THE TRAINING OF PROBLEM-SOLVING AND INQUIRY; I” Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY MICHAEL JAMES LOUPE ,- 1969 This is to certify that the thesis entitled THE TRAINING OF PROBLEM-SOLVING AND INQUIRY presented by Michael James Loupe has been accepted towards fulfillment of the requirements for Ph. D. degree in Education Major professor Date JAM 0469 5 war ..._ 4 s»- In"?! 732‘- f ”I -' ain‘zaa-hl-WH—t I- ’ "‘ ABSTRACT THE TRAINING OF PROBLEM-SOEVING AND INQUIRY BY Michael James Loupe The present writer and Richard M. Piper (1969) collaborated in a two-part experiment, the purpose of which was to improve the inquiry performance and effectiveness of a group of female college sophomores. The goal was to train subjects in either problem-solving strategies (present study) or openness (Piper, 1969) and then determine whether the training improved inquiry in an entirely different situation, the Teacher's In-basket, an in-basket type simulation of a teaching situa- tion. The training situation differed enough from the Teacher's In- basket that transfer to the In-basket could be ascribed entirely to learning of the strategy independent of content. The experiment had two phases: (1) Training and immediate post- testing to determine if the training had any immediate effects, and (2) Administration of the Teacher's In-basket to see if the training had any transfer value. The characteristics of the problem-solving training were derived from a number of sources including classical experimental problem-solving research, equipment trouble-shooting, and inquiry research employing simulation. The training included training in problem sensing, problem formulation, hypothesis generation, hypothesis selection, search and :- Michael James Loupe resolution or reformulation. The training took place in small groups of six to eight and was distributed over three, one-hour sessions. Detective mysteries were used as the basic material in that they could illustrate the process clearly and provide subjects with individual practice, yet maintain independence of content with the Teacher's In- basket. The materials were highly structured but were designed for group use where an instructor elicited contributions from the subjects, compiled the collected contributions of the group and provided verbal guidance and reinforcement. Seeking-style was another major factor in the experiment. Seeking- style, a composite of personality measures (Inventory of Beliefs, Barron's Complexity Scale, Word Associations, Political Position) was expected to predict one's commitment to involve himself in inquiry. Seeking-style was used as a selection variable and thus was considered an entering characteristic. Seeking-style was used in this study to provide replication of former research (Shulman, Loupe and Piper, 1968) and to look for interactions between training and seeking-style. Results indicated that it was possible to train problem-solving skills and that these skills had transfer value into the inquiry situa- tion. Problem-solving training increased the diversity of information consulted in inquiring into a problem. It also appeared to increase the competence with which problems were resolved. It did not result in significant changes in other process measures. Seeking-style did not account for significant differences in inquiry within the Teacher's In-basket. This result was explained by consider- ing the nature of the In-basket task and the entering characteristics of the group. Michael James Loupe Piper, Richard M. Opennessggraining and Facilitation of Inquiry. Unpublished Doctoral Dissertation, Michigan State University, 1969. Shulman, L. S. Seeking styles and individual differences in patterns of inquiry. Doctoral dissertation, University of Chicago, 1963. Shulman, L. S. Seeking styles and individual differences in patterns of inquiry. School Review, 73, 3, 1965. Shulman, Lee S., Loupe, Michael J. and Piper, Richard M. Studies of theignquiry Process, East Lansing: Educational Publication Services, Michigan State University, RR-22, July, 1968. THE TRAINING OF PROBLEM-SOLVING AND INQUIRY By Michael James Loupe A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Personnel Services and Educational Psychology 1969 ACKNOWLEDGMENTS Before acknowledging the many people that gave direct help and assistance on this research project, I wish to thank my major professor, Dr. Lee Shulman. It was through Dr. Shulman that I initially became interested in Educational Psychology. My career thus far has developed directly under his guidance, and it is to him that I credit many of the tools and ideas I now possess. Dr. Shulman is also responsible for help- ing develop this project from the stage of simple ideas to the final written draft. I also wish to acknowledge the help of my committee members, Dr. Joe Byers, Dr. Arthur Elstein and Dr. Robert Craig. Dr. Craig's help in reading and critiquing the thesis was particularly appreciated since it was in the final stages of preparation at a very busy time of the year for department chairmen. This dissertation is half of a joint research project undertaken with Richard M. Piper. Richard Piper and his wife, Joyce, along with my wife, Stefanie, and myself worked together as a research team for many months. Dick's comments, thoughts and help were often of aid to me. Joyce's help in typing subject logs was most deeply appreciated. For preparation of the manuscript I owe many thanks to Mrs. Joyce Stewart and Mrs. Molly Dojka. Their job often involved working nights and weekends through many drafts. ii ———'————fin For permission to reprint the extensive quotations from Sherlock Holmes I wish to thank the estate of Sir Arthur Conan Doyle, M.D. Finally, I wish to thank my wife, Stefanie Loupe, for her assistance and understanding throughout my graduate career. In this research pro- ject Stefanie shared responsibility with me for developing and writing the training materials and for selecting the excerpts from Sherlock Holmes. She observed subjects and spent many hours scoring logs. Her comments and criticism helped shape the manuscript. Her help and support made this project possible. iii _F——*'T ww~ ~, ---~-—— . .— Chapter 1 2 TABLE OF CONTENTS INTRODUCTION . . . . . . . . . . Overview. . . . . . . . . . . REVIEW OF LITERATURE . . . . . . METHODS. . . . . . . . . . . . . Subjects. . . . . . . . . . . Selection Tests . . . . . . . Design. . . . . . . . . . .'. Training. . . . . . . . . . . The Transfer Task . . . . . . Dependent Variables . . . . . Summary of Design and Methods Hypotheses. . . . . . . . . . RESULTS 0 O O O O O O O C O O O 0 Characteristics of the Sample Immediate Effects of Training Characteristics of the Dependent Variables Tests of the Hypotheses . . . . . . . . . CQT’ GOPOAO and Ianiry O o o o o o o o a Summary . . . . . . . . . . . DISCUSSION OF RESULTS. . . . . . Predicted Commitment: Seeking Style. Actual Commitment: Time in Inquiry . Skill in Inquiry: New Learning . Summary . . . . . . . . . . . SUMMARY, CONCLUSIONS AND IMPLICATIONS. Conclusions and Implications. LIST OF REFERENCES . . . . . . . iv Page 21 21 22 24 25 35 4O 44 45 47 47 51 52 57 65 66 68 69 74 76 81 82 85 88 Appendix A Problem-Solving Training . . . . . . . B Problem-Solving Test . . . . . . . . . C Competence Scoring Key . . D Tables of Correlation and Analyses of Variance . . . . E Raw Data . . Page 90 132 142 146 150 Table 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 LIST OF TABLES Page Means and Standard Deviations for Dialectical Didactic Seekers on Beliefs, Complexity, CQT Total and G.P.A. for the Present Study and for the Study by Shulman, Loupe and Piper . . . . . . 48 Correlations Among Seeking-Style Predictor ‘Variables in the Present Study and in that of Shulman, Loupe and Piper . . . . . . . . . . . 50 Means and Standard Deviations Compared for the Problem-Solving and Control Groups on P.S. Quality o o c o o o o o a o o o o o o o a o o 51 Comparison of Means and Standard Deviations of Basic Inquiry Scores for Shulman, Loupe and Piper (1968) and the Present Study (1969) . . . . 53 Intercorrelations Among Dependent Measures of Inquiry . . . . . . . . . . . . . . . . . . . . . 54 Correlations Between Scores on the Problem- Solving Test and Dependent Inquiry Variables. . . 55 Means and Standard Deviations of Inquiry Variables and Post-Test Scores for the Three Training Groups. . . . . . . . . . . . . . . . . . . . . . 58 ‘Means and Standard Deviations for Dialectical and Didactic Seekers on Selected Inquiry Measures . . 6O Correlations Between Seeking Style Predictors and Inquiry variables . . . . . . . . . . . . . . . . 61 Means and Standard Deviations of Inquiry Variables for the Cued and Non-Cued Groups. . . . . . . . . 62 Comparison of Means for the Treatment X Cueing Interaction . . . . . . . . . . . . . . . . . . . 63 vi m”; Wv .I— - ing—g- 4,,— __ {rfi“——._‘ ; rip Table 4.12 4.13 5.1 5.2. 5.3. Page Means for Selection X Cueing Interaction , , , , , , 64 Correlations Between CQT Total, College Grade Point Average and the Dependent Inquiry variables. 0 O O O O 0 O O O O I O O O O O O O O O 65 Comparison of the Correlations of Inquiry Variables and Tempo Variables. . . . . . . . . . . 75 ‘Means and Standard Deviations on Selected Variables Compared for the Problem-Solving and Control Groups . . . . . . . . . . . . . . . . 77 Within Group Correlations of CQT and P.S. Quality with Inquiry Variables . . . . . . . . . . 8O vii LIST OF FIGURES Figure Page 1 Factorial Structure of Experiment Showing the Number of Subjects per Cell. . . . . . . . . . 24 2 Mean Scores for Dialectical and Didactic, Cued and Non-Cued Groups on Bits and Shifting Tempo . . . . . . . . . . . . . . . . . . 71 3 Mean CQT Total Scores for Dialectical and Didactic, Cued and Non-Cued Groups . . . . . . . . 72 viii V mt ': CHAPTER 1 INTRODUCTION Men have asked questions of themselves and of the world around them during all of recorded history. Techniques of systematically posing questions and organizing the information gathered have been codified many times as systems of logic or scientific method. This questioning, gathering and organizing information is part of the pro- cess known to Dewey (1938) as inquiry. It was the purpose of this study to train people to be effective inquirers; to be more effective in gathering and organizing information about problems, and to be more effective in sensing them. Two approaches to the training were taken. The first was an attempt to improve inquiry performance by training in problem-solving techniques which are relevant to a wide range of topics, such as problem formulation and hypothesis formation. That attempt is the focus of this document. The second approach, training in openness, proposed to improve inquiry performance by effecting changes in personality characteristics thought to be determinants of inquiry. The investigation of openness training is treated in detail by Richard M. Piper (1969). Dewey (1938) defined inquiry as the: controlled or directed transformation of an indeter- minate situation into one that is so determinate in its constituent distinctions and relations as to convert the elements of the original situation into a unified whole . . . inquiry is competent in any given case in the degree in which the operations involved in it actually do terminate in the estab~ lishment of an objectively unified existential situation. (pp. 104-105) Lil: ' v . Av." 127%" 5.3“” ' ‘ . 2 He elaborates further by saying that inquiry begins when a situation is judged to be problematic and has progressed well when one can iden- tify the problem or problems presented by the situation. Inquiry reaches its most competent, complete level when the ideas generated during the inquiry are put into actual practice and data gathered from observation of the working system are organized into a coherent whole 'with the other facts. To Dewey, then, inquiry begins when one realizes that a problem exists but does not yet understand its nature. It con- tinues with correct identification of the problem, recognition of the facts relevant to it, and finally with implementation of a trial solu- tion. Inquiry is competent to the degree that the actual problem is solved and the nature of the problem is understood. Two characteristics of this definition are particularly important. First, Dewey notes that although the problem is usually an objective reality, inquiry does not begin until someone senses it. Inquiry begins .with the act of problem sensing. If the problem is directly given to the subject and the solution is then sought, this is not by Dewey's definition a total inquiry. Second, an objective solution to the prob- lem is desirable and indicates competence in inquiry; likewise do understanding and integrating of the facts. Solution itself is not the sole criterion of effective inquiry. What is described here as inquiry differs in two ways from what is normally considered problem- solving. First, inquiry includes the step of sensing the problem, while in problem-solving research, the subject is usually given a problem. Inquiry is well underway before one actually knows what the problem is (Dewey, 1938). Secondly, problem-solving, as is implied by the name, is often evaluated wholly on the basis of the solution which, to be sure, is a criterion of inquiry effectiveness; but it certainly is not the complete criterion. 3 What are the characteristics which affect one's competence as an inquirer? Shulman (1965) reasoned that there were a number of person- ality characteristics which would, in part, determine one's effective- ness as an inquirer. Openness or nonstereopathy should aid in sensing problems and formulating them realisticially in terms of situational constraints, rather than personal biases and sets. Because of the com- plexity of the inquiry process, individuals who preferred complexity would probably be more willing to engage in inquiry than those who did not. Those people having these characteristics (plus others such as word-association fluency, liberal politics, high willingness to risk, and preference for discussion rather than lectures) were called Dialectical seekers after the dialectical mode of teaching which Shulman felt they would be predisposed to use. Those having the opposite set of character- istics were called Didactics. Dialectical seekers were predicted to surpass Didactic seekers in inquiry competence. In order to study inquiry in an experimentally controlled situation which simulated realism and allowed for individual sensing of problems, Shulman (1963) developed the Teacher's In-basket. This is a simulated situation using the in-basket technique (Frederiksen, et. a1., 1957) in which the subject role-plays a beginning teacher (see Chapter 3 for a detailed description of the technique). The Teacher's In-basket was used by Shulman, Loupe and Piper (1968) to check the above predictions regarding the determinants of inquiry, and to describe the behavioral processes of the inquiring individual. Within a sample of teachers-in- training, personality type or seeking-style as it was called, was a significant determinant of inquiry performance. Seeking-style pre- dicted quantitative measures of the inquiry process such as amount of information used, problems sensed, and time spent, as well as the qualitative outcome measure of inquiry competence. The second phase of the research by Shulman, Loupe and Piper (1968) involved analyzing the characteristics of the inquiry behavior itself to see if these characteristics reflected differences in the outcomes of inquiry. One such analysis involved looking at the sequence in which subjects consulted different types of information. The more subjects shifted between types of information, the more effective was their result- ing inquiry. A possible explanation was that this shifting was a result of comparisons being made between different pieces of information. Both the Dialectical-Didactic personality types and the behavioral measures such as shifting were significantly correlated to the dependent measures of inquiry and yet were correlated among themselves at a rela- tively low level. This indicated that the behavioral process measures such as shifting were not as highly related to the personality deter- minants of inquiry as had been hypothesized. These two categories of variables could be considered as relatively independent, both having an influence on inquiry performance and outcomes. Thus, there seem to be two relatively separate components which influence the inquiry be- havior of the individual: A personality component which tends to influence the amount of inquiry in which one engages, and an intellectual component which tends to influence the way one uses information and in- quires while in the situation. These results provided the basic rationale for the present overall experiment. If seeking-style and behavioral strategy are two relatively independent components which affect inquiry and its outcomes, then inquiry performance should be predictably changed by a change in either component. As was mentioned earlier, this study was primarily concerned with changing inquiry via training in general problem-solving strategies. The basic hypothesis was that subjects so trained would surpass those given a control training, both in inquiry competence and in certain measures of the inquiry process such as shifting and information sources used. In order to test the hypothesis, a procedure was developed to give short—term training to subjects in techniques of problem-solving and inquiry. Following training, subjects were observed in inquiry as they role-played a teacher in the Teacher's In-basket. The basic features of the training procedure were derived from a wide range of sources (including Dewey, classical problem-solving literature and applied trouble-shooting) to assure generality of the process. Sub- jects were taught to: (a) look for problems, (b) relate all available information to the problem, (c) develop as many explanatory hypotheses as possible, (d) select a likely hypothesis and determine how to test it, (e) search for pertinent information in accord with the testing plan and reformulate the problem in terms of the new information. These skills were taught in the context of solving detective mysteries, to avoid overlapping in content with the Teacher's In-basket. Thus, if inquiry performance did improve, it could be attributed to learning of the strategy alone and not to familiarity with the content. This particular research project has both theoretical and applied relevance. Theoretically, it is concerned with questions about the effects of relatively short-term training in openness or problem-solving on inquiry performance. Will problem-solving skills learned in one con- tent domain facilitate inquiry performance in a different content domain? Likewise, will teaching a person about openness and how to behave as if he were open, influence his predisposition to inquire? 