l ill; will l l! lull Ill lull l This is to certify that the thesis entitled A STUDY OF THE PROBLEM-SOLVING BEHAVIOR OF TEACHERS AS THEY DIAGNOSE A CHILD"S READING PERFORMANCE AND THE EFFECTS OF EXPERIENCE AND TRAINING ON THAT BEHAVIOR presented by Elaine C. Amon Stephens has been accepted towards fulfillment of the requirements for Ph.D. de 6 Elementary and gr 6 in SpeciaT EcTucation / / Major professor Date February 24, 1978 0-7639 A STUDY OF THE PROBLEM-SOLVING BEHAVIOR OF TEACHERS AS THEY DIAGNOSE A CHILD'S READING PERFORMANCE AND THE EFFECTS OF EXPERIENCE AND TRAINING ON THAT BEHAVIOR By Elaine C. Amon Stephens A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Elementary and Special Education 1978 ABSTRACT A STUDY OF THE PROBLEM-SOLVING BEHAVIOR OF TEACHERS AS THEY DIAGNOSE A CHILD'S READING PERFORMANCE AND THE EFFECTS OF EXPERIENCE AND TRAINING ON THAT BEHAVIOR By Elaine C. Amon Stephens It appears to be an established principle of the reading profession that classroom teachers should be diagnostic teachers of reading. Not so well established, however, are principles to help answer these two major concerns: How are reading problems diag- nosed? and How do teachers learn to diagnose reading problems? what teachers actually do when they diagnose is largely speculative. Specifically, how a teacher uses her problem-solving skills to diagnose a child's reading performance remains unanswered. There- fore, this study attempted to (1) identify and describe the problem-solving behavior of teachers as they diagnosed a child's reading problem, and (2) determine the effects of classroom teach- ing experience and graduate level instruction in reading diagnosis and correction on that behavior. A review of the literature indicated the lack of a suffi- cient model of problem-solving behavior for reading diagnosis based on empirical evidence. Due to the insufficiency of present models, this study, as well as the larger research project of which Elaine C. Amon Stephens it is a part, used a model developed in another discipline which attempts to describe diagnostic problem-solving behavior and which may have application to reading diagnosis. Designated the Inquiry Theory of Clinical Problem Solving, it describes diagnostic problem- solving as a complex form of reasoning which is probablistically determined by the interplay of the problem-solver's previously acquired cognitive capabilities for the problem and specific proper- ties of the problem itself. It suggests that the problem-solver's performance is a function of his store of knowledge about and search strategies for a particular problem. If these can be aided or improved, then diagnostic performance should improve. Previous research in other disciplines seems to indicate that experience and training in diagnostic problem solving improve knowledge and stra- tegy thus resulting in improved performance. This study attempted to apply the Inquiry Theory of Clinical Problem Solving to teachers and their diagnosis of a child's reading problem by hypothesizing that graduate level instruction in reading diagnosis and correction and/or classroom teaching experience would improve teachers' knowledge about reading problems and their search strategies for collecting and processing information, thus resulting in improved diagnostic performance. To this end, thirty pre-service and classroom teachers were assigned to three groups each containing ten subjects. The criteria for selection and assignment to groups were classroom teaching experience and graduate level instruction in the diagnosis and correction of reading difficulties. Under the constraints of procedures which were designed to elicit problem- Elaine C. Amon Stephens solving behavior, each teacher interacted with a set of materials which simulated the behavior of a child with a problem in reading. Written protocols and tape recordings of the teacher's ‘behavior during the interaction as well as statements of the final diagnosis and remediation plan were analyzed to obtain data for designated measures of problem-solving behavior. Both product and process measures of diagnostic performance were used. Product mea- sures were used to measure the final stated diagnosis of the reading problem. Process measures were used to describe the manner in which the problem was diagnosed. One-way multivariate analysis of variance and Pearson Product-Moment Correlation were the statistical proce- dures employed. The major finding of this study was that there was no sig- nigificant mean score difference on the principal product measure of diagnostic performance between teachers with teaching experience and graduate instruction in reading diagnosis and correction, teachers with teaching experience and no graduate instruction in diagnosis, and pre-service teachers without teaching experience or graduate instruction. Given the limitations of the present study, it appears that neither graduate level instruction in reading diagnosis and cor- rection and/or classroom teaching experience seemed to significantly affect the knowledge and strategies employed by teachers in diagnos- ing a child's reading problems under simulated conditions and, therefore, did not result in improved diagnostic performance. To you who have dreamed with me . Hold fast to dreams, for if dreams die, life is a broken winged bird that cannot fly. --Langston Hughes ii ACKNOWLEDGMENTS During my life, I have been privileged to have two Teachers, in the finest sense of that word, who have contributed immeasurably to my desire to know and learn. Although they differ widely in style and manner, Irene Sanderson, my second grade teacher, and George Sherman, Chairman of my doctoral committee, share an excite- ment about learning and a love of teaching which has had a tremen- dous impact upon me. To Irene and George, I express my love and gratitude. My love and gratitude can never be expressed enough to my husband and daughter--to Wes, whose quiet strength has provided the stability for our lives, and to Melinda, whose first comment upon hearing that I had completed my degree was, "Nell, Mom, what's next?" My parents, Wayne and Opal Amon, have always provided the drive and encouragement for the "what's next" in my life. My love and appreciation go to them, also. Additionally, I wish to express my appreciation to my committee members, Dr. John Vinsonhaler, Dr. James Snoddy, Dr. Charles Blackman, and Dr. Louis Romano for their assistance and encouragement. A special bouquet of thanks goes to John Vinsonhaler for his valuable contribution to this study. Additionally, without the help of the CLIPIR staff, Cynthia Zaldokas, and the teachers who participated in the project, this study would never have been completed. I thank each of them for so willingly giving of their time and energy. And lastly, I thank the special friends who listened, sup- ported, encouraged and cheered. iv TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES LIST OF APPENDICES Chapter I. THE PROBLEM . Introduction . . Inquiry Theory of Clinical Problem Solving. The Problem . . . . . . . Research Tasks . . Significance of the Study Definition of Terms . Rationale for the Study Assumptions and Limitations . Overview of the Study . II. REVIEW OF THE LITERATURE Introduction . . . Diagnostic Problem Solving in Reading . Inquiry Theory of Clinical Problem Solving. Factors Relating to Teacher Knowledge of the Teaching of Reading. . . Pragmatic Approaches to Reading Diagnosis Summary . . . . . . . III. PRESENTATION OF THE DESIGN AND METHODOLOGY Introduction . . Design of the Study Sample of the Study The Teacher Population Course Content . Materials and Procedures . Treatment of Data Summary . Page vii viii ix Chapter Page IV. RESULTS . . . . . . . . . . . . . . . . 