ON COMPLEXITY AND RELEVANCE IN EDUCATIONAL RESEARCH FRANZSCHOll TECHNICALUNIVERSITYDRESDEN
The paper by Alan H. Schoenfeld “Toward a Theory of Teaching-In-Context” describes-as the author says himself-an ambitious yet constrained research program. Indeed, it is ambitious yet constrained. It is ambitious for it does not intend less than “to provide a detailed theoretical account of how and why teachers do what they do ‘on line”‘. It is also constrained but not only in the way the author considers it. From my point of view some of these constraints of the “theory of teaching-in-context” are connected with the question of how one deals with complexity and relevance in educational research-an old question which never seems to lose its relevance. Schoenfeld himself indicates in his article that the complexity of educational subjects is a challenge in research he takes up and that by doing this he wants to obtain relevant results for educational theory and practice. Therefore I titled my discussion of his approach “On the Complexity and Relevance in Educational Research”. First, I will discuss Schoenfeld’s approach and then ask to what extent approaches of the kind of Schoenfeld’s can fulfill the challenge of complexity in educational research and obtain relevant results. Now let us have a closer look at Schoenfeld’s approach in three respects: (1) the structure of Schoenfeld’s model, (2) the method of application of this model, (3) the “value added” of this approach-using Schoenfeld’s own criteria to evaluate his approach. Finally we consider this approach as an example to solve the problems of complexity and relevance in educational research and discuss: (4) the scope of approaches of the kind of Schoenfeld’s Theory of Teaching-in-Context.
Direct all correapendence lo: Franz Schott, Institute for Educational and Developmental Psychology, Faculty of Mathematics and Natural Sciences at the Technical University Dresden - Germany, Mommsenstr. 13, D-01 062 Dresden!Ge;enany. Isw~ in Education, Volume 4, Number I, 1996, pages 133-139 All rights of reproduction in any form reserved.
Copyright o 1996 by JAI Press Inc. ISSN: 1060-9724
134
ISSUESINEDUCATION
Vol.4, No.1,1998
ON THE STRUCTURE OF SCHOENFELD’S MODEL In the first section the structure of Schoenfeld’s model is discussed: the three components, the mechanism, the complexity and the novelty of the model. In Schoenfeld’s model three components are used to predict teaching-in-context: “goals,” “beliefs,” and “knowledge, action plans,” each “activated in current context” (cf Fig. 3 of his paper). These three components are related to each other and form the basis for the teacher decision-making. Even though these three components may have a high face validity at the first glimpse, one should ask why especially these three are necessary and sufficient. Let us begin with the goals. Undoubtedly our goals can control our behavior-more or less. In daily life and in science goals, motivation is very basic. For example no murder that is not asked for the motivation of the culprit. In this respect the “goals activated in current context” are of direct face validity as component of the approach. But why do the rest of the components consist just of those two types: “beliefs” and knowledge, action plans”? Why are beliefs conceptually seperated from knowledge? This brings us to the general question of “what is knowledge?“. This question can lead to a very fundamental discussion in epistemology and psychology and cannot be discussed here. At any rate it is useful to take a point of view which abstains from unnecessary preassumptions that can lead to misdirecting prejudices. In such a view of knowledge, individuals have their specific knowledge of the world which enables them to more or less efficiently act in this world. In such a view “knowledge” is understood very broadly and comprises all information of the world an individual has stored in his mind: facts, rules, evaluations, as well as emotions, motivations and volitions associated with objects and events. In daily life and in research we sort this knowledge to understand human behavior. So we sort educational goals and distinguish between cognitive, social-emotional and motor ones as this happens in the taxonomy of educational goals developed by Bloom and his collaborators (cf Bloom, 1956). This distinction corresponds with the old (and actual) point of daily life view of dividing perception and actions into head, heart and hand. In psychology we analyse knowledge as understood in the way described here, from different aspects like factual knowledge, problem solving strategies, motivations, emotions, volitions or whatever. It is, however, one matter to investigate the effect of these aspects on human behavior from each single aspect and to find out more or less important results. Another matter is to fit such results together as components of a model which should predict human behavior. A look into research shows that these aspects are not exactly distinguished in their concepts and that they overlap in some points. But if these aspects become components of a model for predicting human behavior their conceptual base must not overlap. Furthermore, one must have good reasons for their selection. In this case, it is not enough to point out that the single aspects proved important findings in research. Apart from this, it has to be explained why one takes up which components in the model of explanation and prediction of human behavior, But Schoenfeld may have good reasons for choosing his three components. I would have liked to learn more about this.