6 The applied relevance of this research derives from the importance of the inquiry process. If inquiry skills are general enough that training in a set of general problem-solving strategies in one content domain transfers to another, it may be well to consider instruction in these skills per se. For example, Shulman, Loupe and Piper (1968) speculate that inquiry ability and teacher effectiveness are related. If so, and if inquiry is in fact "teachable," then this would imply incorporating instruction in these skills into teacher preparation programs. Overview In Chapter 2 the research relevant to inquiry and training in problem-solving is reviewed, followed by a description of Methodology in Chapter 3. Chapter 4 contains a summary of the experimental results which are discussed in Chapter 5. Chapter 6 contains a final summary, the conclusions, and a discussion of implications. CHAPTER 2 REVIEW OF LITERATURE As was mentioned in Chapter 1, the research by Shulman (1963) and Shulman, Loupe and Piper (1968) was based on the theoretical typology of the Dialectical and Didactic seeking styles. The basic hypothesis in both of the former studies was that this typology, representing the two extreme poles of a continuum of seeking style, would predict inquiry performance. In particular, the basic prediction was that sub- jects identified as Dialectical would surpass those identified as Didactic in their ability to inquire effectively in complex situations. The Dialectical profile included characteristics such as preference for complexity, willingness to take risks, field-independence, fluency of word-association and preference for lecture over discussion. Didactics tended toward the opposite pole on each of these six measures (e.g., field-dependent rather than independent). The theoretical position was that these tests measured personality characteristics which would predispose the Dialectical seeker to be a more effective inquirer than his Didactic counterpart. He preferred the complex, was open to his environment, was willing to risk, and thus should have been more willing to commit himself to inquiry than the Didactic who had none of these characteristics. The investigators viewed these typologies as anchor points on a continuum of seeking style pre- dispositions along which people were distributed. In order to check the viability of the theory, subjects were chosen to represent the extremes 7 8 as closely as possible. They were then placed individually in a situa- tion.which would encourage inquiry and allow reliable observation of inquiry behavior, the Teacher's In-basket. The Teacher's In-basket was designed by Shulman (1963) to provide the opportunity for inquiry within a setting in which the structural cues were minimal (e.g., make out a list of referrals to the school psychologist), and yet the potential situational stimuli were essen- tially the same from one person to the next. The emphasis was on main- taining realism of situation as well as experimental control in order to answer the research questions in a way that would overcome some of the problems of abstraction found in much experimental research (Rund- quist, et. a1., 1965). The Teacher's In-basket was a situational simulation in which the subject was asked to role-play an inexperienced teacher beginning her first teaching job, mid-year. The permanent teacher vacated the class- room two months before and the class had been run by substitutes since. The children were not at the school at the simulated time because it was Record Day, a day to let the teachers catch up on paper work. It was a perfect opportunity for her to deal with administrative tasks and become acquainted with her students' records before meeting them the following day. Potential problems were imbedded in the material which included memoranda and records on each child, among other things. Chapter III will provide a more complete description of the Teacher's In-basket. In the in-basket situation, the subject may or may not sense a problem and, once he senses it, he may or may not choose to inquire further into it. The information employed in inquiry may vary as may the time spent; and of course, the quality of the inquiry itself may 9 vary considerably. These facets of inquiry as well as the subject's sequence of search through the information are quantified. Dialecti- cals and Didactics were predicted to differ on the resulting measures because of differences in their willingness to inquire and their char- acteristic styles of attack once committed to the process (Shulman, Loupe and Piper, 1968). The findings of earlier research (Shulman, Loupe and Piper, 1968) by and large supported the theoretical position: Dialecticals were more effective inquirers than were Didactics. In addition, post-hoc analysis uncovered systematic differences in the behavioral patterns of these two groups, differences which were reflected in the increased effective- ness of the Dialectical over the Didactic inquirers. There was one additional facet to the experiment. Each subject participated in the Teacher's In-basket twice; once each with alternate, but quite equivalent, forms during fall and spring quarters. During the intervening winter quarter, the sixty subjects left campus to student- teach. Inquiry performance was quantified in a number of ways, each designed to tap a somewhat different facet of the process: (a) Problem sensitivity was the number of potentially problematic elements reacted to as problems by the subject. (b) Bits was the number of pieces of information to which the subject attended. (c) Sources was the number of different kinds of information used in inquiring into a specific problem. (d) Competence was a measure of quality of problem resolution reflecting the subject's understanding of the problem. (e) Time was the total time spent in the situation. 10 (f) Shifting was the total number of times the subject moved from one kind of information (source) to another. The results indicated that Dialectical inquirers exceeded Didactic in- quirers in all of the above measures of inquiry during the second admin- istration and, in most cases, during the first administration. They also indicated that as the influence of seeking style increased from the first to second administration, the influence of G.P.A. on inquiry decreased. Specifically, the Dialectical-High G.P.A. group tended to remain constant in inquiry performance from the first to the second administration, the Dialectical-Low G.P.A. group tended to increase their score markedly, and the two Didactic groups decreased consistently on every measure. These results indicate that inquiry performance is, in part, a function of personality style. The fact that shifting was highly correlated with the inquiry variables and relatively uncorrelated with seeking style indicates that inquiry may also be a function of a learned strategy, a strategy which encourages shifting. If so, then inquiry performance should be changeable by changing either the seeking style predispositional component 25 the learned strategy component. As was stated in Chapter I, the focus of this study is on changing the learned strategy component of inquiry performance. Piper (1969) has studied the effect of changes in the a f fec tive component. Since the stated goal of this research is effecting change in inquiry behavior, it is important to specify precisely what is to be considered inquiry. Is it a 22212 set of behaviors which must be con- sidered as a unit as Suchman (1961) has stated, or can it legitimately be analyzed into its components? Let us begin with the model of inquiry offered by Shulman (1965) which views inquiry in terms of four basic 11 processes: problem sensing, problem formulation, search, and resolution. This model is based on Dewey's position (1938) that inquiry is the process of transformation of an indeterminate situation into one of determinancy. Other models of inquiry will be compared and contrasted to this one so we may derive the model which formed the basis for the present study's program of training in inquiry. R. L. Thorndike (1950) reiterated and elaborated upon the basic analysis of inquiry defined by Dewey. Although Thorndike structured his discussion around problem-solving, his definition of the process was broad enough to go beyond the traditional limitations of laboratory studies and was applicable to inquiry as defined in Chapter I. He began by defining a problem in terms of motive: A problem is an un- satisfied motive or need for which the person does not have an immedi- ately available response and which is recognized by the person. Aside from the emphasis on motive, this is not unlike the position of Dewey (1933). Becoming aware of the problem, Clarifying the problem, Proposing hypotheses for solution of the problem, Reasoning out the implications of the hypotheses, Testing the hypotheses against experience (Thorndike, 1950, p. 196). U'Iul-‘tflNH Thorndike acknowledges the abstract quality of his analysis. He also recognizes that these phases or steps rarely occur in such system- atic fashion. Presumably, as research and theorizing in the area of inquiry continue, the conditions under which these various steps occur should be clarified. It is interesting to note the ties between this analysis and other research in problem-solving -- for instance, the research of Meier (1936) dealing with set. The second phase of inquiry is clarifying the problem (problem formulation), analyzing the problematic situation to determine 12 the specific conditions placed on the solution by the situation itself. This, of course, has the function of breaking sets which so often inter- fere with solution (Thorndike, 1950). Phases dealing with hypothesis generation and testing allow one to use all relevant data and logically test out the implications of each, systematically suspending judgment. Another source for analysis of problem-solving is the literature on equipment trouble-shooting. Due to the complexity of electronic equipment it became apparent that if men were to go from observable symptoms to diagnosis and repair malfunctions economically, systematic strategies were necessary. Bryan (1962) derived a nine-stage analysis of trouble-shooting that was essentially an elaboration of the above, stated in terminology pertinent to trouble-shooting: (a) Symptom recognition: Initial awareness of a problem, (b) Symptom elaboration: Exploration of stimulus situation to define the problem, (c) Formulation of hypotheses, (d) Selection of hypothesis to test, (e) Formulation of a testing plan including criteria for accepting or rejecting hypothesis, (f) Acquisition of system status information via use of half-splitting techniques, (3) Interpretation of discrete datum, (h) Organization of all information, (1) Selection of alternative hypothesis, (j) Confirmation of hypothesis. The value of a number of the steps in this analysis has been confirmed by research in trouble-shooting. Moore, Saltz and Hoehn (1955) demon- "c," "d," and "e" by giving subjects strated the value of steps instruction in preplanning. Preplanning consisted of generating as many hypotheses as possible to explain the observed symptoms. Subjects then selected an hypothesis to test and formulated a testing plan complete with criteria for accepting or rejecting the hypothesis. They found, not surprisingly, that the preplanners were significantly better trouble- shooters than the controls. What this implies is that the analysis of 1* l3 trouble-shooting (inquiry or problem-solving) is probably not an analysis of actual observed behavioral sequences but rather a prescription; trouble-shooting will be effective if done in this manner. The above analyses include the same basic set of components: Prdblem sensing, problem definition or formulation, hypothesis forma- tion (not included by Shulman), search and resolution. One might make the generalization that the more closely one's inquiry behavior approxi- mates this analysis (without placing ordering restrictions) the more effective an inquirer such a person will be. This generalization is supported by Fattu (1964) in a review of results of trouble-shooting research. The good trouble-shooter actively looked for symptoms. He was aware that problems would occur and attempted to sense them immedi- ately. In contrast, the ineffective trouble-shooter actively checked symptoms only half of the time and often began remedial action or came to a conclusion too soon and, thus, the solution was ineffective. Finally, when the correction did not return the system to working order, the poor problem solver tended to continue checking within the same subsystem, while the good problem solver broke set and returned to verification of symptoms, considering once again the whole systems From the above discussion one might conclude that there exists a general strategy for problem-solving but, because different types of problems have vastly differing external structures, it might also be that strategies for solution are dependent upon the structural character- istics of a situation, i.e., problem—solving is situation specific. Miller and Slebodnick (1958) approached the structure vs. strategy question from a pedagogical position. Their study was a simple comparison between the effects of training in trouble-shooting strategies and training in the concept of data flow (structure) on subsequent trouble-shooting 14 performance. The trouble-shooting training consisted of instruction in the general strategies of (a) narrowing trouble to a chain of sus- pected components via deduction, (b) making checks via half-splitting, a search strategy similar to conservative focussing, and (c) identifying the malfunctioning unit. Structural training centered on logic and flow of data within the piece of equipment. Results indicated that the group trained in strategy was faster than the structurally trained group. Rundquist, et. a1. (1965) substantiated these findings in a setting very different from the applied trouble-shooting situation: The tradi- tional laboratory concept-learning situation. Subjects were trained (a) to respond to the structural (dimensional) characteristics of a set of stimulus materials identical in structure to that of the criterion task or (b) to use conservative focussing to solve problems such as number series and matrices which were structurally unlike the criterion task. Although results were difficult to interpret because of method- ological problems, the strategy trained group did better than the structure trained group in terms of time to solution (less) and number of correct initial hypotheses. The strategy training discussed above dealt primarily with strategies of hypothesis testing. However, it appears that for complex realistic inquiries, training in strategies of search planning and information organization may be more effective than hypothesis testing strategies such as half-splitting and scanning. Miller and Folley (1953) I prescribed half-splitting as the solution to the complexities of trouble- shooting problems. Simply stated, half-splitting is a search strategy similar to conservative focussing (Bruner, Goodnow and Austin, 1956) in which the subject selects the point at which to make a binary check which will eliminate the maximum number of possible alternatives. Later, _ an; ._. - i-._._.._-o , ‘*" . :4feLu ,, » ,- - a: 15 Goldbeck, et. a1. (1957) trained one group of subjects in half-splitting and a second group in deductive selection of possible trouble units within which the malfunction might be located. Results indicated that half-splitting was successful only when problems were very simple. Otherwise training in logical deduction (including hypothesis genera- tion) was more effective in facilitating trouble-shooting. Finally, Goodnow and Pettigrew (1956), in training subjects to solve a simple alternation problem, concluded that a major factor in problem-solving involves the attention that the inquirer gives to the events or information presented. Instruction or training which has the effect of directing attention toward the events should improve perfor- mance. This was hardly a striking or new finding, and yet they felt it was worthy of reiteration because many training procedures or instruc- tions required the subject to attend so closely to his own responding that attention was directed away from the events from which information was to be gathered. 