53 Introduction . . . . . . . . . . . . . . 53 Major Finding . . . . . . . . . . . . . 55 Analysis of the Data . . . . . . . . . . . 56 Summary . . . . . . . . . . . . . . 66 V. SUMMARY, CONCLUSIONS, DISCUSSION AND CONJECTURES, RECOMMENDATIONS AND POSTLUDE . . . . . . 68 Introduction . . . . . . . . . . . . . . 68 Summary . . . . . . . . . . . . . . . 69 Conclusions . . . . . . . . . . . 73 Discussion and Conjectures . . . . . . . . . 76 Recommendations . . . . 87 Postlude: Observations on Teacher Problem-Solving Behavior in Reading Diagnosis . . . . . . . 89 APPENDICES . . . . . . . . . . . . . . . . . . 98 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . ll4 vi Table hhh-fihhww 0101-th wN d LIST OF TABLES Teachers' Experience and Training Time on Task Design Matrix Means and Standard Deviations: Diagnostic Score MANOVA: Product Measures Means and Standard Deviations: Product Measures MANOVA: Process Measures Means and Standard Deviations: Process Measures Measures of Problem-Solving Behavior: Correlation Coefficients and Levels of Significance . Correlation Coefficients and Levels of Significance Between Grade Point Average of Group 3 PSET) and Dependent Variables . . . . . vii Page 37 44 50 56 6O 60 62 62 64 66 LIST OF FIGURES Figure Page 4.l Distribution of Diagnostic Scores by Group . . . . 57 4.2 Distribution of Diagnostic Scores by Teacher . . . . 57 viii LIST OF APPENDICES Appendix A. B C. D Directions for Observational Session Criterion Diagnosis . Product Formulas . Process Formulas . ix Page 99 106 110 ll2 CHAPTER I THE PROBLEM Introduction Parents send their children to school with the expectation that they will learn to read. Yet statistics released from the U.S. Office of Education as recently as 1974 (Education Briefing Paper, 1974) indicated that: 1. Over 8 million school-age children are not learning to read adequately. Sixteen percent of the enrollment in grades I through l2 require special instruction in reading. In most large city school systems . . . at least half of the students are unable to read well enough to handle their assignments. Each year some 700,000 youngsters drop out of public school. Studies show that the average dropout is at least two years behind his age group in reading and other basic skills. There are more than 3 million illiterates in our adult population. About 8 l/2 million Americans lack the practical read- ing skills necessary to complete simplified application forms for such common needs as a driver‘s license, a personal bank loan . . . . Recent results (Venezky, l977) from a nationwide survey of the reading abilities of 9-, 13-, and l7-year-olds conducted by the National Assessment of Educational Progress seem to indicate that children at age 9 are reading better than their counterparts of a few years ago. However, a close examination of the results reveals that no group improved significantly in any comprehension skill and that children of the economically depressed inner urban areas showed no significant improvement in any skill at any of the three age levels. While it appears that the reading ability of 9-year-old Blacks has improved, nevertheless, it remains 10 or more percentage points below the national average. While we do succeed in helping the majority of children learn to read, the fact remains that many children and adults simply do not achieve a level of literacy which enables them to function successfully and comfortably in our complex society. Probably no other aspect of schooling has provoked such intense national atten- tion and been so extensively researched as reading. Researchers have examined the effects of environmental fac- tors (Robinson, 1946; Deutsch, 1967; Cohen, 1970), physiological factors (Robinson, 1946; Bond and Tinker, 1967), psychological factors (Robinson, 1946; Harris, 1970) and methodological factors (Flesch, 1955; Chall, 1967) on reading achievement. While all of these factors can influence achievement, emerging from this research is a reoccurring theme which suggests that the key variable in reading achievement is the teacher (Ramsey, 1962; Bond and Dykstra, 1967; Harris and Morrison, 1969). While research has not yet provided a complete description of teacher characteristics and behavior which affect reading achievement, many reading authorities endorse the concept of the "diagnostic teacher" (Strang, 1964; Dechant, 1968; Durkin, 1970; Harris and Smith, 1972; Guszak, 1972; Ruddell, 1974; Harris, 1974; Ekwall, 1976). Beginning with Strang's book, Diagnostic Teachinggof Read- jng_(1964), reading diagnosis is described as an integral part of every teacher's job. The term "diagnostic teaching," according to Harris and Smith (1972), is now widely accepted in the reading pro- fession and refers to "broad and continuous assessment of student needs through formal and informal means to the end of differenti- ating reading instruction." Various studies (Feldman, 1974; Rupley, 1975: Schultz, 1975) support this concept by suggesting that reading achievement is fostered by the teacher who has a knowledge of the reading process, can determine the needs of students, and can imple- ment a program based on those needs. Reading diagnosis, however, often carries an aura of mystery about it (Wilson, 1967). While it is no longer unusual for teacher education institutions to offer undergraduate students some form of training in diagnosis and correction (Morrison and Austin, 1977), it is still often viewed as the domain of the advanced level gradu- ate student (Jan-Tausch, 1971). The extent to which classroom teachers are actually diagnostic teachers of reading has not been demonstrated, but it has been observed that the teachers often relegate the responsibility for diagnosis to a specialist outside of the classroom. Yet, as Harris (1974) states, ". . . when it comes to the analysis of children's reading performance so as to determine the skills he needs and how he may best be taught, the . . . teacher should be the diagnostician." It appears to be an established principle of the reading profession that teachers should be diagnostic teachers of reading. Not so well established, however, are principles to help answer these two major concerns: How are reading problems diagnosed? and How do teachers learn to diagnose reading problems? While the field abounds with textbooks designed to teach the "how-to“ of diagnosis, a review of the literature reveals a lack of empirical studies designed to foster an understanding of how teachers collect information on a given reading problem and use that information to reach a diagnosis. What teachers actually gg_when they diagnose is largely speculative. Specifically, how a teacher uses problem- solving skills to diagnose a child's reading performance remains unanswered. Therefore, current approaches to the teaching of diagnosis rest, at best, on unproven ground. Inquiry Theory_of Clinical Problem Solving The diagnosis ofa reading problem and subsequent prescrip- tion of appropriate instruction as currently practiced appears to be a complex process. Diagnostic behavior seems to include both cognitive problem-solving skills and affective interpersonal skills. Presently, there is an attempt by some researchers (Shul- man and Elstein, 1975: Vinsonhaler et al., 1976) to apply the Inquiry Theory of Clinical Problem Solving to the diagnosis of reading problems. A model of problem solving would not, of course, account for all that occurs when a teacher diagnoses a reading problem and prescribes