On Complexiiy and Relevance in Educational Research
135
For making the model work, Alan Schoenfeld adds a mechanism to components. The mechanism remains a bit unclear for me. Compared with the text about components, the text about mechanism is very short. As I took it, the teacher decision in each context is controlled by those representatives of the three components “beliefs,” ”goals,” and “knowledge” which each have highest priority. But how does it happen that especially a certain goal, a certain belief or a certain knowledge have high priority? As it concerns human information processing it would be natural to refer to one of these models of cognitive psychology. In this respect I would have liked to learn from Schoenfeld if he does this or if he sees it as desirable. Another possible alternative would be to design one’s own approach independent of and perhaps better than existing approaches. But even in this case I would like to have a more detailed explanation of the process of this mechanism. For me, an empirical examination of the existence of this mechanism must address the following question: when an observer attributes a topical teacher decision to a certain constellation of the components with their respective priorities, how can the interpretation be proved valid? A solution of this problem could be to assess the valitidy of the interpretation by comparing it with the interpretations of further independent observers-here I would be interested in the correspondence (the reliability) of the interpretations made by different independent observers. Another solution could be obtained from alternative independent sources, for example by information from the teacher. Again, here I would be interested in knowing the level of correspondence of different information sources. There is hardly a possibility to falsify the approach of the three components in case of relying on very few observers or a group of observers only. From Schoenfeld’s point of view his model is complex. I can not comprehend this because for me the model seems very parsimonious: it only consists of the three components, beliefs, goals and knowledge, embedded in a classroomteaching context and a mechanism that says that the highest priorities (however they came about) of each of these components are in effect. From my point of view the structure of the framework (of the types) is not complex. But what is very complex is the process of analyzing or interpreting teaching behavior in terms of the framework (the tokens) Schoenfeld describes. “To our knowledge, the descriptions of teaching and their embodiments in detailed models at the level of mechanism given in this paper are the first of their kind,” wrote Schoenfeld. Naturally, to a certain extent the level of novelty of an approach in the development of science is a matter of interpretation. I have the impression that while the model structure is new the kind of the approach, actually to make detailed investigations on complex behavior in general and teaching behavior in particular, has a longer tradition. Regarding this I have in mind the work of Flanders (1970), Streufert & Streufert (1978) and several approaches in Germany (e. g. Hofer, 1981; Wagner 1977,198l; Wahl, 1981). In addition, aspects of the analysis of behavior in social settings has been discussed by many scientists, in English for example by Bakeman & Gottman (1997), Bakemann & Quera (1995) and Rogosa & Ghandour (1991), in German for example by Wahl(1979).
136
ISSIJESINEDUCATION
Vol. 4,No.1,1998
ONTHEMETHODOFAPPLlCATlONSCHOENFELDUSESHISMODEL In this section the application of Schoenfeld’s model by the observers of teaching behavior is discussed. The model whose structure we have just discussed serves as an interpretive framework for a sequence of teaching behavior. Here, the instruction to be analysed is divided into units which seem to be natural for the interpreter at different levels of parsing and the teaching behavior is reconstructed by interpretation using the above discussed structure of the model . This procedure reminds me of the method of a common direction of educational science in Germany, the so-called “geisteswissenschaftliche Padagogik,” a term that is difficult to translate, but equally means arts-oriented or humanities-oriented pedagogy. Now some comments with regard to this. Impressed by the successes of natural sciences in the last century, disciplines such as theology, history, literary studies were challenged to demonstrate their scientific quality. It followed the distinction between Geisteswissenschaften (which can only be translated insufficiently with “arts” or “humanities”) and natural sciences by some scholars like Dilthey (1894). The subject of geisteswissenschaftlicher Padagogik initially (in theology) is the interpretation of a text. An old text can often not be understood right away. The context of its origin and its time has to be considered. Hypotheses are made about the meaning of single text parts and their fitting together is checked. This is repeated several times what results in a refinement of interpretation. This procedure is also called “hermeneutischer Zirkel” (hermeneutic circle). Furthermore, it was said: “Die Natur erklaren wir, das Seelenleben verstehen wir” (Nature we explain, inner life we understand) (Dilthey, 1894, page 144) A quality criterion for a good approach in natural science is the general validity of laws; a quality criterion for a good geisteswissenschaftlicher approach is the comprehension of interpretation on given premises. Thus, a Marxist historian can offer a different interpretation of the October Revolution in Russia than a historian with a different ideological approach. If one takes these different initial positions the interpretation has to be comprehensible and coherent in itself. Educational scientists transfer this method of interpretation of texts also to the interpretation in instruction and education what they call “Erziehungswirklichkeit” (instruction reality). For me, this geisteswissenschaftliche approach seems to be applied in Schoenfeld’s application of his model. An advantage as well as a danger (of misinterpretation) lies in the fact that the whole expertise of the observer is included; a disadvantage is that this is an ex-post-facto explanation. Is the interpretation comprehensible? Can it be reliable at all or can differently oriented interpreters reach differently coherent interpretations? Is this nothing but an ex-post-facto-explanation? The author indicates that in thought experiments the tape was stopped on which the teacher behavior was recorded and it was tried to directly predict what would happen. Here, I would like to have more information. Schoenfeld’s approach also reminds me of the ‘Handlungsforschung’ (action research) in educational research in the Germany of the seventies and eighties. With great engagement educational processes were interpreted by the researchers making a dive into “reality” and using participant-observer methodology. They
On Complexity and Relevance in Educational Research
137
assured to have individually profited a lot from this research. But what remained for the scientific community? Not much in Germany. I may be mistaken in this regard, but Schoenfeld should be asked how he guarantees that communicable reliable results arise from his investigations which support his theory.