0n the basis of this research (Goldbeck, et. a1., 1957; Goodnow and Pettigrew, 1956), it was decided to eliminate hypothesis testing routines from the present training procedure. In accord with the above research on problem-solving and inquiry, the cognitive training procedure included training in: (1) problem sensing -- careful attention to information presented and inspection of such material to detect discrepancies be- tween the information and one's general expectations for the situation or between pieces of information within the situation. (2) problem formulation -- specification of interrelationships among bits of information within the problematic situation and the specification of limits and conditions placed on the solu- tion by the problem itself. 16 (3) hypothesis construction -- generating a number of possible causes of the problematic situation. (4) hypothesis selection -- selection of an hypothesis to test and specification of a testing plan and criterion for accept- ing or rejecting it. (5) search and resolution -- inspection of information in accord with the testing plan, including re-emphasis of problem sens- ing and problem reformulation. The above training took place in small groups of six to eight. Using simulated problematic situations and anecdotes, each of the steps was explained to the subjects, after which it was discussed and practiced by the subjects. The results of each step were discussed to point out the benefits of proceding in such a systematic manner. Probably one of the most important phases of the training is the problem sensing and formulation, both in the original situation £29 during search. As the subject searches for information to test an hypothesis, he will sense new problems and uncover new information that may add to or change the basic nature of the problem as formulated. He must then reformulate the problem and re-evaluate his hypotheses. The process is analogous to the continual restructuring, which is the basis for the learning by discovery approach as discussed by Bruner (1960). However, the purpose of the proposed research is not to facilitate in- quiry byextended practice in inquiry as done by Suchman (1961) and Crutchfield and Covington (1963). The purpose is to faciliate inquiry by Didactic training in specific components of effective inquiry. Suchman's inquiry training (1962) demonstrated that inquiry perfor- mance could indeed be modified by experience. The particular nature of 17 the experience was extensive practice in asking questions to gather data. The problems set for the students (fifth grade) were movie simulations of apparent scientific anomalies. The task was to discover the cause for the observed discrepancy by asking questions which could be answered with a simple "yes" or "no." The experimenter answered questions and guided the students' questioning by requiring additional formulation when the question was not a simple binary test. This training experience extended over a period of twenty-four weeks and resulted in marked facilitation of inquiry performance on the part of the students. As can be seen from the nature of the training, there appears to have been no explicit model of inquiry underlying the training. The only skill explicitly taught was question asking. This would follow from Suchman's position (1961) that inquiry is a monolithic ability, not properly analyzed into components. The only way for Suchman to teach inquiry is through experience with inquiry. However, as Rundquist (1965) put it, "in the applied setting 133 opposed to the lab abstraction of concept attainmeng7, it is precisely because of the inefficiency or inappropriateness of training-through- doing that pre-training procedures are introduced." (p. 26) Crutchfield and Covington (1963) came closer to training inquiry based on a prescriptive analysis of the process such as developed here. The Productive_Thinking_P;ggram (1966) employed extended practice in inquiry with emphasis on hypothesis generation and strategy planning. One of the distinctive features of that particular approach was that the steps and procedures learned were made explicit, subjects were told to think of a number of possibilities and were rewarded for doing so. However, neither the research by Suchman nor by Crutchfield and Covington demonstrated changes in the inquiry process itself within a situation distinct from that of the training. Facilitation of classroom 18 discussion was noted, but no clue was given as to changes in the inquiry process itself. The present study examines changes in this process by testing trained subjects with the Teacher's In-basket, a situation entirely independent of the problems and setting of the training situa- tion. Another type of approach to changing inquiry behavior that has been incorporated into the present study involves manipulation of temporary fluctuations in performance via cueing (as opposed to basic changes in capability via training). Colgrove (1968) demonstrated that in a problem- solving situation, subject performance could be significantly improved by providing a cue. Subjects were given the Changing Work Procedurgs problem in which they were given data on work efficiency of three men on an assembly line. The problem was to make a series of administrative decisions involving the men in order to maintain work efficiency and resolve dissention among the workers. Subjects cued to be original offered more solutions to the problem than those not cued. In addition to the training of problem-solving skills, the present study includes the cueing of specific subjects to examine whether changes in capabilities resulting from the training are made more manifest in subsequent performance on the Teacher's In-basket when a cue is provided. A practical problem which had to be dealt with in constructing the training concerned transfer. It was necessary that the training include no content material which overlapped the Teacher's In-basket (the transfer task). However, transfer is a function of stimulus and response similarity in the training and testing situations. This posed a problem because to the degree that the situations (training and testing) did overlap, the experimental conclusions would be ambiguous. The problem involved main- taining task similarity without content overlap. was“ ._ ....,. “PM -VW‘H -. ~ v ‘-:.. :1 .:_.._.....£..s—-x-’14 .r a ~._‘~,.. nan-h... .4..- ,, ~0- . . .~ 19 When Andrews, Cronbach and Sandiford (1950) wrote their article on transfer, similarity had been defined in as disparate terms as identi- cal elements and perceptions of similarity. Andrews, Cronbach and Sandeford (1950) made some suggestions about maximizing transfer of instruction which seemed to consider both of these positions. Accord- ing to them, transfer is facilitated if: (a) students expect to use what they are learning in different situations, (b) the generalizations are meaningful and are practiced as behaviors, not just verbalisms, and (c) students can recognize the new situation as a special case of the old. These suggestions were accounted for in the present training pro- cedure by (a) emphasizing that the strategies being taught were pertin- ent to teaching, (b) providing materials with which to practice and refine the skills, (c) emphasizing the structural nature of the problems to be solved rather than that of the specific situations in which they were cast. The procedure itself is described in full in the following chapter. The present experiment was composed of two phases. Phase I included training of problem-solving skills followed by an immediate post-test. Phase II was administration of the Teacher's In-basket and analysis of the observed inquiry performance. Stated generally, the hypotheses were that subjects trained in the problem-solving strategy would exceed sub- jects receiving an irrelevant training in a number of facets of the in- quiry process. They should exceed controls in problem sensitivity because the training emphasized looking carefully at information in order to find problems. Generation of hypotheses and planning of a search plan should lead subjects into a more diverse inquiry pattern and, thus, 20 should increase the number of sources of information consulted and the amount of shifting engaged in. Finally, this type of inquiry should result in increased competence of problem resolution. CHAPTER 3 METHODS Subjects Sixty female students in the beginning undergraduate course in Educational Psychology at Michigan State University were selected from approximately 160 females taking the screening battery. Subjects were not selected at random from a large population but were selected pur- posely to represent two ideal types here called Dialectical and Didactic. A Dialectical subject was defined as a person scoring above the median on at least three of the four screening tests. Conversely, a Didactic subject was defined as one who scored below the median on at least three of the four tests. Since our interest was in selection of extreme subjects, subjects with all four test scores either above or below the median were chosen first. Only then were subjects who scored consistently on only three tests chosen. These subjects (with one deviant score) were selected on the basis of the proximity of their deviant score from the median. Thus the thirty subjects most extreme in the Dialectical direction and the thirty most extreme in the Didactic direction were selected from the 160 subjects taking the selection tests. vAt the time of testing all subjects were informed that only certain people would be contacted on the basis of their test scores (cognitive styles) for further participation. They were told that those who agreed 21 -. «w- "W- :’~w - 22 to participate would be paid $1.50 per hour for approximately six hours of work over a seven-day period and that, in addition, they would learn some things relevant to teaching. Subjects were contacted by telephone and approximately 90% agreed to participate. Those declining generally did so because of very heavy work and class schedules and/or family responsibilities. All subjects were paid $1.50 per hour and a $2.00 bonus if they kept all appointments. They were also told that their work was valuable only if they completed all phases of the study and, thus, they would be paid only if they completed all work. Selection Tests The selection battery consisted of four group-administered inven- tories and tests which took approximately thirty-five to fifty.minutes to complete. The four measures comprising the battery were Word Associa- tions (Getzels and Jackson, 1962), Complexity (Barron, 1967), Political Position (Shulman, Loupe and Piper, 1968) and the Inventory of Beliefs (Stern, Stern, and Bloom, 1956). The Word Associations test consists of a list of twenty-five words, each of which have a large number of distinct meanings (e.g., coil, punch, host, etc.). The subjects' task was to give as many meanings for each word as possible within a time limit of ten minutes for all twenty-five words. The instructions suggested that definitions need not be written out but that in most cases, meanings could be indicated by simply writing down one word (e.g., bark -- tree, dog, seal, boat). The test differs from the traditional word association techniques which ask the subjects to give words which the stimulus word brings to mind, not necessarily words which suggest a definition of the stimulus word. The Word Associa- tions score was a composite of the total number of different meanings (three 23 in the above example) and the total number of responses given (four in the above example). Word Associations correlated .42 with the MSU Reading Test and .45 with CQT Total in research by Shulman, Loupe and Piper (1968). The Complexity scale consists of thirty statements designed to elicit a statement of preference for either simple or complex situa- tions (e.g., "I like to have a place for everything and everything in its place." "Some of my friends think my ideas are impractical, if not a bit wild.") The subject either agrees or disagrees with each item. A high score indicates preference for complexity, ambiguity, the unpredicted, the assymetrical, etc. Complexity correlated .50 with the Inventory of Beliefs and -.06 with CQT Total in research by Shulman, Loupe and Piper (1968). The Politics Scale is a simple, four-item self-report form asking subjects about the political positions of their parents and themselves. Each of the four items is scored in the direction of liberal politics, thus a high score indicates reported liberalism. This Political Scale is a revision of the one used by Shulman, Loupe and Piper (1968). In that research, Politics correlated .23 with the Complexity Scale, .36 with the Inventory of Beliefs, and .39 with Word Associations. The Inventory of Beliefs is a collection of 100 items or statements based on cliches and slogans of ten heard in the 1950's dealing with such topics as politics, labor and religion. The inventory was devel- oped by Stern, Stern, and Bloom to measure stereopathy -- nonstereopathy, or roughly the equivalent of openness and non-authoritarianism. The Inventory of Beliefs correlated -.42 with Rokeach's Dogmatism Scale (1960) 24 and .08 with CQT Total (Shulman, Loupe and Piper, 1968). The median reported reliability is .86 (Stern, Stern and Bloom, 1956). Design The experimental design consisted of three factors: Training, cueing, and seeking-style. The three levels of training were problem— solving training, openness training, and control. Cueing and seeking- style were both dichotomous (cueing vs. no cueing, Dialectical vs. Didactic). As indicated in the description of the subject sample, subjects were selected to most closely approximate the Dialectical and Didactic seeking-style types. Subjects in each of these two groups were assigned randomly to one of six combinations of the treat- ment and cueing factors. Thus the total experimental design consisted of three crossed factors (3 x 2 x 2) having twelve individual cells and five subjects per cell. Problem-solving, Openness Control cue 5 5 5 15 Dialectical no 5 5 5 15 one cue 5 5 5 15 Did ti 8° ° no 5 5 5 15 cue 20 20 20 60 Figure l. Factorial Structure of Experiment Showing the Number of Subjects Per Cell 25 Training All training took place in groups of six to eight subjects which met for four one-hour sessions, the first three of which were devoted to training and the last to testing. Problem-Solving As stated in Chapter 1, the central problem of this document is facilitating inquiry performance through training in problem-solving. The problem-solving training consisted of reading, discussion, and both individual and group practice in the components of inquiry: prob- lem sensing, problem formulation, hypothesis generation, hypothesis selection, search and resolution 25 redefinition. The objectives of the training were that, given a situation with unidentified potentially problematic elements, the subject would: (a) carefully inspect the situation, looking for discrepancies either between her expectations and the information or between bits of information in the situation, (b) formulate the problem by systematically relating all available information tothe problematic situation, (c) generate a number of possible hypotheses which could explain the problem as formulated, (d) select an hypothesis to test and specify a testing plan including criteria for accepting or rejecting the hypothesis, (e) search the available information in accordance with the testing plan and either resolve the problem and cease inquiry or use the resultant information to reformulate the problem. The instructional materials were primarily excerpts taken from Sherlock Holmes mysteries. These were used both as illustrations and as a basic materials for instructional exercises. In addition to the mystery stories, puzzle type problems were used as illustrative exercises. fi—mn--,i(ie 26 Three additional problems were constructed in which there was a brief, one-paragraph statement of a problematic situation follwed by sixteen to twenty-three pages, each of which contained an additional piece of information (see Appendix A). The