instruction. Still, it appears useful for developing an understanding of certain crucial aspects of diagno- sis, namely, the complex reasoning processes involved. Elements of the Clinical Problem-Solving Model A model of clinical problem solving contains certain basic elements. The behavioral domain of the model is the diagnostic encounter. It is a set of events which occur during an interaction between a problem solver or clinician and another individual. The purpose of this interaction is to solve the individual's problems. It is characterized by the identification of the problem and its causes followed by the selection of treatment suitable for the problem. Thus, the study of the diagnostic encounter for reading diagnosis encompasses the cognitive problem-solving behavior which occurs when a teacher is presented with the problem of determining a child's reading difficulties and planning instruction. It involves two major decisions: What are the problems? and What can be done about them? The conceptual framework for studying a diagnostic encounter is based on two principles. The first principle is that the cre- ation and use of a simulated problem case ensures a necessary degree of scientific objectivity and control while enabling realis- tic, valid problem-solving behavior to occur. The work of Turner and Fattu (1960) conceptualized teaching as a problem-solving activity for which problem situations or teaching tasks in reading and arithmetic among other areas could be simulated. Simulations have also been used effectively in medical research (Elstein et al., 1976) to study the problem solving of physicians during diag- nosis and in psychology (McDermott, 1975) to study the diagnostic decision making of school psychologists. The second principle is that what occurs during the diag- nostic encounter is based on an interaction between specific proper- ties of the problem itself and the problem solver's previously acquired cognitive capabilities for that problem. That is, the problem-solving behavior which the clinician demonstrates with a particular problem is probablistically determined by the interplay between his search strategies for gathering and processing informa- tion and his store of knowledge about the problem, as well as by the salient features of the problem itself. The conceptual framework for clinical problem solving also includes two important corollaries about evaluation and instruction which have implications for educators who are concerned with help- ing teachers improve their diagnostic problem-solving skills. The first has to do with evaluation of the clinician's diagnostic behavior. It has already been stated that what occurs during the diagnostic encounter is probablistically determined by the inter- action of the problem with the clinician's knowledge and search strategies. Therefore, the score obtained by the clinician on valid and reliable measurements should be an indication of his knowledge and strategies for a given problem. Evaluation of the encounter should include (1) an accurate, effective problem case simulation, (2) criteria for describing the clinician's behavior, and (3) criteria for comparing the clinician's behavior with the behavior of experts on the given problem case. The second or instructional corollary has to do with the influence of experience and training on clinical problem solving. Again it is noted that the clinician's performance is a function of knowledge and search strategies. If these can be aided or improved, then diagnostic performance should improve. It appears that experience and training in clinical problem solving may improve knowledge and strategy. It follows, therefore, that as experience and training increase, performance during diagnostic problem-solving encounters should improve. While there may be several sources of experience and train- ing in clinical problem solving for teachers, traditionally an emphasis has been placed on two major sources. The first is actual teaching experience. The extent and quality of reading diagnosis in the classroom, however, is difficult to determine. It may occur intermittently, it may occur with or without accurate feedback, or, for many teachers, it simply may not occur at all. Still, the potential exists for reading diagnosis to occur within the confines of the classroom. The second source is specific instruction in the diagnosis and correction of reading problems which teachers may elect to take from teacher education institutions. Courses which are designed to provide teachers with experience and feedback in clinical problem solving in reading should also result in improved knowledge and strategy thus resulting in improved diagnostic performance. The Problem The purpose of this study is to investigate the problem- solving behavior of teachers as they diagnose a child's reading problem. If the Inquiry Theory of Clinical Problem Solving is representative of the nature of diagnostic problem solving as per- formed during reading diagnosis, and if training and experience improve this performance, it may be expected that teachers at successive levels of training and experience would differ on mea- sures of diagnostic performance. Specifically, it may be expected that as training and experience increase, diagnostic performance would improve. This study will build upon research currently being con- ducted by the Clinical Information Processing In Reading (CLIPIR) Project of the Institute for Research on Teaching at Michigan State University. One component of the CLIPIR is the identification and description of the problem-solving behavior of expert reading diag- nosticians. The present study, like the larger research project of which it is a part, limits the scope of its investigation to cogni- tive problem-solving processes and the way they are used to arrive at a diagnosis. Research Tasks The main areas of this study are: 1. To identify and describe the problem-solving behavior of teachers as they diagnose a child's reading problem. 2. To determine the effects of classroom teaching experi- ence and graduate level instruction in reading diagnosis and correction on the problem-solving behavior of teachers as they diagnose a child's reading problem. Significance of the Study It appears to be an established principle of the reading profession that teachers should be diagnostic teachers of reading. However, very little is known about how reading problems are diag- nosed and how teachers learn to diagnose. A review of the litera- ture reveals the lack of a sufficient model of problem-solving behavior for reading diagnosis based on empirical evidence. This study is intended to provide data on the diagnostic problem-solving behavior of teachers and how classroom teaching experience and graduate level instruction in diagnosis affects this performance. These data will help researchers as they seek to evaluate the application of the Inquiry Theory of Clinical Problem Solving to reading diagnosis. It will aid educators who are concerned with helping teachers improve their diagnostic problem-solving skills. Lastly, it will add to the growing body of knowledge on clinical problem-solving behavior. 