the two subjects solved post-test problems off-line via dynamic mental simulations, and were not dependent on the computer to run the simulation. This seems crucial if students are to take deep understandings away with them after using a computer simulation. Thus, there is evidence that appropriate mental simulations of target problems can be fostered by interventions consisting of using a computer simulation in a predict observe explain activity. Our hypothesis then for why the intervention was successful for these two subjects is as follows. ( 1) They were initially stimulated to develop a new understanding by the recognition that their prediction was false after viewing the simulation. ( 2) They then constructed some new understandings by explaining the simulated event, especially for the case of the car whose apparent direc-tion of travel reverses. ( 3) (...) there is evidence that they were able to use their memory of the simulation in an analogue fashion to generate correct imagery in the target situation. ( 4) There is also some evidence that this imagery was dynamic and allowed them to run a qualitative mental simulation of the target situation. ( 5) In quantitative problems, the subjects used qualitative comparisons of relative speeds or directions of travel to determine appropriate additions or subtractions to perform at a quantitative level."
"Examination of this case leads to several hypotheses concerning the difficulties that may be encountered by students who use recollection of a simulation to guide their solution of target problems. First, the student may not understand the base, i.e. the computer simulation. This may be due to a lack of conceptual understanding of concepts and/or terms central to the computer simulation; without such understanding the student may not be able to accurately process the feedback provided by the simulation. Second, the student may inaccurately transfer knowledge gained during interaction with the simulation to an analogous case which is not truly analogous."
Nog op te lossen probleem dat ze in de discussie opvoeren is ten aanzien van het belang van het werken met de simulatie als stap naar een "general schema or model":
"Forming a visualizable model with such general properties would comprise what White (1993a) refers to as an intermediate abstraction i.e. a representation that is more general than single examples, but which is not as abstract as mathematical formalisms. Such visualizable models are considered by many to be central in the thinking of practicing scientists; there is growing evidence that they may also be central for understanding in science students."
Tot slot in critique and further research:
"If this last view has merit, it leads to questions like: What makes a computer simulation memorable enough but basic enough that it can act as a key exemplar? We need to explore this idea more carefully with reference, e.g. to Minstrell s (1996) idea of benchmark examples in physics teaching, and perhaps to Kuhn s (1977) idea of paradigmatic exemplars in science."
"At the methodological level, imagery has always been extremely difficult to study because it has so few behavioural indicators, none of which are very direct. From these case studies we cannot predict the frequency with which these will be exhibited. However, the cases do serve to provide existence exemplars establishing the possibility of such indicators, which was our purpose here, rather than the purpose of providing estimated average frequencies of such indicators over a particular population. The development of methodologies for recognizing imagery indicators would give us a powerful tool for detecting the form of student's knowledge structures and aid in discriminating between rote algorithms and conceptual understanding. Appropriate application of experience gained through interaction with computer simulations is a complex skill. Its full development would undoubtedly require more support than was offered in our interventions. Further research should be conducted to identify and describe additional factors that affect students ability to solve and visualize problems following computer simulations in order to develop improved pedagogical strategies for using them. An implication of these results for curriculum development is that formative evaluations that include data on the conditions under which students visualize target problems following use of a simulation can and should be an early component of development projects. Such data should guide and improve both the developer's design of educational computer simulations and the teacher's design of learning activities which employ them."