subject matter of the problem and the indexing system which gave access to the materials were different for each of the three problems according to the purpose for which it was constructed. The first of the three dealt with the problems of a playground supervisor. Each subject was given a complete table of contents for the problem. The purpose of this problem was not to instruct or test but to demonstrate the relevance of strategies to problem-solving before instruction began. Thus, it did not matter that the subject's hypothes and questions might have been structured by the index. The second problem, a murder mystery, was organized formally in the same manner as the first; however, no indexing aids were provided the subject. Because subjects worked with this problem for thirty to forty-five minutes with continuous feedback and discussion, it was important that the content of the mystery problem did not overlap with the content of the Teacher's In-basket to avoid facilitating their in-basket performance simply by teaching them to use the particular content contained therein. It was also important that the subjects were not constrained in their hypothesizing and questioning of a list of sources. When using this problem, subjects raised hypotheses and asked the instructor for the location of the information source thought to contain the answer. The third problem (see Appendix B), involving an unemployed sales- man, was designed as a test of problem-solving ability to be used as a training post-test for all subjects. To function as a test it was desirable that the constraints to hypothesizing and questioning be 27 minimal, that the same set of information be equally available to all subjects and that the indexing be self-contained to allow completely individual work on the problem. Thus, the material was indexed as a branching program with general categorical indexes and specific sub- indexes available to the subjects by following particular branches. After each piece of information or index the subject was given the option of seeking additional information or completing the program by proposing a solution. This particular training employed materials of high intrinsic interest, techniques of individual questioning and constant immediate feedback to maintain high subject involvement and reinforce the target behaviors of the training. As was mentioned above, the training took place in three one-hour sessions. Each session was ended by raising a problem for subjects to consider overnight. They were also given two brief puzzle problems to do between sessions. Although the time re- quired out of class totalled only about thirty minutCS, subjects reported working on problems with friends and roommates and were very anxious to find out the "answers" and gain closure. The purpose of the unfinished problems was to take advantage of the lack of closure to maintain curiosity and aid recall of the past day's activities as in the Zeigarnik phenomenon. The homework problems were simply to gain an additional exposure to training related materials and provide additional practice and illustration of the principles discussed and practiced during the day. First Training Session At the beginning of the first training session subjects were told that their individual participation was necessary. They were encouragai 28 to ask questions or make comments at any time. They were then told that the purpose of the next few meetings was to train them in problem- solving skills. The next few minutes was devoted to arousing interest and motivation. They were told that their time would be well spent since traditional teacher training did not adequately prepare teachers either to sense or solve problems in the classroom. Brief anecdotes to illustrate the importance of problem-solving abilities to both the preactive and reactive (Jackson, 1966) phases of teaching were elicited from the group. Then the subjects were asked to takeffiye to ten minutes to solve the problem (by finding the cause of the situation) of a play- ground director who was having difficulty getting children to come to her ghetto area playground (see Appendix A). The problem was in book type format but was administered in the same way as the Rimoldi-type tab tests (1960). Each subject was asked individually what she thought the problem was and a group answer was compiled. This was compared to a model answer. The problem was constructed so as to lead an individual into stating any one of three possible solutions. To reach the "best" answer, which involved an interaction among at least three factors, subjects were told they must follow a particular strategy and this was briefly illustrated. Subjects were then told that research had shown that the best problem solvers in many fields (medicine, teaching, trouble-shooting, etc.) all seemed to use a particular technique or strategy, the strategy that they would learn over the next three days. Further, those people identified as good problem solvers were considered to be superior practitioners within their fields of endeavor. At this point the formal instruction began. Subjects read a four- page explanation of the steps (see Appendix A) stopping after reading as 29 about each step to discuss important points and ask questions about the examples. The balance of the first meeting was concerned with explain- ing and practicing the phases of problem sensing and problem definition with the major emphasis being placed on attention to detail. As an example, subjects were asked what they could tell about a person simply from examining a pair of his (or her) glasses. The instructor asked each subject for contributions and listed them on the chalk board. Subjects then read a brief excerpt from Sherlock Holmes' "The Adven- tures of the Golden Pincenez" in which Holmes describes the culprit down to details such as size, posture and facial expression from cues found in a pincenez. A pair of eyeglasses was passed around among the subjects. The group was instructed to look for characteristics of this particular pair of glasses which would aid in describing the owners. Each subject was asked for at least one contribution. A11 points were listed on the chalk board and the group then proceeded to relate the pieces of information into a tentative description of the wearer and the circumstances under which the glasses were worn (definition of the problem). The first day of each training session ended by passing out a list of clues found by Sherlock Holmes at the scene of a murder in "The Lauriston Garden Mystery." Subjects were to take the list home and "define the problem" or describe the culprit and the scene as carefully as possible by interrelating the clues provided. They were also given two short puzzle-problems to solve overnight as illustrations of probknn sensing and problem defining. — __.......-M.—. -- _, o. 7 ,, .--, if. 1-....- 7, 30 Second TrainingiSession The second day began by discussion of the home assignments. Each subject was asked what he thought was the important information in each problem and how the information and the problem statement were related in order to reach a solution. From the contributions a group description of the Laureston Garden case was generated. Subjects then read the description offered by Sherlock Holmes. Comparisons were drawn between the group description and that of Holmes. The manner in which both the group and Holmes related the discreet data into a meaningful description was briefly discussed. At this point, the focus shifted from problem sensing and problem definition to hypothesis generation. Subjects reread the description and examples of hypothesis generation and selection (see Appendix A) and briefly discussed the purpose of these steps. It was emphasized that by generating as many hypotheses as possible which could account for the problematic situation, they would avoid the error of unknowingly committing themselves to a particular solution before having gathered sufficient information to make a decision logically. Subjects were told that the good problem solvers -- troubleshooters, teachers, physicians, etc. -- consciously and consistently use this technique (hypothesis forma- mation) and it has proven to be very successful. A simple medical example of hypothesis formation was used as illustration. The balance of the second session was devoted to group practice in hypothesis formation. Key passages were selected from two Sherlock Holmes short stories, "The Adventure of the Norwood Builder" (pp. 498-500, 503), and "The Adventure of the Three Students" (pp. 596-597, 600-601), to present sets of information and clues from which the problem could be 31 defined and hypotheses generated. The excerpts were selected to present sequential sets of critical information necessary to solve the mysteries. The group read the information one set at a time. After reading, individuals were asked what the basic problem was and what important information and clues were presented. The instructor listed these on the board until the set was exhausted then asked selected subjects to define the problem by relating all of the clues found. Subjects were then asked to generate as many hypotheses that would account for the information as possible. Each hypothesis was acknowledged and reinforced without being evaluated, much as in a brainstorming session. The next step was to read the subsequent excerpt, extract the important clues and details from it, redefine the problem in terms of the complete set of information and generate additional hypotheses. The group then compared the solution generated by Sherlock Holmes to the hypotheses generated by the group and noted the points of correspondence and difference. This proved to be a very effective technique of provid- ing feedback and reinforcing the generation of multiple explanatory hypotheses. This (second) session ended with the group having proceeded part- way through the second exercise of the type described above. They were asked to continue reading -- noting important details, redefining the problem and generating explanatory hypotheses. They were to return to the next session ready to discuss the problem. In addition they were given a third such exercise taken from "The Yellow Face" (pp. 353-358) to do overnight. 32 Bird Training Session The third and final training session began with a discussion of the homework problems. Individual subjects were asked what clues they considered important and what hypotheses they had. These were listed on the board and the group compared their hypotheses to Holmes' solu- tions. The major purpose of this session was to recombine the steps in the problem-solving analysis which were practiced in previous sessions; namely, problem sensing, problem definition and hypothesis generation, and complete the process through hypothesis selection, search and resolu- tion or redefinition. A mystery story was constructed to act as the vehicle for the group to practice these skills. The mystery was presented in booklet form, each page containing a single piece of information in the form of an interview, letter or document. Subjects read the first page, a one-paragraph description of the problem situation, and were told that their task as a group was to decide who committed the crime and why. All the available information had been gathered and compiled in their booklets but was available only upon request from the instructor. They were to request a source of information and the instructor would then tell them which page to consult. After they had read the new informa- tion the instructor asked the following questions: (1) Were there any new problems in the information? (2) What were the important details presented? (3) How did the new findings relate to the problem definition (problem redefinition)? (4) Was the hypothesis under test confirmed or disconfirmed? (5) Were there any new hypotheses? '1IIIllllIllllllllllllllllli======E::—— -u-iiw,i -iiiiiiiiimw, 33 (6) Which of the possible hypotheses should we test and where would you expect to find relevant information? The answer to the last question became the request for the next source of information. The instructor maintained a cumulative list of informa- tion, clues and hypotheses on the chalk board before the group. This was the final training exercise and was undertaken to integrate the skills practiced previously with practice in hypothesis selection, search and redefinition or resolution into a concentrated guided group experience in systematic problem-solving. Openness Training The openness training consisted of discussion, lecture and role- play aimed at teaching subjects the behavioral and attitudinal compon- ents of the model "open" person. The procedure was very value laden in an attempt to arouse a commitment to openness as well as teach the attributes of the concept (see Piper, 1969 for further elaboration). Control Training The control training consisted of highly structured (programmed) lecture and discussion about concept learning, generalization and dis- crimination, transfer, and suggestions as to how teachers could implement these principles and facilitate the learning of concepts by their own students. The material was taken directly from School Learning by Stephen Yelon (1969). The purpose of this training was to develop experimental involvement and thus control for the Hawthorne Effect. The subject matter was potentially interesting and valuable to pro- spective teachers yet considered to be enough removed from the experi- mental task (Teacher's In-basket) that there would be no transfer. C) .31 It (I) D 5“" 34 Post-Testing The fourth meeting was essentially the same for all subjects regard- less of training. Subjects met as a group, one to three days after training, for post-testing. There was no additional instruction at this meeting. Two post-tests, the Dogmatism scale and the problem-solving test (see Appendix B) were administered to all subjects. In addition, control subjects took a short examination on concept learning and in- struction, the topic of their training sessions. The Dogmatism scale was used as a post-test for the openness train- ing because an Open person should not be dogmatic. The Dogmatism scale was scored in three ways: (1) The simple algebraic sum, which is essentially a scale from "right" authoritarianism or dogmatism through "openness" to "left" authoritarianism; (2) The sum of the absolute deviations toward either left or right authoritarianism from an optimum score of "openness”; and (3) The number of items with which the sub- ject disagreed (agreeing with an item is considered a dogmatic response). The reader will recall that the problem-solving test included a statement of a problem and approximately twenty pages of information, indexed as a branching program, which was available to be used in Obtaining a solution (see Appendix B). In taking the problem-solving test, a subject was instructed to do the following: First, write down the number of each page as it was read; and second, when personally satisfied that she knew what the problem was, record her proposed answer on the sheet and stop working. Thus, two scores were generated from the Problem-solving test which were analogous to two scores (bits and com- PEtence) derived from the Teacher's In-basket. The first, Problem- sOlving Steps (P.S. Steps), was the total number of content pages con- SU1ted in solving the problem (analogous to bits). The second, Problem- SOLving Quality (P.S. Quality), was a rating of the solution offered aBainst a standard model (see Appendix B). 