10 Definition of Terms Definitions for the key terms in this study have been estab- lished to provide a common basis for understanding. Reading Diagnosis: This term refers to both (1) an act or process and (2) a decision rendered, or product. (1) The process whereby information is obtained, analyzed, and interpreted relative to a child's reading per- formance in order to determine his strengths and weaknesses for the purpose of prescribing suitable instruction. (2) The product which is a conclusive statement of the decision rendered regarding a child's reading per- formance. Simulated Case: A collection of materials and procedures of moderate fidelity designed to approximate an actual case of reading difficulty so that samples of problem-solving behavior may be obtained under controlled laboratory conditions. Cue: An item of information pertaining to a factor or factors related to a child's reading performance coupled with the procedure for obtaining it. Hypothesis: A statement regarding the condition of a fac- tor or factors related to a child's reading performance which is inferred from cues during the process of diagnosis. Rationale for the Study The rationale for this study is based upon two premises: first, that classroom teachers are teachers of reading and should, therefore, be able to diagnose the reading performance of individual children, and second, that the diagnostic behavior of teachers ll involves, in part, cognitive problem-solving behaviors which can be learned. Assumptions and Limitations The following statements serve as the basic assumptions of this study: 1. 2. A reading problem can be simulated. The particular case of reading difficulty selected for this study was accurately and effectively simulated. Teacher performance on a simulated problem case approximates performance on an actual reading problem. Appropriate measurement tools exist for data col- lection and analysis. The method for selection of subjects was appro- priate. The clinical problem-solving model is representa- tive of the nature of diagnostic problem solving. The findings of any study are limited by the existence of potential defects which can arise from the set of assumptions upon which the study is built. Additional caution must be exercised in the interpretation of the findings of this study in that volunteer subjects were used to diagnose only one simulated reading problem under a specificed set of conditions. The findings are, therefore, generaliZable only to other subjects selected in a similar manner diagnosing a similar reading problem under the same set of simula- ted conditions. 12 Overview of the Study A review of pertinent literature is presented in Chapter II which examines (l) empirical studies describing diagnostic problem solving in reading, (2) selected findings in other areas and from other disciplines relative to the application of the clinical problem-solving model to reading diagnosis, and (3) pragmatic approaches to reading diagnosis as described by the reading pro- fession. The research design and methodology of the study are described in Chapter III. The data which were collected, treated and analyzed for this study are presented in Chapter IV. A summary of the study and appropriate conclusions are pro- vided in Chapter V. Implications and suggestions for future research are also included. CHAPTER II REVIEW OF THE LITERATURE Introduction The purpose of this chapter is to review the existing literature relative to the nature of problem solving in reading diagnosis. The following areas will be examined: (1) empirical studies describing diagnostic problem solving in reading, (2) selected findings in other areas and from other disciplines relative to the application of the clinical problem-solving model to reading diagnosis, and (3) pragmatic approaches to reading diagnosis as described by the reading profession. Diagnostic Problem Solving in Reading Turner and Fattu at the Institute of Educational Research at Indiana University (1960) conceptualized the teaching of reading as a problem-solving activity. Based upon Liberman's description of a professional as someone who places an emphasis on intellectual rather than physical techniques, particularly on defining and col- lecting evidence to resolve problems, Turner and Fattu stated that the teacher's function was "to define and to resolve problems within the domain delimited by their objectives." According to Turner and Fattu, a problem arises whenever there is a discrepancy between the actual behavior of a student and 13 14 the behavior the teacher desires. This problem is then analyzed into parts or tasks. Turner and Fattu considered the diagnostic task to be the epitome of teaching. Diagnostic tasks go beyond mere evaluation to include identification of the particular aspect of the goal which the student has not achieved, examination of his errors en route to the goal and the prescription of remedial procedures. Working with Turner and Fattu, Burnett (1961) designed a set of paper and pencil problems of low fidelity simulation intended to provide a measure of proficiency at problem solving in the teach- ing of reading which would provide some insight into why individuals differ in this proficiency. The problems were based on two children who had been reading clinic referrals and who had reading diffi- culties similar to what any elementary teacher could normally expect to find in the classroom. Modeled upon tests used in medical diagnosis and electronic:troubleshooting, an attempt was made to measure five levels of teacher operation in the use of diagnostic procedures. Seeking to simulate information which would be avail- able to any classroom teacher, test data, background information, and school records were collected. The subjects were required to perform five tasks: 1. Pick critical information from a pool of data. Select a means of securing additional data. Interpret the data. awn Make recommendations for improving instruction. 15 5. Finally, after all of the remaining data were made available, re-evaluate the recommenda- tions. The subjects were required to rank four responses on each task in terms of how well they met the specifications called for in the problem. The choices were so weighted that the total score reflected consistency on all the tasks. Burnett. found that reading specialists significantly out- scored experienced teachers and that experienced teachers signifi- cantly outscored undergraduate students. However, the following variables did not result in significant mean score differences on the problems: teaching experience which was exclusively at the primary or intermediate level, size of school system, number of years of teaching experience beyond the third year, age, or the possession of a master's degree. Burnett concluded his study with the recomendation that additional investigations be conducted to shed more light on the strategies which elementary teachers use in solving reading diffi- culties. He suggested giving diagnostic problems to individual teachers and recording their vocalized thought processes as they attempted to solve the problems. A review of the literature in the 15-year interval since Burnett's study fails to reveal any significant studies: which describe diagnostic problem-solving behavior in reading in this manner. It'appears appropriate, therefore, to examine selected findings in other areas and from other disciplines relative to a 16 model which attempts to describe problem-solving behavior and which may have application to reading diagnosis, namely, the Inquiry Theory of Clinical Problem Solving. Inquiry Theory of Clinical Problem Solving The clinical problem-solving model attempts to describe and account for the cognitive problem-solving behavior which occurs when a clinician diagnoses the problems of a particular individual and prescribes treatment. There are certain elements of this model which are of particular import for the present study. The first is that the problem-solving behavior of the clinician is probablis- tically determined by the interplay between his search strategies for gathering and processing information and his store of knowledge about the problem, as well as by the salient features of the problem itself. The second has to do with the influence of experience and training on clinical problem solving. It is these elements which may be helpful in addressing two major concerns: How are diagnos- tic problems solved? and How does one learn to solve diagnostic problems? How Are Diagnostic Problems Solved? For most kinds of problem solving the crucial elements are the initial representation of the problem and the systematic use of mental operations under the influence of a plan (Posner, 1973). Newell and Simon's theory (1972) of human problem solving includes these two fundamental propositions: 17 l. The task environment or the problem is repre- sented internally as "problem space" by the problem solver. 