THE PAEJQMQRPHZC REPRESBWTATEW 0F TEACHER 353% EC-N ééAé-fiENG AS A PREDéCTOR OF iNQUKRY PERFGRFéAfiCE Thesis for fihe Degree of Pin. D.‘ Mlflfiifim SMTE UNWERSITY FéERLYN MILDRED MONDOL ‘ 1973 ‘ . 5 illlllllllHIllllllllllllllllllllill\Hllllllllllllllllllllll - -- _ 3 1293 10396 6135 ‘7 1 Michigan State #1 University This is to certify that the 3 3 “'THE PARAMORPHIO REPRESENTATION OF TEACHER DECISION MAKING As A PREDICTOR OF INQUIRY PERFORMANCE thesis entitled presented by Merlyn Mildred Mondol has been accepted towards fulfillment of the requirements for BLED. degree in Counseling, Personnel Services and Educational Psychology 735/me1 Major professor Date July 3. 1973 0-7 639 ABSTRACT THE PARAMORPHIC REPRESENTATION OF TEACHER DECISION MAKING AS A PREDICTOR OF INQUIRY PERFORMANCE By , Merlyn Mildred Mondol The skill with which humans integrate information based on uncertain data is acquired over a long period of practice and experience. Previous studies have shown that ignorance and uncertainty about one's own cognitive judgmental processes are at the root of the problem of ineffective application of knowledge in judgmental tasks. Therefore. one solution to this problem would be to devise procedures to make explicit the characteristics of persons' judgmental systems and relate these to the characteristics of the judg- mental task. The major purpose of the present study was to investi- gate the possibility of training people to modify their judgmental policy preferences by using a form of cognitive feedback and a discussion and reflection training which would make them sensitive to the relevance of the informa- tion sources to their judgments. It was also intended to study the effects of such training on subsequent judgmental tasks. A second phase of this eXperiment was to study the relationships between judgmental policy preferences and inquiry performance. The beta weights assigned to the various sources of information utilized to make the judg- ments were used to predict the inquiry behavior of the subjects. Selected personality tests were also given to all subjects and the relative contributions of personality and cognitive variables to inquiry performance were studied. Fifty-four female college students were selected and randomly assigned to three treatment groups. namely: the discussion and reflection training group. the in—basket followed by discussion and reflection training and the control group. A repeated measures design was used in which a pre-judgmental task. requiring all the subjects to rate the likelihood of hypothetical students having in- structional problems in a classroom setting was administered. After the training was completed the judgmental task was repeated. For this particular study the judgmental post test revealed significant differences between the control group and experimental groups taken together. Further analysis suggested that the differences between the two experimental groups was not significant. Turning to the prediction of inquiry performance. results showed that the judgmental weights were not potent predictors of inquiry. Of the personality variables. internal locus of control was the best predictor. Although neither group of variables predicted inquiry very well. all the independent variables together predicted the dependent variables significantly. The prediction was significant especially for the inquiry dependent variables of bits. . competence and problem sensitivity. The results have at least two implications for a theory of judgment. First, the equation is a "paramorphic” rather than "isomorphic" representation of subjects' judg- mental policies therefore,any inferences regarding the actual sequence or process of information utilization.might not be warranted. 7 Second. training involving cognitive feedback and discussion and reflection on the subjects' weighting poli- cies can modify the policy preferences of teachers. For education the implications seem to be that optimal weighting patterns could be deveIOped for the judgments teachers have to make regarding instructional problems of students. Training programs could then be developed to facilitate the learning of effective and efficient decisionp making in the classroom. THE PARAMORPHIC REPRESENTATION OF TEACHER DECISION MAKING AS A PREDICTOR OF INQUIRY PERFORMANCE By 12” Merlyn MildredPMondol 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 May. 1973 'U’ ACKNOWLEDGEMENTS It was my privilege to work with Dr. Lee Shulman, my major professor, in his study of