ON THE ‘VALUE ADDED’ OF SCHOENFELD’S APPROACH Schoenfeld asks for the ‘value-added of his approach and establishes three criteria for estimation: Explantory power, predictive power and scope. Let us begin with the latter: here, Schoenfeld is completely right. As the building blocks of his model are totally independent of each kind of teaching and as they can be used for any prediction of human decisions and actions due to their composition they have a broad scope. I have some difficulties with the “explanatory power” criterion. I have already explained this in connection with the problem of interpretation as it refers to the question of reliability and validity of the observers’ interpretations. I would also like to learn more of how Schoenfeld tests the predictive power of his model systematically and what are the results of these tests.
ON THE SCOPE OF APPROACHES OF THE KIND OF SCHOENFELD’S THEORY OF TEACHING-IN-CONTEXT Finally I turn to the extent that an approach of the kind represented by Schoenfeld’s Theory of Teaching-in-Context can handle the complexity of the research object “teaching” and provide relevant results. Schoenfeld says about complexity: “One must begin with a necessary caveat: Human behavior, especially in complex social settings, is not predictable in the sense of predictions made in the physical sciences. However, it is possible to make predictions of a certain type using the kind of models developed here, and to judge the model by the accuracy of those predictions.” I cannot agree with this point of view. In my opinion a fundamential issue is raised. I used to tell a fairy-tale (Schott, 1980) to make this clear: The King does not like it that leaves lie around in his castle grounds in the fall. Thus, all the leaves are collected. But there is one leaf left on a tree. “If I knew where this leaf will fall to the ground I could put a vase at that place and the leaf would disappear,” the King said to himself. For this reason he had a great research project started on which the best natural scientists of the kingdom worked together, especially physicists, metereologists and specialists in the field of aerodynamics. The scientists obtained quite empirically verified results such as ‘a leaf does not fall far from the tree’ or, more exactly: ’ The probability of the leaf’s landing on the ground gets lower according to a certain mathematical function with a greater distance from the tree’. But as you can imagine it was not possible to predict exactly where the leaf would fall. The court jester wept bitterly. He did not get research funds although he had constructed a leaf-collecting-machine which collected the leaf at the desired moment and took it to the desired place. - The moral of this
138
ISSUESINEDUCATION
VoL4,No.1,1998
story is that it roughly describes the difference between descriptive and prescriptive research. Two points should be discussed here: First, one cannot compare physical laws which are formulated on ideal conditions with the analysis of complex social conditions which were not idealized like Schoenfeld does. Even the prediction of nature in the present non-idealized condition is difficult, think for example of the problem with the leaf. Second, instead of examining the fall of the leaf or teaching behavior in descriptive ways, a prescriptive approach can lead to the design of a leaf-collecting-machine or most efficient teaching methods. In this sense Schoenfeld’s method reminds me of the discussion of Stolurow (1965): ‘Model the master teacher or master the teaching model’. I agree with the latter. If Schoenfeld had succeeded in developing a model for the prediction of teaching behavior in context in spite of the doubts I expressed in sections (1) to (3), his approach would be a great success in psychology as it would have been provided a model to predict human behavior in complex social settings. But, would educational science profit from it for improving instruction? In contrast to Schoenfeld I think it would not as it can only describe present instruction. The futile search for the qualities of a good teacher in educational research has shown that it should be a relatively futile attempt to find them by means of Schoenfeld’s model. In my opinion, a prescriptive research approach would be more effective for progress in instruction. Furthermore, one should not restrict teaching behavior to include only those elements described in the Schoenfeld model; rather one should consider instruction as a whole including the students and the entire setting in which the instruction takes place as it is tried to be represented by the Universal Constructive Instructional Theory (UCIT, Schott & Driscoll, 1997). Whichever view is preferred, an interesting question in connection with complexity and relevance in educational research has been raised. I am curious about the view of the other commentators and about Schoenfeld’s reaction. Whatever the final result will be-Schoenfeld’s extensive work which is presented in a very engaging way, initiated an intense discussion of this question.