2. The structure of the problem space determines the information-processing activities which the problem solver uses in his search for a solution. Since the potential size of the problem space is enormous, some way must be found to limit the size of the space to be searched. The open, ill-defined problem of "what is wrong" must be transformed into a set of closed, better-defined problems (Barlett, 1958). Recent research in clinical problem solving in medical diagnosis (Elstein et al., 1976) seems to indicate that a major strategy used by physicians to define this problem space is the early generation of tentative diagnostic hypotheses. This hypothesisaggjded approach to clinical problem solving (Gordon, 1973) represents a departure from the traditional view of diag- nostic problem-solving behavior. Traditionally, diagnosis has been described as a sequential activity in which large amounts of information are thoroughly and systematically gathtered and thgn_ana- lyzed and synthesized into a conclusion. Research findings seem to indicate, however, that in actual practice, physicians behave quite differently. It appears that what they do in a diagnostic encoun- ter is generate a few tentative diagnostic possibilities after initially collecting a very small amount of information or cues. Instead of proceeding further with a routine, systematic collection of infbrmation, the physician is then guided by his hypotheses to 18 gather data which confirms or disconfirms the diagnoses he is con- sidering. The hypothesis-guided approach suggests that diagnostic problems are solved by the iterative processing of the following tasks (Elstein et al., 1976). l. Cue acquisition--the process of gathering and collecting data. 2. Hypothesis generation--the process of generating alternative formulations of the problem. 3. Cue interpretation--the process of interpreting the evidence collected in the light of these hypotheses. 4. Hypothesis evaluation or judgment--the process of combining information to reach a diagnostic decision. The formation of hypotheses appears to direct the clini- cian's search strategies and acts as the organizing principle for retrieving information. The formation of hypotheses is limited by the clinician's knowledge of the problem stored in his long-term memory. The number of hypotheses considered at any one time are limited by the capacity of the clinician's short-term working memory. Thus, it appears that some diagnostic problems may be solved through a process of generating and testing hypotheses. These hypotheses then direct the clinician's use of search strate- gies and retrieval of information about the problem. If a clini- cian's diagnostic behavior is a function of his knowledge and search strategies, what are the implications for learning to solve diagnostic problems? 19 How Does One Learn to Solve Diagnostic Problems? The clinical problem-solving model suggests that if a clini- cian's knowledge and search strategies can be aided or improved, then diagnostic problem-solving behavior should improve. It appears that experience and training in clinical problem solving may improve knowledge and strategy. As noted earlier, Burnett (1961) found that reading spe- cialists significantly outscored experienced teachers and experi- enced teachers significantly outscored undergraduate students on measures of proficiency at problem solving in the teaching of read- ing. Using school psychologists at successive levels of training and experience, McDermott (1975) demonstrated that they could be differ- entiated on the basis of measures of diagnostic decision making. In medical research, clinical performance seems to improve as a function of experience with feedback (Neufeld, 1976; Barrows, 1976). Elstein (1976) noted that diagnostic performance was related to the amount of experience during and after medical school. Additionally, in an extensive review of the literature pertaining to problem solving in chess, logic, and medicine, Elstein stated that the dif- ferences between expert and weaker problem solvers are more to be found in the repertory of their experiences, organized in long-term memory, than in differences in the planning and problem-solving heuristics employed. Thus, it appears that one may learn to solve diagnostic problems through some form of experience and training which provides 20 practice with feedback. One aspect of the present study is an attempt to understand how the experience and training which teachers receive relate to their skills in diagnosing reading problems. It has already been noted that the empirical studies in this area are very limited. There is, however, a body of literature pertaining to teacher knowledge of the teaching of reading and the factors related to that knowledge. A review of this body of literature may provide some insights into the factors which affect the learn- ing of diagnostic problem-solving skills with implications for teacher education. Factors Relating to Teacher Knowledge of the Teaching of Reading Investigations of teacher knowledge of the teaching of read- ing are based largely on the results of teacher performance on paper and pencil test items. Artley and Hardin's Inventory of Teacher Knowledge of Reading (1971), for example, uses 95 multiple- choice items to measure teacher knowledge in these seven areas: reading readiness; word perception; comprehension and critical reading; differentiating reading instruction; silent and oral read- ing; evaluation, diagnosis and correction; and goals of instruction. The factors affecting this knowledge which have received the most study are teaching experience and educational training. VanRoosendaal (1975) conducted an extensive review of the literature pertaining to these and other factors which may con- tribute to teacher knowledge of the teaching of reading. She found 21 that the literature is about evenly divided as to whether or not experience contributes to knowledge of the teaching of reading. Her own study using Artley and Hardin's Inventory of Teacher Knowledge of Reading concluded that experience appears to make a significant contribution to teachers' knowledge of reading. VanRoosendaal's review of the literature found coursework to be a contributing factor to teacher knowledge of the teaching of reading. Her own study, however, found no significant difference in knowledge between teachers having only one reading course and those with two or more reading courses, except in the area of word perception. Studies of the effect of student teaching on teacher knowl- edge, according to VanRoosendaal, are so limited as to preclude drawing any conclusions. In conclusion, while studies related to teacher knowledge of the teaching of reading seem to produce some contradictory results, it appears that training and experience may be significant factors effecting knowledge. Ekwall (1973), in a summary of review of research on the effectiveness of teacher-training programs, states: Although studies on the effectiveness of teacher-training programs do produce results that tend to be contradictory, the important point seems to be that certain types of teacher training are highly successful while other teacher- training programs appear to be of little or no value. Our research in improving the training of the reading teacher should no longer concern itself with whether or not train- ing is effective, but rather with what type of training is most effective with teachers with various degrees of training and experience. 