inquiry for about two years. To him I credit many of the ideas and tools of research that I now possess. Dr. Shulman has guided me throughout this project with his ideas, questions, criti- cisms and support. I am deeply grateful to him for his time and expert advice which he gave unstintingly with warmth and friendship. I wish to thank him for guiding me through all stages of my doctoral program. I also wish to thank my committee members, Dr. Joe Byers. Dr. Arthur Elstein and Dr. William Schmidt for their valuable help, especially during the time when the thesis proposal was being developed. My appreciation and thanks to Dr. Howard Tietelbaum for participating in the orals during the absence of Dr. Elstein. Dr. Schmidt's guidance in the analysis of the data was invaluable. Dr. Byers suggestions and criticisms were very helpful in rewriting the results chapter. His guidance is deeply appreciated. This project involved several hours of testing and scoring. I wish to acknowledge the help of Nancy Beeman for scheduling and administering tests: Coleen Briggs, for tran- scribing the tapes; Peter Miceli for helping with the test- ing and scoring of the logs and protocols. Last, but not least, I would like to thank Gertrude Allen for her depend- able help, in all phases of the project. Gertrude also ii typed hundreds of pages of rough drafts, as well as a major portion of the final draft, I am most deeply appreciative of the work Gertrude Allen and Betsy Frank did on the Final Draft. For permission to use the testing rooms at the Child Guidance Clinic in Saginaw, I owe grateful thanks to Jim Royle, Director of the clinic. I am also grateful to the many people at the Saginaw Valley College eSpecially I Dean Samuel Levine for allowing the use of the facilities at the college whenever I needed them for this project. Finally. I wish to thank my children, Rajendra. Rajkumari and Jitendra for all their love, support and understanding during periods of stress and tension. Their backing made this project possible. iii II III TABLE OF CONTENTS INTRODUCTIONOOOOOOOOOOOOOOOOOOOOOOOOOO Rationaleeeeeeeeeeeeeeeeeeoeeeeeeee Information Processing in JudgmentiflOOOOOOOOOOOOOOOOOOOCOOO Brunswik's Lens Model.............. The IN - Basket Technique and Teacher Decision Making.......... The variableSOOOOOOOOOOOOOOOOOOOOO. Cognitive Feed Back Training and TransferOOOOOOOOOIOIOI'O0.0.0.... REVIEW OF LITERATUREOO0.00.00.00.00... Literature on the Judgmental PrOCBSSeeeeeeeeoeeeeeeeeeeeeeeeee Empirical Studies of Multiple Cue Probability Learning or Cognitive contrOlOOOOOIOOOC0000000000000... Research on the Inquiry Process Using the Teacher's IN-Basket.... Literature Related to Internal Versus External Locus of Control As a Personality Construct.......... METHOD0.00000000000000000000.000...... Subjects........................... Pre Tests.......................... Design............................. DeveIOpment of Cardexes............ Training........................... Discussion - Reflection Training... iv PAGE 10 10 1h 1“ 25 27 30 35 36 36 38 39 42 CHAPTER PAGE The IN-Basket Experience.......... 46 Dependent Variables............... 49 Measures of Performance, Experimental Questions and Hypotheses........ 51 Questions Regarding Modifiability of Judgmental Policy Preferences Due to Training................. 51 Questions Regarding Prediction of Inquiry Performance............. 53 Hypotheseso....................... 5# IV RESULTSOO0.0.0.0.0....OOOOOOOOOOOCOOO 57 The Judgmental Policy Variables... 58 Reliability of the Instrument..... 60 Judgmental Policies and Stability Of Beta weightSOOOOOOIOOOOOOOOO. 60 Judgmental Policies and Training.. 65 Search for An Ideal Policy........ 7# Comparison on Post Deviations from Ideal Policy Using Pre Deviations as Covariables.................. 76 Judgmental Policies and Inquiry... 79 Inquiry Process Variables and Personality Variables........... 88 Relative Contributions of Personality Variables and Beta Weights of Policy Preferences as Predictors of Inquiry Performance 91 Examination of Individual Decision making PrOfileSOOOOOOIOOOOOCOO... 98 C ER PAGE v DISCUSSIONOCOOOOOOOOO0.0.0...00.00... 108 Judgmental Policies and Training.....flé'.OOOOOOOOOOOOOO. 110 Judgmental Policies and InQUiry....C...‘............... ll“ Cognitive Shifting and Equality Of Beta weightSOOOOIOOOOOOOOOOO 116 Personality Determinants of InquiryOOOOIOOOOOOOOOOOOOOOOOOO 118 Individual Decision Making Profiles, Pre and Post......... 120 IV SUMMARY AND CONCLUSIONS.............. 