35 Competence was scored by matching the characteristics of the offered solution with a model solution. The score was simply the total number of matching characteristics. In order to avoid the possibility that personal characteristics of the experimenters would influence the sub- jects, both experimenters shared responsibility for all three types of training groups. The Transfer Task The basic purpose of this study was not to test the feasibility of training people in problem-solving techniques and strategies, but to test the effect of such training upon subjects' inquiry performance in real- istic, teacher-relevant situations. A major research concern was main- taining external validity, an approximation of real-world tasks in the experimental situation. As Rundquist et. a1. (1965) put it, "concept attainment tasks are much too simple for the purpose of answering applied questions regarding the efficacy of different kinds of extensive train- ing." (p. 26) Thus, in order to test the hypotheses it was necessary to maintain reality of task (external validity), and at the same time, experimental control (internal validity) over the stimulus inputs into the situation. Further, it was necessary that the cognitive processes of the inquirer be as accessible as possible to the observer in order for him to record the process of inquiry as well as the solution to the problem situation. To these ends the Teacher's In-basket was employed. The In-basket provided the realism necessary to maintain external validity and was composed of a specific set of materials into which problems were built to stimulate maximum inquiry. Role-playing was used to encourage a 36 degree of emotional involvement with the situation to further simulate a real teaching situation. Subjects were asked to think aloud in order to make their thoughts available to the observer. Thus, it was possible to determine what general information source the subject was accumulat- ing, whether the subject perceived the situation or information as problematic, and whether the information assisted the subject in resolv- ing a problem. The Teacher's In-basket consists of the complete set of current and past records on a fictitious class of fifteen sixth graders in Ridge Forrest Elementary School, Ridge Forrest, Michigan. The In-basket materials include memoranda, telephone messages, tasks and lists to be completed, test scores, schedules, etc. Finally, a school principal and "reference memory" are available for consultation over an intercom with programmed answers to anticipated questions. The subject was brought into a one-way observation room and told that this was her new classroom in Ridge Forrest School. She was seated behind the teacher's desk and asked to wear a small microphone. On the desk was an intercom, the in-basket, a folder containing current report cards, cardexes, attendance book and discipline anecdotes, and pencil and paper. She was then briefly introduced to the situation; that she was a new teacher taking over this sixth grade class in December after a succession of substitutes have been presiding for two months. There were many things to do while she was role-playing a teacher in this situation. She was asked to make the situation as real as possible and act as if this were indeed her class. Next she was told that the suc- cess of the research depended entirely on her ability to think out loud. Since the research interest was in following her thoughts throughout 37 the situation it was necessary for her to make all thoughts verbal, whether she deemed them relevant or not. After this rather brief verbal introduction the subject was given a short period of training in thinking aloud. She was given three line sketches of teachers in various situations and was asked to tell a story about each. The object was not to train the subject to tell stories but to get her verbalizing and she was told so. Following the stories, she was given a written description of the situation (see Appendix D) and was asked to read it aloud. Questions about the situa- tion itself were answered but requests for direct structuring of the task such as "But what should I do?" were met with the reply, ”Do what you would do if you were this teacher.” Subjects were then shown samples of all information sources and were instructed in the use of the intercom and reference memory. Reference memory was to provide the subject with information which she probably would know if she were the teacher, such as information about school policy, testing methods, test interpretation and general "scuttlebut." Because most of the subjects were taking their first education course and had no knowledge of testing methods or ability to interpret test scores, this function of reference memory was emphasized in an attempt to make up for a lack of content knowledge. It was mentioned that subjects had use of the cumulative records and medical records for each student. These contained information on achievement and aptitude tests, intelligence tests, attendance records, family data, health records, hearing and vision tests, immunization records and special help records from the time of the students' entrance into the school systenu Transfer records were also available as a part of the cumulative records. It was explained to the subject that 38 since these were usually found in the principal's office or in the guidance office filed with all students' records, their use of these records would be restricted to simulate this condition. They were placed on a table across the room from the subject and she was told to use them as she wished with the restriction that she not take the folder of records back to her desk but select the record(s) for the particular student(s) in whom she was interested. The material on her desk were for her unrestricted use; she could do anything except write on them. The current materials included progress reports for each student listing grades in academic subjects, citizenship and work habits, atten- dance, age, height and weight for the current year. Permanent elementary school records of achievement tests, intelligence tests, family data and significant comments for the current year were available on cardex file cards for each child. The teacher's record book contained an up- to-date attendance record for each child for the year plus eight discipline reports on selected children clipped to the record book. The subject was not restricted in her use of reference memory or the school principal as potential sources of information. On the desk was the in-basket containing memos and jobs which had been accumulating for lack of a permanent teacher. Imbedded within the in-basket were tasks, letters and memos designed to structure the situa- tion to the degree that all subjects would at least attempt to under- take the same basic set of tasks: Identify those students who must see the school psychologist for testing and interviews, make out a list of students to be referred for special enrichment or remediation tracks, etc. The purpose was to give the subject a set of tasks which would lead her into inquiry, which would require that she make use of student records and thus expose herself to potentially problematic information. "7“?— ' rv¢__’ 39 The choice of whether to continue inquiring or not was up to the subject but the in-basket tasks were constructed to expose her to a minimal amount of problematic information and allow the opportunity of continued inquiry if the subject so chose. The subject was observed in the above situation by a single observer who also functioned as reference memory and the school principal, and thus was in contact with the subject via intercom. The observer was separated from the subject by a one-way mirror. The "teacher's" desk was situated in such a way that the observer could see everything the subject was doing, what information she was reading and, even what she was writing. Everything the subject said was heard by the observer. The observer's job was to dictate a complete log of everything the subject did during her two to three hour stay in the situation. Everything the subject said, looked at, read, wrote, was recorded by the observer so that the log was, in fact, an accurate compressed time record of the subject's behavior. Interpretation of the subject's behavior was not attempted during observation. Scoring and interpretation were reserved until after all observations had been made. A number of steps were taken to reduce the possibility of observer bias: (8) The observer was not aware of the cognitive characteristics of the subject used in selection and blocking, nor was he (she) aware of the academic ability of the subject as indicated by his G.P.A. (b) In addition to the two people responsible for the group training, a third person was involved in the observation and every attempt was made to schedule subjects so that the person observing the subject was not the same person who had trained the subject. Although it was not possible to maintain a perfect "double blind," scheduling overlaps were such that the observer could not be certain of the training treatment 40 which the subjects had undergone. Finally, the method of recording afforded much greater accuracy in observation than had been possible in the earlier study (Shulman, Loupe and Piper, 1968). Dependent Variables A number of scores were abstracted from the observations in order to represent in quantitative form the inquiry behavior of the subject. A number of potential problems were imbedded in the in-basket and in the student records. They were of two basic types, the first being a simple violation of expectancy or deviation from the subjective "average" situation. An example of this type problem is the fact that Ronald Mayer, one of the characters developed in the in-basket, has straight A's on his report card. One simply doesn't expect a student to get such grades consistently, rather'the expectation for the "average" student is more likely to be that he get B's and C's, maybe even an occasional D. The second kind of problem is also a violation of expectancy. This expectancy is generated by the information itself, i.e., it is a conflict between sources of information. Ronald Mayer had straight A's on his report card but his permanent record indicates his last tested IQ was 91. One simply doesn't expect a student with an IQ of 91 to get straight A's. The experimenters and the writers of the Teacher's In-basket had arbitrarily identified some 250 such potential problems, potential be- cause a situation is not truly a problem until sensed. Although iden- tification of these problems was somewhat arbitrary in terms of average expectancies it was also quite exhaustive. On occasion a subject would sense a legitimate problem not formerly identified; these were carefully evaluated to see if they did in fact meet the criteria specified above “‘1‘!" ‘ ,_ 11 ,4’73‘: 3' Pro' I bl I . . ,A . ta . .. l.I.‘....I.l.. .. . 3 .m nu 9.» Ct AL. I Li S HUI 'i (A it. U .1 Out 41 for a potential problem. If a subject indicated through her behavior that she sensed a problem, it was scored as such. Thus the Problem Sensitivity Score was simply the total number of problems sensed by a particular subject. Since use of information is important to the process of inquiry and problem-solving, two measures of information usage were used. gigs was simply the number of times a subject consulted any source materials, including her own written notes, for information. For example, if a subject looked at a report card, a cumulative record, the sociogram, the same report card again, and the sociogram (in that sequence) this would be considered five "bits" even though he returned to two sources. This score than was an index of the total amount of information. Information sources, on the other hand, was a measure of diversity in information usage. For scoring purposes the Teacher's In-basket was divided into ten basic problem areas. A record was kept of the differ- ent sources consulted within each area. There were 12 sources categories: In-basket, sociogram, CTP summary, community maps, reference memory, report cards, cardex files, cumulative records, medical records, atten- dance book, discipline anecdotes, and newsletter. Thus, if a subject looked at a report card, a cumulative record, the sociogram, the same report card again, and the sociogram, the number of sources consulted is only three (if they all concern a particular problem). For any one problem area a source was tallied but once. Two scores were generated from the sources data: Total Sources, the total sources used across all ten problem categories; and Mean Sources, the total number of sources used divided by the number of problem areas into which the subject inquired. Because all subjects did not inquire into the complete set of problems, the two scores were not simple transformations of each other. 42 The only score calling for a qualitative interpretation of the sub- ject's performance was Competence. For each of the ten above-mentioned problem areas the problem resolution reached by the subject was judged as to the understanding of the problem it demonstrated. Model resolu- tions to each problem were written out at varying levels of complexity, the lowest of which was simple recognition of a contributing problem and the highest of which represented the fullest understanding of the problem that could be achieved with available materials (see Appendix C). These model solutions were then rated 1-5 according to their complexity and the effort necessary to gather appropriate information. The sub- ject's problem resolution was compared to this standard in scoring for competence. Her competence score on each problem was simply the scale value (1-5) of the model answer which most closely matched the character- istics of the offered resolution (see Appendix C). As a general index of inquiry performance the scores for problem sensitivity, mean sources and competence were standardized, and summed yielding the score called General Inquiry. Timg was also variable, since a subject remained in the situation until she called on the intercom and said she was done. Time varied considerably across subjects with a range of 90 to 180 minutes. The problem-solving training made no attempt to influence subjects to spend more time inquiring but simply attempted to improve their skills and make their efforts more effective. It is possible that such training would allow subjects to reach a tolerable resolution more quickly, thus they would take less time than if untrained. It is also possible that a factor which influences termination of inquiry is the frustration of finding one's self ineffective in work. If this is the case, training should increase time spent simply by making the subject more effective. 43 Thus two sets of scores were used, the raw scores as described above and ratios of the scores with time, to separate efficiency and effective- ness in inquiry from commitment. This same rationale was used to generate yet another score, M333 Competence. Because the training only attempted to increase competence of solution within a problem and did not attempt to influence subjects to make an added commitment to inquire, competence was averaged only for the general problem areas into which the subject inquired. The final score derived from the logged observations was more a reflection of the cognitive processing style of the subject than were the former measures. Operationally, Shifting was the total number of times the subject shifted his search from a bit of information in one source category to information in another (see page 41 for listing of sources), e.g., if a subject compared a report card, cardex, and the attendance book, this would be scored as two shifts, from report card to cardex and cardex to attendance book. However, if the same subject looked through the set of 15 report cards, no shifts would be involved. Shifting analysis was developed in the former study (Shulman, Loupe and Piper, 1968) in order to operationalize the theoretical description of the Dialectic. A major characteristic of the Dialectic as a style of argument or persuasion is that differing positions are developed, compared, juxtaposed and finally synthesized. Thus a Dialectical seeker should employ this style in her inquiry; she should do a lot of comparing and contrasting sources of information. This would tend to lead her into much shifting among information sources. In the Teacher's In-basket, information regarding any particular child was scattered throughout a number of sources and was never contained wholly within one source. However, it was possible to gather information 44 either by shifting or by progressing systematically through the information sources. Summary of Design and Methods Before the hypotheses are stated, it would be well to review and summarize the design and methods to clarify the tie between the statisti- cal hypotheses and the experiment itself. A two-stage experiment was developed to test the effects of particular types of training on subse- quent inquiry performance of sixty female subjects taking an introductory course in Educational Psychology. Subjects were selected in order to fit two theoretical types which were predicted on the basis of prior research to determine differences in inquiry performance and effective- ness. The types, Dialectical and Didactic, were operationalized in terms of a composite profile of four measures: Preference for com- plexity (Barron, 1967), liberality of political position, word associa- tion fluency, and degree of non-stereopathy (Inventory of Beliefs, Stern, Stern and Bloom, 1956). Subjects selected were randomly assigned to the six treatments which were combinations of two factors. The first factor of training had three levels: Openness training, problem-solving training, control training. The second treatment factor was presence or absence of a specific cue in the instructions during the second stage of the experiment. Following training, all subjects took a series of post-tests to judge the immediate effect of training. The second stage of the experiment involved placing each subject individually in a simulated teacher situation for a period of two to three hours and observing her inquiry behavior. A number of measures of inquiry performance were derived from the observations including: ‘7_ I 'Y I‘ m . . _ .. =7:- -.....~.-.r~.—1’__._—, . —‘:-vr':: - ' , - ,. __- .4...,_.-.....--:-ww 7 , — , -_....... ....,, .