22 In applying the findings from these studies to the effects of experience and training on diagnostic problem-solving behavior in reading, two cautions must be observed. First, it must be remembered that these studies deal with acquisition of knowledge about the teaching of reading rather than application of this knowl- edge particularly in a diagnostic context. In support of this dis- tinction, Hammond and Summers (1972) state: Although learning theorists have long emphasized the dis- tinction between learning and performance, little atten- tion has been given to skill in the application of knowledge in tasks which do not involve motor performance. Rather, there is an implicit assumption that once knowl- edge has been acquired, the application of this knowledge is largely dependent on certain experimental circum- stances . . . . The position taken here, however, is that acquisition and application are independent components of learning in cognitive tasks as well as psychomotor tasks. Second, it is important to note that the kind of knowledge measured in these studies of teacher knowledge of the teaching of reading may or may not be the same kind of knowledge required for clinical problem-solving behavior. Clinical problem-solving behav- ior seems to require a knowledge of specific problems, the cues associated with these problems, and appropriate instructional prac- tices. The extent of teacher knowledge in these areas has not yet been amply demonstrated. In summary, these points emerge: first, empirical studies describing diagnostic problem-solving behavior in reading are very limited; second, while the Inquiry Theory of Clinical Problem Solv- ing developed in another discipline provides a description of diagnostic behavior, it remains for research to demonstrate its 23 applicability to reading diagnosis; and third, studies pertaining to teacher knowledge of the teaching of reading may or may not be related to diagnostic problem-solving behavior. There remains, however, another area to investigate. That area concerns the tra- ditions which exist within the reading profession of what consti- tutes "good" diagnostic behavior. Pragmatic Approaches to Reading Diagnosis In an attempt to understand the nature of diagnostic problem-solving behavior in reading, it appears appropriate to examine diagnosis as described by the reading profession. A description of the pragmatic approaches to diagnosis which are held to be good diagnostic practices and are taught to students can be derived from a review of several well-known authorities in the field. Such approaches appear to be based on practical experience or rational thought rather than extensive empirical evidence. Traditionally, instruction in reading diagnosis has gener- ally involved specifying a set of procedures to follow, a series of tests to administer, and guidelines for interpretation and recom- mendations. Developed by Helen Robinson and Helen Smith (1968), pion- eers in the field of reading diagnosis, this plan for diagnosing readers involves the following steps: 1. Secure a detailed case history. 2. Estimate as accurately as possible the level at which the person might be expected to read. 24 3. Determine the current level of achievement in reading and related areas. 4. Evaluate the information obtained in the three pre- ceding steps to decide if the subject is a retarded reader. If individuals are reading at their expected levels, they are considered not to be retarded read- ers. For these individuals appropriate recommenda- tions may be made at this time. For those who are retarded in reading, the next steps are followed. 5. Make a detailed analysis of the problem in reading. Consideration is here given to the four major areas of reading: word recognition, vocabulary, compre- hension, and rate. 6. Identify inhibiting factors. The chief concern is to identify all malfunctions that may currently be inter- fering with the person's learning to read. 7. Collate all the data secured during the preceding steps and interpret them accurately. 8. Make appropriate recommendations for all aspects of remedial therapy. This approach emphasizes the thorough and systematic gath- ering of a relatively large amount of data which are then interpre- ted and evaluated in order to reach a diagnosis. Bond and Tinker, ReadingDifficulties, Their Diagnosis and Correction (1973), expand and modify the traditional approach by describing three levels of diagnosis: general, analytical, and case-study. They feel that many children's reading problems need only a general diagnosis while others may require an analytical or case-study approach. Therefore, they stress diagnosing only so far as is necessary to prescribe treatment. General diagnosis is used to identify children who are doing relatively poor work in reading compared with their work in 25 other areas or with measures of their mental ability. Information is obtained on the children's ages, general intelligence, and achievement in other curricular areas from group tests and cumula- tive records. Analytical diagnosis is designed to systematically identify a child's strengths and weaknesses in reading. It specifically locates problem areas such as limited word recognition or general comprehension difficulty. Case-study diagnosis provides an in-depth appraisal of a child. It includes not only the child's reading skills and abili- ties, but his mental, physical, and sensory characteristics, his attitude towards reading, and his environment. The emphasis is upon the collection and detailed study of all the requisite informa- tion before prescribing treatment. Bond and linker indicate that information must be sought in these areas: 1. It must be determined whether the child is correctly classified as a disabled reader or if some other prob- lem of growth and development is the basic difficulty. 2. The specific faulty learning which is impeding reading progress must be discovered in order to specify the nature of the training needed. 3. It must be determined where remediation can most effectively be provided based upon the nature of severity of the problem. 4. Based upon the nature of the problem, the most effi- cient methods for improving the child's reading must be determined including level and types of material, length and frequency of remedial lessons, indepen- dent activities, and means for indicating progress to the child. 26 5. Any condition within the child which might detrimen- tally influence his reading growth must be located. 