REFERENCES Bakeman,R., & Gottman,J. M. (1997). Observing interaction. An introduction to sequential analysis. Second Edition, Cambridge University Press. Bakemann, R., & Quera, V. (1995). Analyzing interaction: Sequential analysis with SDIS and GSEQ. Cambridge University Press. Bloom, B. S. (Ed.). (1956). Taxonomy of educational objectives: The classification of educational goals. Handbook I. Cognitive Domain. New York David McKay. Dilthey, W. (1894). Die geistige Welt. Gesammelte Schriften. 8. Auflage (1990). Band V, S. 144. Goettingen: Vandenhoeck & Ruprecht. (Mental world. Collected works. 8th edition. Volume V, p. 144). Flanders, N. A. (1970). Analyzing teaching behavior. Reading, MA: Addison-Wesley. Hofer, M. (Ed.). (1981). lnformationsverarbeitung und Entscheidungsverhalten von Lehrern. Beitriige zu einer Handlungstheorie des Unterrichtens (Information processing and decision
On Complexity and Relevance in Educational Research
139
behavior of teachers. Contributions to an action theory of instruction). Miinchen: Urban & Schwarzenberg. Rogosa, D., & Ghandour, G. (1991). Statistical models for behavioral observations. Journal of Educational Statistics, 16 (3): 157 - 252. Schott, F. (1980). Zur Anwendung verhaltensmodifikatorischer Prinzipien - diskutiert am Beispiel “Anwendungsfeld Unterricht”. In: Belschner, W., Dross, M., Hoffmann, M. & Schott, F.: Verhaltenstherapie in Erziehung und Unterricht. Band 2: Anwendung. 169 - 210 (On the application of principles of behavior modification - discussed by way of the example “area of application: instruction” in: Behavior therapy in education and instruction, volume II: application, 169 - 210) Schott, F., & Driscoll, M. (1997). On the architectonics of instructional theory. In S. Dijkstra, S. Schott, F.N. Seel, & Tennyson, R. (Eds.), instructional design-lnternational perspectives. Hillsdale, NJ: Erlbaum. Volume I: Theory and Research, Part A: Theoretical Foundations of Instructional Design, Chapter 7,135 - 173. Stolurow, L. M. (1965). Model the master teacher or master the teaching model. In J. D. Krumboltz (Ed.), Learning and the educational process, (pp. 223247). Chicago: Rand McNally & Company. Streufert & Streufert, (1978). Behavior in the complex environment. Washington: Winston. New York: Wiley. Wagner, A. C. (1977). Die Analyse von Unterrichtsstrutegien mit der Methode des “Nachtraglichen Lauten Denkens” von Lehrern und Schiilern zu ihrem unterrichtlichen Handeln. (Analysis of instruction strategies with the method of “Later Loud Thinking“ of teachers and students about their actions in instruction). Unterrichtswissenschaft (5), 244ff. Wagner, A. C. , Maier, S., Uttendorfer-Marek, I., & Weidle, R. H. (1981). Unterrichstspsychogramme. Was in den Kiipfen von Lehrern und Schiilern vorgeht. (Instructional psychograms. What is going on in the minds of teachers and students. Reinbek: Rowohlt. Wahl, D. (1979). Methodische Probleme bei der Erfassung handlungsleitender und handlungsrechtfertigender subjektiver psychologischer Theorien von Lehrern. (Methodical problems in the collection of action-leading and action-justifying subjective psychological theories of teachers.) Zeitschrift fur Entwicklungspsychologie und Padagogische Psychologie, 11,208 - 217. . (1981). Methoden zur Erfassung handlungssteuernder Kognitionen von Lehrern. (Methods for the collection of action-controlling cognitions of teachers.) In Hofer, M. (Ed.), Informationsverarbeitung und Entscheidungsverhalten von Lehrern. (Information processing and decision behavior of teachers.) 49 - 77. Miinchen: Urban & Schwarzenberg.