- , _ 45 1. number of Problems Sensed; 2. number of information Bits consulted; 3. number of Information Sources categories consulted; 4. number of Shifts made between source categories; 5. amount of Zimg_spent inquiring; 6. rated Competence of problem resolution. A number of scores were derived from these primary scores to equate for time spent and problem areas investigated. Hypotheses Hypotheses stated here are general to the total experiment or specific to the problem-solving training, those specific to the open- ness training are found in Piper (1969). l. The mean for the Dialectical group will exceed the mean for the Didactic group in all of the above primary measures of inquiry. This hypothesis is based on theoretical positions and empirical findings of Shulman, Loupe and Piper (1968) using a similar test composite to predict inquiry performance. 2. The mean for the problem-solving group will exceed the mean for the control group in the following measures of inquiry: problem sensitivity, competence, information sources, shifting. The reader will recall that the problem-solving training designed to give the subjects skills to enhance their inquiry ability. There was, how- ever, no direct attempt made in this training to develop an increased commitment to inquire. One question of interest will be to see if in- creasing subjects' skills in inquiry results in their spending an 46 increased amount of time in inquiry. However, this will not be put to planned statistical test as the experimenter has no theoretical justi- fication for predicting one result over the other. Thus, the above hypotheses will be tested in two ways: With the primary inquiry scores and with these scores adjusted to account for differential time expended and problem areas investigated. Problem sensing and shifting will be adjusted by dividing the scores by total time; competence and informa- tion sources will be divided by the number of problem areas investigated. 3. The mean for the cued group will not differ from the mean for the uncued group. 4. Cueing will interact with training such that the mean for the cued and trained groups will exceed the mean for the uncued, trained groups; however, the two un- trained groups will not be different from each other. The above two hypotheses simply express the author's feeling that the very brief and simple verbal instruction or cue was not adequate to pro- vide an enduring set or to recall a particular behavior pattern unless it was familiar and well-practiced as was the case with trained subjects. 5. Training will interact with seeking-style such that the difference in group means between trained and untrained Dialecticals will be less than between trained and untrained Didactics. The results of the experiment including tests of the above hypotheses are presented in Chapter 4 and interpretation of the results follows in Chapter 5. CHAPTER 4 RESULTS The basic purposes of this chapter are to inspect the data carefully, codify them, and organize them into a meaningful form which can be interpreted according to the purposes of this experiment. The following scheme will be used to organize and report the data in order to aid the reader in interpretation. First, the characteristics of the present sample will be described. The degree to which this sample replicated that of Shulman, Loupe and Piper will be discussed. Second, the effectiveness of the training in achieving improvements in problem- solving ability will be explored in terms of the problem-solving post-test. Third, the intercorrelations between the dependent inquiry variables measured in the transfer situation will be explored. Fourth, the experi- mental hypotheses will be tested and finally, the relationships between seeking-style, inquiry and intellectual measures such as CQT Total and G.P.A. will be explored. Characteristics of the Sample The importance of this particular section is two-fold. First, the results of the experiment will be more meaningful when the reader understands some of the psychometric characteristics of the sample for these would certainly have influenced how the subjects learned and responded in the experimental setting. Second, a very central part of this research design involves replication. The experimental hypotheses predicted replication of a number of findings regarding seeking-style and inquiry 47 48 reported by Shulman, Loupe and Piper (1968). Further, the inquiry behavior of groups fitting two carefully identified personality types was described in detail by Shulman, Loupe and Piper (1968). The present study set out to make predictions on the basis of those descriptions. The experimenter had a set of expectancies against which to compare the results. This allowed the possibility of considering non-replication of results as a function of non-replication in sample characteristics. A number of comparisons are made in this chapter with the results of Shulman, Loupe and Piper. These compari- sons are always with the results of the figsp administration of the Teacher's In-basket. TABLE 4. l MEANS AND STANDARD DEVIATIONS FOR DIALECTICAL AND DIDACTIC SEEKERS ON BELIEFS, COMPLEXITY, CQT TOTAL AND G.P.A. FOR THE PRESENT STUDY AND FOR THE STUDY BY SHULMAN, LOUPE AND PIPER* -————_~... Present Study Shulman, Loupe and Piper (1969) (1968) Dialectical Didactic Dialectical Didactic N=30 N=30 N=21 N=25 Beliefs Mean 68.00 52.07 65.57 54.72 S.D. 7.71 6.47 7.20 7.80 Complexity Mean 19.13 11.90 16.09 8.28 S.D. 4.24 3.85 4.22 2.25 GPA Mean 2.68 2.55 2.68 2.61 S.D. .56 .51 .48 .53 CQT Total Mean 131.60 127.63 133.76 123.96 S.D. 24.85 18.60 17.81 25.25 *Word Associations and Politics are not compared here because slight differences in scoring and item content make the raw distribu- tions non-comparable. — 4.-.... 7 49 Table 4.1 shows the means and standard deviations on the Inventory of Beliefs, Complexity, CQT Total and G.P.A. for the present sample and for that of Shulman, Loupe and Piper (1968). In the present sample Dialecticals and Didactics appear to be only slightly more differentiated with regard to Beliefs and G.P.A. than the former sample while at the same time less differentiated with regard to CQT Total. None of these differences approach significance. In general, the samples appear to be quite similar with regard to seeking-style and intellectual ability. However, it must be remem- bered that CQT Total is not a measure of current status. CQT Total was taken from the University Registrar's files and indicates one's score at the time of entgy to Michigan State University. The sample used by Shulman, Loupe and Piper consisted exclusively of college seniors in elementary education. The present sample consisted of sophomores interested in both elementary and secondary education. Thus, the sample was replicated psychometrically but differed in terms of college status and occupational interests. It has been demonstrated that the mean status on a number of var— iables was approximately the same in this sample and in the sample used by Shulman, Loupe and Piper (1968). The present group repli- cated the former group in measured entering characteristics. Another method of checking for replication would be to compare intercorrelations among the selection measures for the two groups. Although there is no absolute criterion against which to determine similarity of inter- correlation pattern, if the intercorrelations appear to be similar this is evidence that the two samples have similar characteristics. — l -“ _-'-Q.—--_." .0- ’_._.“~ ‘_ _ _, .c 50 TABLE 4.2 CORRELATIONS AMONG SEEKING-STYLE PREDICTOR VARIABLES IN THE PRESENT STUDY AND IN THAT or SHULMAN, LOUPE AND PIPER* Present Study Shulman, Loupe and Piper (1969), N=60 (1968), N=56 Beliefs l x _ l x Complexity 2 .62 x 2 .50 x Politics 3 .58 .29 x 3 .36 .23 x Word Association 4 .37 .36 .20 x 4 .36 .29 .39 x 1 2 3 4 1 2 3 4 *For a N of 60 the probability of a correlation = .25 occurring by chance =..05, for N = 56, the critical value of r .26. Table 4.2 shows the intercorrelations among the four seeking- style predictors for the present sample and for that of Shulman, Loupe and Piper. A high degree of intercorrelation existed in both samples among the four variables. The patterns of intercorrelation appear to be quite similar although the sample correlation values with Beliefs appears to be slightly higher in the present sample than in the former sample. In general, the replication of the former sample in the present study is striking, particularly considering three factors: (a) The Word Associations and Politics scales were modified slightly in the present study, (b) the samples differed with regard to college status and homo- geneity of major, and (c) the campus political climate has changed quite drastically in the past two years. These factors would tend to mitigate against replication; yet, in terms of tested entering char- acteristics, the two samples appeared quite similar. 51 Now that the replication of sample characteristics has been established, the focus will change to consider the effects of the training. The entering characteristics of the sample will be con— sidered again when the experimental hypotheses are formally tested. Immediate Effects of Training The reader will recall that the present study was a two stage experiment: the first stage being the training and the second being transfer of the learned skills to the Teacher's In- basket situation. Before any differences in performance on the Teacher's In-basket can be attributed to the training, it must first be established by immediate post-test that the training was effective in achieving the stated objectives. In order to determine the effect of the training, the problem-solving group was compared to the control group on the problem-solving post-test. Since the hypotheses stated that the problem-solving training would increase competence or quality of problem resolution, that measure (P.S. Quality) will be used for the comparison. TABLE 4. 3 MEANS AND STANDARD DEVIATIONS COMPARED FOR THE PROBLEM-SOLVING AND CONTROL GROUPS ON P.S. QUALITY Problem-Solving Control t p N=20 N=20 P.S. Quality Mean 4.45 3.45 2.04 .025* S.D. 1.67 1.43 *One-tailed test. 52 Table 4.3 indicates that the problem-solving training did, in fact, succeed in increasing inquiry competence within subjects so trained. The trained group differed significantly from the group given control training. Thus, any transfer effects of the training can be justifiably attributed to the effects of the training itself. Similarly, if there are no effects, it will be because the skills did not transfer, not because the training did not work. But, before the transfer effects of training are tested, it would be well to consider the characteristics of the dependent variables themselves. Characteristics of the Dependent Variables There were a number of measures of the inquiry process which were observed and recorded in the Teacher's In-basket. The variables measured aspects of the quantity of inquiry engaged in (Bits, Time), diversity of materials used (Sources, Mean Sources), amount of comparison between sources (Shifting), campleteness of inquiry (Competence), and the sensitivity to the discrepant (Problem Sensitivity). To ascertain that the dependent variables have essentially the same meanings as in prior studies (Shulman, Loupe and Piper, 1968), we can compare the means and standard deviations on the variables of the present study with those of Shulman, Loupe and Piper. Comparison of the means and standard deviations for four basic inquiry variables indicates that the Teacher's In-basket functioned basically the same way in the present study as in the past. Likewise, the correlations among the inquiry variables within the present study replicated those found in Shulman, Loupe and Piper (see above section on correlations.) The means and standard deviations varied little 53 TABLE 4.4 catPARISON OF MEANS AND STANDARD DEVIATIONS OF BASIC INQUIRY SCORES FOR SHULMAN, LOUPE AND PIPER (1968) AND THE PRESENT STUDY (1969)* Shulman Loupe and Piper Present Problems Mean 67.81 77.1 S.D. 17.48 16.75 Sources Mean 5.84 5.77 S.D. 1.09 1.17 Bits 'Mean 215.42 186.55 S.D. 48.16 50.88 Time Mean 120.00 129.35 S.D. 20.37 31.77 *Competence is not compared because of scaling differences in the two studies. despite two important differences in the studies: (a) The subject population in Shulman, Loupe and Piper (1968) consisted of senior elementary education majors; in the present study the population had a much broader base; (b) The recording methods were refined for the present study allowing more accurate scoring of most of the variables; (c) Subjects were trained for the purpose of changing inquiry performance in the present study. The procedures used in the present study prob- ably allowed for more accurate scoring of Problem Sensitivity while at the same time slightly underestimating Bits. Evidently, then, level of the dependent variables measured in the Teacher's In—basket remained relatively constant across samples. Thus, not only did the present sample replicate the selection characteristics 54 of the sample in Shulman, Loupe and Piper, but it also replicated the overall inquiry performance of the former sample. To aid the reader in interpreting the results presented later in this chapter and in Chapter 5, it would help to discuss the intercorrelations between the dependent variables measured in the Teacher's In-basket. TABLE 4.5 INTERCORRELATIONS AMONG DEPENDENT MEASURES OF INQUIRY, N=60* Bits l x Time 2 .54 x Total Sources 3 .65 .53 x Mean Sources 4 .56 .48 .84 x Shift 5 .90 .48 .75 .68 x Problems 6 .41 .41 .60 .52 .51 x Competence 7 .21 .35 .39 .29 .33 .63 x Mean Competence 8 .00 .20 .25 .38 .09 .37 .71 x *For N-60, the probability of a correlation .25 occurring by chance = .05. The correlations in Table 4.5 indicate that the dependent measures of inquiry were intercorrelated at a rather high and significant level. The high correlations among Shifts, Bits and Sources reflect the depen- dency between these measures in scoring. Bits placed the upper limit on Shifts and Sources defined the lower limit. If a person never con- sulted two consecutive pieces of information in the same source cate- gory Shifts would equal Bits. Likewise, if a person gathered all desired information in a particular problem from a source category and 55 never returned to it, Shifts would equal Sources. The other correlations indicate significant relationships between the process and outcome measures of inquiry with one notable exception. ‘Mean Competence, one's competence in resolving those problems into which he inquired, was sig- nificantly correlated with Mean Sources (.38) and Problem Sensitivity (.37). It was uncorrelated with Bits (.00) and Shifting (.09). Evidently Competence in problem resolution was related to the diversity of one's search primarily with regard to the problem under investigation. The amount of overall searching (Bits) gave no clue as to one's effectiveness in problem resolution (Competence). Table 4.5 also shows the correlations of Problem Sensitivity, Bits, and Shifts With Time. The correlations of the primary scores are very similar to the corresponding correlations found by Shulman, Loupe and Piper (1968). This indicates a rather high constancy of these parti- cular relationships over time and across subject samples. Problems, Bits, and Shifting were significantly related to the time spent in inquiry. The more time, the more problems sensed, the more bits of information used the more cognitive shifting in which the subject engaged. Finally, before the hypotheses are tested, it is necessary to define the relationship between the problem-solving test and the dependent inquiry measures of the Teacher's In-basket. To the degree that they measure the same things one would expect training to effect performance on the In-basket. Table 4.6 shows the correlations between the scores derived from the problem-solving test, P.S. Steps and P.S. Quality, and the inquiry variables of the Teacher's In-basket. The problem-solving test was a paper and pencil situation in which the subject was to take on the role 56 TABLE 4.6 CORRELATIONS BETWEEN SCORES ON THE PROBLEM-SOLVING TEST AND DEPENDENT INQUIRY‘YARIABLES, N=60* P.S. Steps P,S, Quality Bits .31 ..i6 Time .24 .48 Total Sources .36 .39 Mean Sources .33 .37 Shifts .31 .24 Problems .19 .33 Competence .28 .39 ‘Mean Competence .22 .33 General Inquiry .33 .45 *For N860, the probability of a correlation .25 occurring by chance 8 .05. of an employment counselor and find out why the man in the situation could not keep a job. The basic problem was stated. Information to solve the problem was available through a branching index system. P.S. Steps was the number of pages of information consulted and P.S. Quality was the rated completeness of the solution. All but three of the eighteen correlations between the problem-solving test and the In-basket measures are significant. Interestingly, P.S. Steps was most highly correlated with Bits, and Sources, measures of amount of inquiry. P.S. Quality was more correlated with Time, Competence and General Inquiry. In the problem-solving test the task.was well defined; in the Teacher's 57 In-basket, it was not. The problem-solving test included but one basic problem, the Teacher's In-basket included many. There was no direct content overlap between the two situations other than both were in the realm of social sciences. However, evidently the structural nature of the two tasks was similar; both required inquiry through use of various source materials in an unspecified sequence, with point of termination and amount of inquiry needed to be determined only by the participant. Thus, there was some reason to suspect that the dif- ferences between training groups found in the training post-test would be found in the transfer task as well. The next section will test the experimental hypotheses to see what effect training, seeking-style and cueing had on subsequent inquiry performance. Tests of the Hypotheses Before testing the hypotheses it would be well to review the nature of this particular experimental design. The research reported here is only part of a broader research project being carried out by this writer and Piper (1969). There were two training procedures, problem- solving and openness, to be compared with the control. In addition there were two other experimental factors, seeking-style and cueing, each of which had two levels. The comparisons between the two levels of seeking-style (Dialectical and Didactic) and the two levels of cueing (Cued or Non-cued) were germane to both studies. However, tests for transfer of training for individual treatments could only be made post-hoc. The hypotheses of the two studies overlapped with regard to the dependent variables and thus orthogonal planned comparisons were not possible. Hypotheses concerning training were tested by Analysis of'Variance for an overall training effect. In this section, each hypothesis will be tested separately. MEANS AND STANDARD 58 TABLE 4.7 DEVIATIONS OF INQUIRY VARIABLES AND POST-TEST SCORES FOR THE THREE TRAINING GROUPS Total Sources Mean Sources Shifts Problems Total Competence Mean Competence General Inquiry Problems/Time Bits/Time Shift/Time Control Problem-Solving Openness N=20 N=20 N=20 Mean 9.25 8.90 9.70 S.D. 5.30 3.82 5,23 Mean 3.45 4.45 4.20 S.D. 1.43 1.67 2.61 Mean 170.85 186.70 202.10 S.D. ~46.40 48.90 54.61 Mean 119.10 129.25 139.70 S.D. 29.37 24.78 37.91 Mean 53.40 58.75 55.90 S.D. 12.29 12.65 11.80 Mean 5.61 6.27 5.44 S.D. 1.22 1.07 1.11 Mean 113.25 137.10 137.70 S.D. 42.04 48.39 54.42 Mean 75.50 78.65 77.15 S.D. 16.61 14.90 19.19 Mean 25. 