6. The entire learning environment of the child must be appraised in order to locate any conditions which might interfere with his progress in reading. The person conducting the diagnosis works in this manner: He usually starts by giving survey and achievement tests, individual mental tests, personality appraisals, and may continue until he has measured such details as how many independent letters are unknown to the child or which of the important digraphs he does not know. A study of the possible limitations should go as far as and no farther than is necessary to formulate the nature of the reading instruction needed. Ekwall, Diggnosis and Remediation of the Disabled Reader (1977) and Teacher's Handbook on Diagnosis and Remediation in Read- jng_(l977), represents another modification of the traditional approach. Using a scope and sequence of reading skills as the framework for diagnosis and remediation, the teacher must know not only EDEE to diagnose but also understand nnen_to expect each stu- dent to have mastered each of the reading skills. Ekwall stresses these operational principles: 1. When deciding upon the amount of diagnosis to conduct before starting remediation, choose somewhere between the position that it is better to do a great deal of diagnosis before remediation is begun and the opposing position that it is better to do only enough diagnosis to initiate remediation and then continue diagnosis while teaching. 2. Consider each individual in terms of the type of prob- lem he obviously exhibits when deciding upon tests to administer. Do not fall into the trap of administering the same battery of tests to every student regardless of his or her reading level or apparent reading prob- lems. Ask yourself whether giving any particular test is likely to change the course of the student's remedi- ation. 27 3. When examining any diagnostic test, keep in mind this important question: Does the student have to perform in a situation similar to what he would have to do when actually reading? A major problem with most group diagnostic tests and with many individual tests is that they do not really measure what they purport to measure. 4. There is little value in diagnosing factors for which we either do not expect to provide remediation or for which remediation has not proven effective in the past. 5. Diagnosis for a seriously disabled reader should involve more than an appraisal of educational factors. Each of these authorities writes about the nature and causes of reading difficulties. Robinson and Smith provide descriptions of case studies. Additionally, they all write extensively about test instruments to be used in diagnosis. Bond and Tinker empha- size standardized instruments describing their characteristics and what information can be gained from them. Ekwall's approach is to first describe the reading skill to be diagnosed and then to suggest various instruments, especially informal devices, for diagnosing that skill. Appropriate instruction for various reading problems is also dealt with in varying degrees of detail by these experts. In conclusion, several approaches to diagnosis have been reviewed in an attempt to understand the nature of reading diagnosis as described by authorities within the reading profession itself. Each of these authorities places a relatively heavy emphasis on the negn§_by which information is collected on a child's reading prob- lem. Although varying somewhat in details of the process, still the major focus appears to be on what instruments to use, how and when to use them, and what information they will provide. How to 28 analyze and synthesize this information so as to arrive at a diag- nosis is left largely unspecified. The hypothesis-guided approach to clinical problem solving being examined in this study uses an iterative process of cue acqui- sition, hypothesis genration, cue interpretation, and hypothesis evaluation in solving diagnostic problems. Reviewing the selected approaches to reading diagnosis in the light of these tasks, it appears that cue acquisition, the process of gathering and collect- ing data, is thoroughly treated. This does not appear to be so for the remaining three tasks: hypothesis generation, the process of generating alternative formulations of the problem; cue interpreta- tion, the process of interpreting the evidence collected in the light of these hypotheses; and hypothesis evaluation, the process of combining information to reach a diagnostic decision. Addi- tionally, no other set of tasks or processes based on an alternative approach or model is readily apparent. It appears, therefore, that pragmatic approaches to reading diagnosis which are held to be good diagnostic practices and are taught to students fail to address the central problem of how to EDIEE about the information which is being collected on a child's behavior so as to arrive at a diagnosis. Research on physicians (Elstein et al., 1976) indicates that problems of integrating and combining information so as to arrive at a diagnosis are more important sources of error than a lack of thoroughness in collect- ing information. It remains for research to demonstrate whether this is also true for reading teachers. 29 Summary The purpose of this chapter has been to review the existing literature relative to the nature of problem solving in reading diagnosis. From this review, four important points emerge. 1. Very limited empirical data exist describing problem- solving behavior in reading diagnosis. 2. Findings relative to a model of clinical problem solving developed in another discipline suggest that some diagnostic problems may be solved by the iterative processing of these tasks: cue acquisition, hypothesis generation, cue interpretation, and hypothesis evaluation. Experience and training in clinical problem solving appear to improve diagnostic performance. 3. Experience and training appear to affect teacher knowl- edge of the teaching of reading. These findings may or may not be applicable to teacher problem-solving behavior in reading diagnosis. 4. Pragmatic approaches to diagnosis as described by authorities in the reading profession place a relatively heavy emphasis on the collection of information pertaining to a reading problem. These approaches appear to slight descriptions of how to analyze and synthesize that information into a diagnosis. In all, therefore, this review of the literature appears to indicate the lack of a sufficient model of problem-solving behavior for reading diagnosis based on empirical evidence. Current research efforts which are directed at determining the applicability of the Inquiry Theory of Clinical Problem Solving for reading 3O diagnosis may help to fill this void by providing data on the problem-solving behavior of teachers as they diagnose reading problems and how this behavior is acquired. CHAPTER III PRESENTATION OF THE DESIGN AND METHODOLOGY Introduction The purpose of this chapter is to describe the research design and methodology which were used to obtain data relative to the nature of problem solving in reading diagnosis. The main areas of this study are: 1. To identify and describe the problem-solving behav- ior of teachers as they diagnose a child's reading problem. 2. To determine the effects of classroom teaching experience and graduate level training in reading diagnosis and correction on the problem-solving behavior of teachers as they diagnose a child's reading problem. Design of the Study In order to study the problem-solving behavior of teachers as they diagnosied a child's reading problem, 30 teachers were assigned to three groups. The criteria for selection and assignment of subjects to groups were classroom teaching experience and gradu- ate level instruction in the diagnosis and correction of reading difficulties. The following groups were designated, each containing 10 subjects: 31 32 Group 1: CTWT (Classroom Teachers With Training)--Experi- enced elementary school teachers with graduate level instruction in the diagnosis and correction of reading difficulties. Group 2: CTNT (Classroom Teachers With No Training)-- Experienced elementary school teachers without graduate level instruction in the diagnosis and correction of reading difficulties. Gronp 3: PSET (Pre-Service Elementary Education Teachers)-- Pre-service elementary education teachers without full-time teach- ing experience or graduate level instruction in the diagnosis and correction of reading difficulties. Under the constraints of procedures which were designed to elicit problem-solving behavior, each teacher interacted with a set of materials which simulated the behavior of a child with a problem in reading. Written protocols and tape recordings of the teacher's behavior during the interaction as well as the teacher's written statement of the final diagnosis and remediation plan were analyzed to obtain data for the designated measures of problem-solving behavior. Sample of the Study Subjects The subjects were elementary school teachers and pre-service ,elementary education teachers who volunteered to participate in this study. Because of the lengthy data collection and analysis proce- dures, the sample size was limited to 30 subjects. Selection and assignment to groups was on the basis of classroom teaching 33 experience and graduate level instruction in the diagnosis and cor— rection of reading difficulties. A description of each group follows. GrOpp 1: Classroom Teachers With Training(CTWT).--The sub- jects assigned to this group consisted of 10 elementary school teachers presently teaching full time with a minimum of two years of classroom teaching experience and a minimum of two graduate level courses in reading diagnosis and correction. A list was obtained of teachers who had completed the course in Clinical Practices in Remedial Reading at Michigan State University during the previous year. A prerequisite for this class is a course in the diagnosis of reading difficulties, thus ensuring that the subjects would meet the stated requirement of a minimum of two graduate level courses in diagnosis and correction. The teachers were contacted by telephone and invited to participate in the study subject to the specified criteria. Of the ten teachersvdu1were assigned to this group, six had completed a master's degree in reading improvement from Michigan State Univer- sity. 0f the remaining four, three were enrolled in master's degree programs also majoring in reading instruction. The fourth was enrolled in a master's degree program in elementary education majoring in general classroom teaching. Several of the teachers in this group had had additional experiences worthy of note. Two of them had worked as supervisors in the Reading Clinic at Michigan State University. Their 34 responsibility was to supervise other practicum students who were tutoring children with reading problems. One of the teachers had worked for five years as a reading consultant in the public schools before returning to classroom teaching. The teachers in this group had a mean of 5.5 years of class- room teaching experience. For seven of the ten, the majority of their teaching experience was in grades 1, 2, and 3. The remaining three had taught primarily in grades 4 through 8. Five of the teachers were teaching in suburban schools near Michigan State Uni- versity, three were in a metropolitan school system near the Uni- versity, and the remaining two were in nearby rural/small town school districts. All of the subjects were teaching in public schools. Group 2: Classroom Teachers With No Training(CTNT).--The subjects assigned to this group were 10 elementary school teachers presently teaching full time with a minimum of two years of class- room teaching experience and no graduate level courses in reading diagnosis and correction. An attempt was made to select teachers whose number of years of teaching experience, teaching assignment, and teaching location would approximate those of Group 1. Thus, the composition of the two groups would be similar except for gradu- ate level instruction in diagnosis and correction and reading difficulties. None of the teachers in this group had taken any graduate level courses in the diagnosis and correction of reading 35 difficulties. Seven of the ten had taken or were presently enrolled in a foundations course in reading instruction at Michigan State University. Four of the teachers were enrolled in master's degree programs in elementary education majoring in general classroom teaching. One of these was working on a specialist degree. The remaining four, while not enrolled in a master's degree program, had all taken some graduate level courses in education. The teachers in this group had a mean of 6.3 years of class- room teaching experience. For eight of the ten, the majority of their teaching experience was in grades 1, 2, and 3. The remaining two had taught primarily in grades 4 through 8. Six of the teachers were teaching in suburban schools near Michigan State University, three were in a metropolitan school system near the University, and the remaining one was in a nearby rural/small town school district. All of the subjects were teaching in public schools. Group 3: Pre-Service Elementary Education Teachers (PSET).-- The subjects assigned to this group were 10 pre-service elementary education teachers who had passed the required undergraduate courses in reading and student teaching at Michigan State University but had not yet received their teaching experience. The subjects were contacted through an education course which all students are required to take following student teaching. From a list of students who volunteered to participate in the study, 10 subjects were selected on the basis of grade point average. The mean grade point average for the subjects was 3.1. Although it 36 has been impossible to obtain the exact mean grade point average of graduating elementary education majors from the University, it is estimated by University officials to be between 3.2 and 3.4. All of the pre-service elementary education teachers in this group had completed at least one undergraduate course in reading. Additionally, six of the subjects had completed a second undergradu- ate course in reading. The second course had become a requirement for the undergraduate elementary education curriculum within the past year. Therefore, not all of the subjects had been required to have it on their programs. Table 3.1 illustrates the most pertinent information per- taining to the experience and training of each group. The Teacher Population The teachers for this study were drawn from Michigan State University and the area public schools. Michigan State University has a total student population of 42,000 with approximately 2,500 enrolled in the Elementary and Special Education Department of the College of Education. It is located near a medium-sized midwestern metropolitan area with suburban and rural communities surrounding it. The population is predominantly white, middle class with a 15 percent minority comprised of Blacks and Latinos. Course Content Since one aspect of this study is concerned with how gradu- ate level instruction affects diagnostic problem-solving behavior, 37 .m_nm—Pm>m Ho: cavemecomcw u <\z« ~m.o _.m o.F 0.0 0.0 o.o o.o Hume <\z <\z <\z “.0 o.o ¢P.m m.m pzeo <\z <\z «<\z m.m o.~ om.¢ m.m thu m .m .m .m .w m .M mmmcm>< mcwummm cw memummm cw comuumcgou a memocmmwo mocmwcmaxm were; mmmgaou mumzumcm mmmczou mumaumcw mcwummm cw mmmgzou mewsumwh mo anco mecca -cmuca eo consaz mo topazz Papa» muuaumcu we amassz mgmm> eo .oz .mcwcwmsh new mucmwgoaxm .mcmzommhnl._.m mnm