45 26.40 26. 70 S.D. 3.36 3.65 3.97 Mean 2.74 2.90 2.71 S.D. .26 .34 .39 Mean 144.90 153.90 151.45 S.D. 24.91 21.69 28.33 Mean .65 .63 .58 S.D. .15 .16 .14 Mean 1.47 1.46 1.52 S.D. .33 .37 .31 Mean .95 1.08 1.02 S.D. .25 .38 .29 59 Training We have observed that the problem-solving group was significantly higher than the control group on P.S. Quality, the competence measure of the problem-solving test. demonstrating that the training did have an immediate effect on those trained. Table 4.7 summarizes the scores on the problem-solving test and the Teacher's In-basket for the three training groups. The only contrast in which there was a clear effect for training was for Mean Sources (F - 2.98, p - .06). Thus, the training succeeded in increasing the diversity of the source base which subjects consulted during inquiry. This was expected. However, differences in Mean Competence, Problem Sensitivity and Shifting were also hypothesized. Closer inspection of Table 4.7 shows that, within this sample, the problem-solving group exceeded the control group on those meaSures and all other dependent measures of inquiry (however, it must be remembered that many of the differences were quite small). The interpretation is difficult-~the hypotheses were supported with regard to Mean Sources and formally rejected with regard to the other variables although a definite trend was observed. Seeking-Style Another of the hypotheses regarded seeking-style. Dialectical seekers were expected to surpass Didactic seekers in all measures of Inquiry. Table 4.8 lists the means and standard deviations necessary to test this hypothesis. Table 4.8 shows that, within the sample, Dialecticals surpassed Didactics in all measures of inquiry. However none of the differences on measures from the Teacher's In-basket were either psychologically or — _fl__,e_,._,m ,.___._ “.31.- ._ -_ 60 TABLE 4.8 MEANS AND STANDARD DEVIATIONS FOR DIALECTICAL AND DIDACTIC SEEKERS 0N SELECTED INQUIRY MEASURES Dialectical Didactic N=30 N-30 Problems Mean 78.10 76.10** S.D. 16.35 17.37 I‘ll ‘Mean 189.87 183.23 * S.D. 52.14 50.27 Time Mean 133.07 125.63** S.D. 35. 92 27. 15 Mean Sources Mean 5.84 5.7 * S.D. 1. 24 1. 11 Competence Mean 26.33 26.03 * S.D. 3.05 4.21 Shifts Mean 129.60 129.10 .S.D. 53.71 44.80 General Inquiry Mean 152.20 147.97 * S.D. 24.23 25.97 P.S. Quality Mean 4.46 3.60 S.D. 2.27 1.57 *This contrast was significant beyond alpha-.05 during one administration in the study by Shulman, Loupe and Piper (1968). **This contrast was significant beyond alpha=.05 during both administrations in the study by Shulman, Loupe and Piper (1968). statistically significant. Interestingly, Dialecticals performed sig— nificantly better on the problem-solving post-test (P.S. Quality) than did the Didactics. Thus, with regard to the problem-solving test, the hypothesis was correct. With regard to the Teacher's In-basket the hypothesis was rejected. 61 To describe the relationship between seeking-style and inquiry further we can look at the intercorrelations between the individual variables pooled into the seeking-style predictor and the inquiry var- iables themselves. use 4.9 . CORRELATImS 3mm SEEKING-STYLE PREDICTORS AND INQUIRY VARIABLES, N=60* Beliefs Complexity Politics Asszzigtion Bits .14 .14 .14 .19 Time .02 .13 .05 .14 Total Sources .12 -.04 .00 .35 Mean Sources .03 -.10 -.05 .26 Shifts .07 -.01 .07 .11 Problems .02 -.04 -.04 .14 Competence .02 -.02 -.05 .18 Mean Competence -.04 -.08 -.13 .13 General Inquiry .07 -.02 -.04 .26 *For N-60 the probability of a correlation .25 occurring by chance - .05. Table 4.9 shows that the only significant correlations between the seeking-style predictors and inquiry are with Word Associations. This, of course, is counter to the findings of Shulman, Loupe and Piper. Chapter 5 will address this discrepancy and attempt to explain it in terms of sample characteristics. The next section will be concerned with the effect of cueing and interactions. Cueing and Interactions _.,__: . .. I ~— ~.-..—- 62 TABLE 4.10 '-.. .— 7‘... MEANS AND STANDARD DEVIATIONS OF INQUIRY VARIABLES FOR THE CUED AND NON-CUED GROUPS Bits Time Total Sources Mean Sources Shifts Problems Competence Mean Competence General Inquiry Bits/Time Shifts/Time Problems/Time Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean Cued N=30 184.77 45.08 126.73 33.74 55.13 11.08 5.75 1.15 127.27 44.20 76.30 17.59 26.10 3.76 2.81 .38 148.77 25.36 1.53 .36 1.05 .35 .64 .17 Non-cued N=30 188.33 56. 83 131.97 30.06 56.90 13.44 5.79 1.21 131.43 54.20 77.90 16.13 26.27 3.59 2.75 .29 151.40 24.98 1.44 .30 .98 .28 .61 .14 63 Table 4.10 presents the means and standard deviations necessary to test the third hypothesis that the means for the cued grour>will not exceed those of the non-cued group against the alternative that the cued group will exceed the non-cued group. As can be seen from the table, the means for the two groups did not differ. In fact, they were almost identical. If any trend is to be seen, it is that the cued group scored lower on the basic inquiry measures and moved at a quicker tempo. However, the sample differences are very small indeed. These results tend to confirm the present writer's feeling that a simple cue alone would be insufficient to influence one's performance in a situation as complex and open as the Teacher's In-basket. Although no effects for cueing were expected, the fourth hypothesis predicted that cueing would interact with training such that the means for those groups both cued and trained would exceed those trained but not cued. No such difference was expected within the control-trained groups. TABLE 4.11 COMPARISON OF MEANS FOR THE TREATMENT X CUEING INTERACTION* Cued Non-Cued Problem Problem Control Solving Open Control Solving Open N-10 N-lO N810 N=10 N-IO N=10 General Inquiry 149.7 158.6 138.0 140.1 149.2 164.9 Problems 80.4 81.1 67.4 70.6 76.2 86.9 Mean Sources 5.74 6.57 4.95 5.47 5.98 5.92 *In considering the interactions, means will be presented only for those interactions significant beyond alpha-.10. 64 Analysis of variance showed that there were statistically sig- nificant interactions between training and cueing for three inquiry variables, General Inquiry (F = 3.47, p = .04), Problem Sensitivity (F . 4.50, p = .02) and Mean Sources (F = 2.56, p = .09). However, the nature of the interaction was contrary to that stated in the hypo- thesis. Comparison of the means in Table 4.11 shows that cueing facili- tated performance of the control group and the problem-solving group. This is contrary to the author's expectation but is consonant with other experimental literature on cueing in less complex situations (see Colgrove, 1968). However, inquiry performance of the cued, openness group was considerably lower than that of the non-cued openness group. TABLE 4.12 MEANS FOR SELECTION X CUEING INTERACTION Dialectical Didactic Cued Non-Cued Cued Non-Cued N-15 N-15 N=15 N=15 Bit/Time 1.42 1.55 1.59 1.37 Shift/Time .91 1.07 1.15 .94 The final hypothesis that training will interact with seeking- style was not supported. None of the contrasts were significant. One additional result which does not relate to the experimental hypotheses involves the interaction of selection x cueing. Table 4.12. shows the means for Bits/Time (F - 3.10, p - .08) and Shifts/Time (F = 4.09, p = .05). The net affect of this interaction was that cueing tended to slow the inquiry tempo (the rate at which the inquirer used _.._v~-1_,’ mg a, -‘ 0g," a“ .._ 1' _ . ,, 65 information, shifted, etc.) for Dialectical seekers and increase the tempo for Didactic seekers. There were very few differences noted in the results and yet many were expected. Before the results are summarized it might be well to look at the relationships between CQT, G.P.A. and inquiry to see if this might provide grounds for interpretation of the results in the next chapter. CQT, G.P.A. and Inquiry TABLE 4.13 CORRELATIONS BETWEEN CQT TOTAL, COLLEGE GRADE POINT AVERAGE AND TEE DEPENDENT INQUIRY VARIABLES, N=60 CQT Total Qagaé; Bits .10 -.05 Time -.07 -.09 Total Sources .25 .22 Mean Sources .28 .28 Shifts .04 -.02 Problems .31 .24 Competence .14 .15 Mean Competence .18 .24 General Inquiry .28 .24 Table 4.13 shows the correlations of inquiry with CQT Total and G.P.A. The correlations between G.P.A. and mean sources (.28), pro- blem sensitivity (.24), mean competence (.24) and total sources (.22) indicate that G.P.A. was more related to quality of inquiry within ‘ '5‘"..- “32"?“3‘7: 66 a problem than overall commitment to inquire. Interestingly, CQT Verbal and Numerical scores correlated with different inquiry measures. CQT Verbal correlated significantly with Problem Sensitivity (.29) and Competence (.37). CQT Numerical correlated significantly with Mean Sources (.28) and Problem Sensitivity (.31). Overall General Inquiry correlated .22 with CQT Verbal, .25 with CQT Numerical, and .28 with CQT Total. Within the present study it appears that there was more relationship of inquiry with intellectual ability than with seeking-style. However, although significant, these correlations were very small indeed. SummaEy (1) Results of the experiment demenstrated consistency of the Teacher's In-basket as a research instrument. Means of inquiry measures and their intercorrelations remained stable across subject pools and years. Likewise, the concept of seeking-style types was confirmed by replication of the inter- correlations between seeking-style components. (2) However, seeking-style was not a significant determinant of inquiry performance in the present study. The differences between Dialectical and Didactic group means were in the pre- dicted direction, but the magnitude was not sufficient to make them reliable. (3) The training groups differed significantly only in Mean Sources. However, in this sample the problem-solving group differed from the control group on all measures of inquiry and all differences were in the predicted direction, 67 although the differences were not large enough to justify generalizing. (4) Cued subjects did not differ significantly from non-cued subjects in overall inquiry performance. (5) Cueing interacted with training such that it tended to de- crease inquiry performance of those trained in openness. (6) Selection interacted with cueing such that inquiry tempo (Bits/Time, Shifts/Time) was increased for Didactics and decreased for Dialecticals. In Chapter 5 the author will look more closely at some of these findings and attempt to explain and interpret the results of this experiment. CHAPTER 5 DISCUSSION OF RESULTS One of the basic premises of this research was that inquiry effectiveness was a function of two very different factors; commit- ment to inquire, which determined the quantity of inquiry in which one engaged (Time, Bits); and skill or effectiveness once engaged in inquiry (Mean Sources, Mean Competence, Shifting). Commitment to inquire was thought to be determined by personality characteristics such as openness and preference for the complex. Thus, the person- ality composite was called seeking-style predisposition. Skill or effectiveness in inquiry was thought to be related to the skills of information organization and search variously connected with problem- solving, scientific inquiry, etc. Both of these factors (commitment and effectiveness) were felt to contribute to the overall outcome of inquiry; namely, competence of problem resolution. This experiment also distinguished between entering characteristics and newly learned skills and attitudes. Commitment to inquire considered as an entering characteristic, was measured by using a composite of personality measures collectively called seeking-style. It was mani- pulated as a learned set of predispositions by Piper (1969) through the use of training in openness. Skill or effectiveness in inquiry was manipulated in the present study through training in problem-solving, but in neither study was skill in inquiry explicitly considered as an entering characteristic. 68 69 This distinction between entering characteristics and new learning is of prime importance when considering the results. Inquiry performance measured in the Teacher's In-basket was explicitly considered a function of commitment to inquire and skill in inquiry. Both facets of commit- ment (entering characteristic and new learning) were considered. Only one facet of inquiry skill was considered: the newly learned strategy. The other facet of inquiry skill, knowledge and skill possessed by the subject at the beginning of the study, was ignored and was presumed to be randomly distributed. However, the results indicate that these enter- ing skills may have had a major effect on subject's performance. We will attempt to reconstruct the group nature of these skills later in this chapter (particularly with reference to seeking-style). However, the principle emphasis of this chapter will concern commitment as an entering characteristic (seeking-style) and the newly learned strategies of problem-solving. A third factor in the experiment of less central interest was cue- ing. It was thought that cueing might aid the subject in transferring newly learned abilities to the Teacher's In-basket and thus aid the experimenter in distinguishing between inquiry performance and capability. With cueing, performance would presumably be closer to the level of capability than without. Results showed no overall main effect for cueing; thus, it will be discussed only in terms of its interactions with entering characteristics and new learning. Predicted Commitment: Seeking Style One of the basic hypotheses of the study was that Dialecticals would be more effective inquirers than Didactics and that this would be due largely to a greater commitment to inquire on the part of the 70 Dialecticals. It was thought that this commitment would be reflected in increased Time and greater inquiry effectiveness. However, the means of the Dialectical and Didactic groups differed very little on any of the inquiry variables. Dialecticals spent an average of seven min- utes longer in inquiry than Didactics. In the research by Shulman, Loupe and Piper (1968) this difference was more on the order of twenty minutes, or 2 1/2 times the present differential. Dialecticals in the study by Shulman, Loupe and Piper (1968) were Significantly better inquirers during the first administration of the Teacher's In-basket in terms of.Prob1ems sensed and Time Spent. In that study it was not until the second administration that the Dialecticals were clearly superior inquirers to the Didactics on the other inquiry criterion measures. How- ever, the present sample of Dialectical seekers did not even exceed the Didactics in Time, the most clearcut indication of commitment to inquire. The problem is to consider a number of alternative explanations for the discrepancy. The first possibility is that differences in intelligence or aptitude in favor of the Didactic group may have confounded the findings. The results did show a small, but significant, relationship between CQT Total, G.P.A. and inquiry. However, there were no differences between Dialecticals and Didactics with regard to those measures. Another alternative to consider is that the effect of cueing may have interacted with seeking-style to mask the effects of seeking-style. Cueing did, in fact, interact with seeking-style and in a disordinal manner. The effect of cueing on Dialectical subjects was precisely the opposite of its effect on Didactic subjects. This is most apparent on the measures of inquiry tempo. 71 Recall that tempo was the rate at which one inquired while in the situation. Tempo measures were derived by simply dividing Time into another inquiry measure such as Bits or Sources (Bits/Time, Sources/Time). The resultant ratio described the number of Bits, Sources, etc. per min- ute used by the inquirer and thus gave an indication of the pace at which she inquired as opposed to her overall amount or quality of inquiry. __—_-. Bit-Tempo Shift-Tempo 1.60 r 1.15 - Dial. 1.55 h 1.10 I— Dial. 1.50 . 1.05 t 1.45 — 1.00 - Did. id- ]..40 '- .95 I’ 1. 35 L’ 1 090 '- L I Cued Non-cued Cued Non-cued Figure 2. Mean Scores for Dialectical and Didactic, Cued and Non-Cued Groups on Bit and Shifting Tempo Figure 2 shows that for Dialectical subjects, cueing had the effect of slowing inquiry tempo and for Didactics it had the opposite effect of increasing tempo. Both of these interactions were significant (Bits/Time, F = 3.10, p = .08; Shifts/Time, F = 4.08, p = .05, see Tables D-6 and D-7). Cueing may have had this strange effect because the cue itself was too ambiguous or, because when cued to behave in a rather undefined manner, one assumes that this requires a basic change in behavioral style, away from one's preferred style. This interaction can also be explained by inspecting the CQT scores. Figure 3 shows the mean CQT Total scores for Dialectical and Didactic, cued and non-cued groups. Notice that the pattern is exactly the oppo- site from that of the tempo scores. It could be that intellect has the ..'., _ _ -, __ r — -_ any—u- lL’f: 72 effect of slowing one's tempo thus allowing careful consideration of information. This is borne out to some degree by the correlation be- tween CQT Total and Bit tempo (r=.31). 135 r Did. Ial. 130 P 125 ~ 120 I I Cued Non-cued Figure 33. Mean CQT Total Scores for Dialectical and Didactic, Cued and Non-Cued Groups However, as we have already seen, CQT or G.P.A. do not supply legitimate explanations for the findings of no differences between Dialecticals and Didactics, nor do interactions or sampling errors. In Chapter 4 we established very clearly that the present sample was a close replication of the sample used by Shulman, Loupe and Piper in which Dialecticals and Didactics did differ in commitment to inquire (Administrations I, II) and competence of inquiry (Administration 11 only). .Yet the present sample just as clearly did not replicate those findings. The only entering characteristics not replicated by this sample were status measures such as age, number of years in college, college major, etc. Presumably, such differences in Status would indicate differences in status and variability on skills which would greatly influence ability to inquire effectively in the Teacher's In-basket situation. In particu- lar, skill in problem-solving and familiarity with the late elementary 73 educational setting were probably different in the two samples. The members of the present sample, for the most part, were beginning students in education considering education as a professional choice. They cer- tainly were not as committed to their choice in terms of time spent and energy exerted as the members of the previous sample. Also, their inter- ests in education were not necessarily with elementary aged children. It is fair to assume that a group homogeneous in its commitment to elementary education would probably have more knowledge about the nature of children's problems at that age than a group whose members were not so homogeneous in interest. Another facet of non-replication involves specific knowledge about the In-basket. Students in their final year of teacher preparation would, most likely, have some knowledge about testing procedures and would be able to interpret simple normed testing scores. Likewise, they would probably have some average expectancies regarding social interactions, physical appearance and academic abilities of an hypothetical group of students. It is doubtful to this writer that a group of college sophomores in the introductory course in education would possess such sophistication. It is very possible that in a situation as complex as the Teacher's In-basket, supplementary stores of information such as Reference Memory are simply not sufficient to make up for a deficit in knowledge. With- out a minimal amount of this particular knowledge it is doubtful that one could even see a problematic situation. As such, commitment to in- quire and perceptual sensitivity to the discrepant would be irrelevant. Seeking-style is a viable dimension only when working in a realm with which all are familiar. This generalization is supported by the finding that Dialecticals were superior to Didactics in P.S. Quality, a measure 74 of competence on the problem-solving test (see Table D-2, F=2.84, p=.10). In a situation in which all had sufficient entry skills to operate adequately, seeking-style was a relevant dimension. Actual Commitment: Time in Inquiry Prediction of commitment to inquire via the seeking-style dimension did not meet with notable success in the present study although there appeared to be at least one cogent explanation for the lack of results. However, we have not yet addressed a more central question: To what extent does actual commitment to inquire influence inquiry performance? Time spent in inquiry will be used as an indication of actual commitment to inquire. The reader will recall that time spent in inquiry was correlated with other inquiry variables, such as Bits, which tended to reflect overall amount of inquiry. Results showed that these remaining in the situation longer also chose to inquire into more problem areas. Time indicated overall amounts of inquiry. However, there were also qualita- tive differences in inquiry which were related to time spent. The correlations between Time and the ratio of the primary measures with Time disclosed a pattern of tempo which may be a characteristic of the process of inquiry. Table 4.1 shows these correlations. Notice the high and significant negative correlations between the Bits and Shifts ratios and Time. This pattern can be interpreted in two ways. First, there may have been a ceiling effect such that once a person spent a certain amount of time in the situation she had used all of the available bits. Second, those that took more time in the In-basket may simply have moved at a slower pace. The second alternative is supported by the -.37 correlation between Shifts Tempo and Time. Clearly, there is no ceiling for shifting. Shifting is simply a style as. i 75 TABLE 5.1 COMPARISON OF THE CORRELATIONS OF INQUIRY VARIABLES AND TEMPO VARIABLES WITH TIME, N=60* _T_1_g_1§_ Bits .54 F Shifts .47 Problems .41 ' BitS/Time -.56 Shifts/Time -.37 3 Problems/Time -.12 *For N=60, the probability of a correlation of .25 occurring by chance=.05 (two-tailed test). measure and gives no direct index of quantity of information used. Notice, however, the correlations between Problem Sensing tempo and Time (-.12). There is no systematic way in.which overall time spent influences one's tempo in sensing problems. Those spending more time in the In-basket moved at a slower pace and yet sensed more problems. To borrow an analogy from Shulman, the film speed (intelligence) of the cameras (inquirers) may have been the same, but the light was rather dim (structure of the task), so a slowing of the shutter produced a better picture. ‘Moving quickly and spending little time may have proved an inefficient strategy because it did not allow sufficient exposure to sense problems. Thus, commitment to inquire (Time) had its effect on inquiry performance (increased Sources, Shifts, Problem Sensed, and Competence increased with Time) not only through a 8irnple increase in Time spent. Comitment to inquire seemed to determine ‘Ilnalitative aspects of the inquiry process. Those spending a long time 76 in inquiry were slower moving but this slow tempo evidently enabled them to make more effective use of the information available. Thus, they sensed more problems and resolved them more competently. Changes in commitment to inquire as a function of training and the effects of such changes on subsequent inquiry behavior are discussed in full by Piper (1969). The basic findings of that research indicate that short-term training can affect changes in commitment. Subjects so trained spent more time and used more bits of information during inquiry than subjects not given training. However, these changes in commitment were not reflected by concomitant changes in inquiry competence. This, as well, may be due to a masking of effects by large variation in sub- jects' entering characteristics. That a change in commitment to inquire can be induced was demonstrated; that such a change has an effect on the outcomes of inquiry was not. Skill in Inquiry: New Learning The other phase of the present study involved training subjects in problem-solving in order to improve the effectiveness of their inquiry efforts. As was mentioned in the introduction and review of literature, the characteristics of the problem-solving training were derived from a number of sources to insure generality and transfer from the training to testing situations. Also, there were two stages to the experiment; the first to establish if the training procedure worked and the second to check for transfer of the skills into the Teacher's In-basket situation. In Chapter 4, the results were analyzed by analysis of variance to test the main experimental hypotheses. Additional analyses presented :ln this chapter to compare the problem-solving group to the control will make use of _t_:_ tests. Because there were two training treatments to be A! us - 77 compared with one control, all comparisons of interest were not possible within an orthogonal set of planned comparisons. Therefore, the £_tests presented in this chapter comparing the groups on the Teacher's In-basket were done only for exploratory purposes and cannot be considered fully legitimate tests of hypotheses. Because they are not orthogonal with a similar set of F tests presented by Piper (1969) the probability levels reported are adjusted to 2 1: alpha. Thus, the actual a‘lpha level of the test is less than or equal to that reported. To review the hypotheses, the problem-solving group was expected to surpass the control group in P.S. Quality (the problem-solving post-test), Mean Sources, Shifts, Prob- lems and Mean Competence. TABLE 5.2 MEANS AND STANDARD DEVIATIONS ON SELECTED VARIABLES COMPARED FOR.THE PROBLEM-SOLVING AND CONTROLS GROUPS** Problem-solving Control t p N=20 N=20 _ _ P.S. Quality Mean 4.45 3.45 2.04 .025 S.D. 1.67 1.43 Mean Sources Mean 6.27 5.61 1.82 .10‘* S.D. 1. 07 1 22 Shifts Mean 137.10 113.25 1.66 .20 * S.D. 48.39 42.04 Problems Mean 78.65 75.50 .63 .80 * S.D. 14.90 16.61 Mean Competence Mean 2.90 2.74 4.52 .002* S.D. .34 .26 General Inquiry ‘Mean 153.90 144.90 1.30 .20 * S.D. 21.69 24.91 '*Because these tests are not orthogonal with a similar set of ¢=cH pcooapm Him mqm<9 150 151 AAA ANH AH AN HA AAN AA HAH AAH AA AH :: AAN AA HAH ANH AH AN AA AAN AA AAH AAH HH HH HA AAA AA AHH AHH AH AH AA HAN AA AAH ANH AA AA NA AAN AA AHH AHH AH HN AA ANN A AAA ANH AH HN AA AAA A AAH AHH AH AH NA AAN NA HAH NAA NH NH AA AAN HA HAH ANH NH AH AA AAN AA NAH AHH NH AA A AAN A: AAH NAA NH AH A ANN A: AAH AAH AH AA AA AAH A: ANH AHH AH AN AA HAA A: AAH HHH AA NH AA AAA A: ANH NAA AA AH AA AHA A AAH AAA AH AH AA AHA : AAA AAH AH HN AA AAN NA NAH AAH AH AA A: :HN H: AAH AAH AH AN AA AAA AA AAA AAH AH AH AA AAA AA HAH AAH AH AN AA AAN AA HNH ANH AH HN AA AAN AA AAH AHH AH AN AA AAN AA AAH HAH AH AH :A AAN AA «AH AHH AH AH NA AHN AA AAH AAH AA AH AA ANN AA AAH AHH AH AH AA AAH NA AAH HAH NH AN NA. AAA HA 3mm coHpmHooAA< AoHpHHom AAonamaoo mmoHHom mo 4&0 .aonsaz SAHAAsmoQ Ago: Asapco>cH acocaum dam mamHom wcH>Hom unocaum amanoam soanopm Nam mqmde in. AAA HAH AAH A.N HN AAA AA NAA AAA AAH. A.N AN HAA AN AAA AAH HAH A.N AN NAH N AAA AAH AAH A.N AN AAA AN AAA HAA AAH A.N AN AAA AN AAA AAH AAH H.A HA AAH AN HAA AAH ANN A.N AN AAA N AAA AAH AAH A.N AN AAH AN HAA HNH AAH A.N AN AAA NN AAA HHH AAH H.A AN NAH HN AAA AAH AAH H.A HA AAA AN HAA AAA AAA A.A AA AAA AH AAA AAH AHN A.N HN AAA AH % NAA AAA HAH .A.A AA AAA AH 1 AAA HAA HAH A.N AN AAA AH AAA AAA AAH A.N AN NAA AH NAA AAH AAH A.N AN AAA H AAA AHH NAH A.A AA HAA H AAA AAA ANH A.A HN NAA NH AAA AAA AAH A.N AN AAA HH NAA AAA AAH A.N AN HAA AH AAA AAH HAH A.N AN HAo A AAA AAA AAA H.A HA AAA A AAH AAA NAH A.N AN AAA A NAA AAA AAA A.N AN AAA A AAA AAA AAH A.N AN AAA A HAA AAH AAH A.N AN AAA A AAA AAA ANH A.A AA AHH A AAA HAA AAA A.N AN AAA N NAA HHH AAH A.N AN AAA H esHe osHA oaHe oosopommwo ommepoasoo AAAMAAHAcom panama \AAHAHAHAAAA \AAAHAA \AAHA emu: soHpoAA. AeoAAAA emanoam A-A wands an, 1: ——_——_——————-——=_— _.._...=__._—.— ——v.-_—_—.—_ 155 AAA AAA NHH A.N AN NAA AA NAA AAA NAH A.A AA AAA AA AAA AAA AAH A.N HN AAA AA AAA NAA AHH A.N AN AAA AA AAA NAA HAH A.N AN AAA AA AAA NAA AAH A.N AN AAA AA AAA HAA AAH A.A AA AAA AA HAA HAH AHN A.N AN AAA AA AAA AAH HNN A.N AN AAA NA AAA NAH AAH N.A NA AAA HA AAA ANH AAH A.N AN AAA AA AAA AAA AAH H.A NN AAA AA AAA AAA AAH A.H AH AAA AA AAA HAH HAH A.N HN AAA AA AAA ANH HAH A.N AN HAH AA NAA HAA AAH A.N HN AAA AA AAA AAH AAH A.N AN AAA A NAA AAH HAH A.N AN HAA A AAA AAA AAH A.N AN AAA NA HAA HAH AAH H.N AH AAA HA AAA AAH AAH A.A AA AAA A AAA HAA AAH A.N HN AAA A AAA NAH AAH H.A HA AAA AA AAA AAA AAH H.A AN NAA AA AAA AAH AAH A.A AA NAA AA AAA AAA AAH A.N AN AAA AA AAA NAA AAH A.N AN AAA A AAA AAA AAH A.N AN AAA A AAA ANA AAA A.N AN AAA NA AAA A A AAH A.N AN AAH HA 05am. 08.3. 08.3. vofikooflu—OO ova—0909960 Hudbduamflmm 900.952 \AAH>HAHAAAA \AAAHAA \AAHA awe: aoHpoam unoozum Emanoam Aim wands 3!"...9. HHH AN AA AA AN ANH AA AAH AH AA AA AA HAH AN AAH NA AA AA NA AAH AN ANH AH A NA HA AAH AN AHH AN A NA AN AAH AN AAH AA AA AA A AAH AN A A H AA AA NA A HAH AN ANH AN AA AA NA AAH AN AHH AA AA AA AA HAH NN AAH AA HA NA HA NAH HN HAH AA HA AA AN AAH AN 6 ANH AH A AA AA HAH AH 5 AAA AN A . AA AN AHH AH HNH AN HA AA AA AAH AH HAH AA AA AA AA NAH AH AAA AH AA AN AN AAH AH AAA .. AN NA AA AH HAH AH NAH AA HA A AA AAH AH AAH A A A AA AHH NH AAH A A AA HN ANH HH AAH HN AA AA A ANH AH AHH AN NA AA A A H A ANH AH AA AA AA AH A AAH AN AA NA AA AH A AAH AN AA AA AN HNH A HAH AN AA NA AA NAH A NNH HA A HA NN AAH m AAH AN A AA AA AAH NAH NN A . AA AA AH N AAH AA H . AA AA AH H HAAAB HmoHpossz noHpmaAowcH Hann0> igacdom ”.35qu 539352 AAA AAA AAA Ado Am: Hwnocoo pacesum 30m mam<8 HAH A AA AA AN HAH AA AAH AN AA AA NA AAH AA AHH AN AA AA AN AHH AA NAH AN AA AA AN AAH AA AHH HN AA AA HA AAH AA _ AAH NA AA NA AA HAH AA A AAH AH AA HA AN HAH AA A AAH AN HA AA AA AAH AA AHH AN NA AA AN AAH NA AAH , AA AA AA AA AAH HA NAH NN AA AA HA AAH AA 7 AAH AN AA AA AA AAH AA 5 ANH AN A A AA HN AAA AA ANH HN A HA AN AHH AA AAH AH AA AA AA NAH AA AAH AA NA AA AA ANH AA HAH AN AA NA AA AAH AA ANH . AN AA AA AA AAH AA AAH AA AA AA AA AHH NA NNH AN HA AA AN AAA HA AAH AA AA AA NA AAH A AHH AH A NA HA AAH A AAH AH A HA AA AAH AA AAH AA AA HA A AAH AA ANH AN AA AA H HAH AA HNH HN AA AA AA AAH AA AHH HA HA AA AN ANH AA AAA HN AA NA AH ANH AA AAH AA AA A AN A H NA AAH AA NA . A AA AAH HA Hmpoe HAoHpmesz [moHpAsnougH amnmo> mmHummm NuHsde 9098:: AAA AAA AAA AAA AA: HauocoA pcocspm aim ”Huge 158 AN AN AN AN NA AA AA AN AN HH NA HA AN NN HN AN AA AA AA AH AA AA AA AH AH AA AA NA AH AH AH HA HA AH AH NH HH AH AA AA AH A A A A AH AA AA A AH AA HA A AN AA AA A N H oAAAAA HH onHHoA w,onAmmm panama AmoHHom unconnm A.A AAAAA Ho on an AA AA AA AA AA HA Md “\0 (\oo 00 \nvnnxnunnxnxo NA AA AA AA HA AA HA AA AA 159 AH AA AA AA AA AA AA NA NA AA AA AA AA AA HA A vwcmso HH mmoHaom H mmoHHum pmnasz AAoHaom . . acoASAm A-A AAAAA u I I n“ I?- \ 5V 160 AAAA Aoan oosono n wchso AHAAAAHAHH AHAAA HAoHAoAHAHAnA - AAHAAAA Ammccmaoum AAH>HAA eoHpopAnH Hoppnoouo u mchHmnB AAA AAHAAAHAHAAAAH pomnnsm o H o o H o 0N H o H mN o H 0 AN H H H 0N H o N mN o H H N H H N N o o H NN o o H HN A H H AN 0 o N 0H 0 H N NH O o 0 AH H o N 0H H o o mH o H N H o H H H o o H NH H o H HH H H N OH H o o m H H H A H H o A H _H o A H o N m H H N m o H N o H o N H H o H meozo omhum \mcHnHApe bones: AAonom AcoAAAA A-A AAAAA 0|!qu u.~..~vH<.~. 1 . . 161 ooze pozuH omsouo - mchso oHpomoHnuH mHHpm HacHuoeHmHQuo - mconom mmmccomoum mzH>Hom smAnoumnH Aoaucoono u wchHwne |hfl‘{u!u.m.... . an...“ s... h ham CoApmoAmApcocH pomfinsm OHOOHHOHHHHOOOHHOOHOHOOOHOHOHH OOHOOOOOOOHOHHHOHHHOHOOOHOHOHO OHHNOHONOOONONHHfiNq—GHNHONOONNHN NM: “\ONQOO oarutvvocxabowaea -3:rdm3:rdu3:iUNVVAU\VHAUqunuvo mmllwo r! a) w-m mgmde wwwmm wconmm mecmee TRTE UNIV. LIBRQ 11111 111 I11“ 3101561789 HICHIGRN s “WWI" 3129