125 Conclusions and Implications..... 128 BEFERENCES..................................... 133 APPENDIX A Pilot Tests To Develop The Categories Of Information SOUI‘CQSeooeeoeeoeeeeeee 135 B Selected Personality Tests............ 139 C Cardex - Rating Scale................. 148 D Sample Cardexes....................... 150 E Discussion - Reflection Training...... 153 F Scoring Sheet for IN - Basket......... 160 G Competence Scoring Key................ 162 H Analysis of Covariance................ 166 I Sample Imbedded Problems in Teacher's IN - Basket........................... 168 J Judgmental Policy Summary Sheets...... 171 K Raw Data.............................. 179 vi “.3 h.“ u.5 “.6 11.7 ”.8 4.9 4.10 "'0 11 “.12 LIST OF TABLES Means, Standard Deviations of Stan- dardized Beta Weights X 1003 And Stability Coefficients Of Beta Weights of Five Cues on Pre and Post Tests measureSOOOOOOOOOO0.0000000000000000000000 Regression Analysis For the Prediction of Post Beta Weights With Five Pre Beta Weights as Predictors..................... Control Group XE- Average of Treatment Groups. F Ratio For Multi-Variate Test of Equality 0f Mean veCtorSeeeeeeoeeeeeeeeeee Means, Standard Deviations and Ranges For Standardized Beta Weights. For Information Sources on Pre and Post Judgmental Tasks For Three Groups......... Adjusted Least Squares Estimate of the Contrast. H1 = Control Group - Average Of Treatment GroupSIOOOOOIOOOOOOOOOOOOOOO. Significant Beta Weights in Pre and Post Judgmental Tasks Within Three GroupSOOO...0.0...IOOOOOOOIOOOOOCOIOOOOOOO Average D28 From the Ideal Policy and Differences Between Pre and Post Average D23 in the Three Treatment Groups ........ Means and Standard Deviations For Total Deviations of Policy Profiles For Each Subject From the Ideal Policy on the Pre and Post Judgmental Tasks For the Three Treatment GrouPSOlOIOOOOOOOOOOIOOOOOOOOOOO Correlations Between Pre Judgmental Beta Weights and Dependent Inquiry Variables... Intercorrelations Among Dependent measures 0f InQUiryeeeeeeeeeeeeeoeeeeeeeee Means, Standard Deviations and Ranges For The IN - BaSket Variables................. Correlations Between Post Judgmental Beta Weights And Dependent Inquiry Variables... vii PAGE 62 6h 68 69 7O 73 78 79 80 81 82 83 4.111 4.15 4.16 “.17 4.18 n.19 “.20 4.21 4.22 #.23 5.1 Correlations Between the Pre Beta Weights and Problems Sensed Within Corresponding Categories in Teacher's IN-Basket......... 85 Means, Standard Deviations of Problems Sensed Within Five Information Sources and Total Potential Problems Within Those categorieSeeeeeeeeeeeeeeeeeeeeeeeeee 86 Correlations Between Pre Beta Weights and Bits of Information Attended to Within Corresponding Categories of Information sources in IN‘BaSketeeeoeeeeeeeeeeeeeeeeee 87 Correlations Between Four Personality Predictor Variables and Inquiry Process Criterion measureSeeeeeeeeeeeeeeeeeeeeeeee 89 Intercorrelations Among Personality variableSOO00.000000000000000...0000...... 90 Statistics For Regression Analysis With Nine Predictors...COOOOOOOOOOOOOOOOOOOOOOO 92 Step-Wise Regression To Analyse the Percentage of Additional Variance Accounted For in Inquiry Variables By Addition of Each Independent Variable..... 94 The Percentage of Additional Variance Accounted For in Inquiry Variables By Addition of Each Independent Variable as Shown By a Step-Wise Regression AnalYSiSOOOO0....00.0.0000.000.000.0000... 95 Relative Contributions of Judgmental Beta Weights and Personality Variables in the Prediction of Inquiry Performance..... 96 Step-Wise Regression Analysis Adding Predictor Six (Internal Locus of Control). 97 Standardized Beta Weights For Six Re- presentative Subjects on Pre and Post Judgmental TaSkSOIOOOOIOOOOOO00.00.0000... 99 Correlations Between Problems and Bits of Information Attended Within Corre3pond- ing Categories of Information Sources in IN’BaSketeeeeoeeeeeeeeeeeeeeeeeoeeeeeeeeee 118 viii LIST OF FIGURES FIGURE PAGE 1 Brunswik's Lens Model................... 6 2 Chart of Experimental Procedures........ 40 ' 3.A Histograms of st for 53 subjects on Pre JUdgmental TaSkeeeeeeeeeeeeeeeeeeeeeeeee 61 3.B Histogram of R25 for 53 subjects on Post JUdgmental TaSkOIOOOOO00.000000000000000 61 4 Means of Beta Weights of Five Cues on Pre and Post Test Measures.................. 63 5.A Standardized Weighting Patterns for Six Representative Subjects on Pre and Post Judgmental TaSKSOOOOOOOOOOOOOOOOOOOOOOOO 100 to 103 a 5.8 Standardized Weighting Patterns for three Representative Subjects................. 104 6.A Means For The Three Treatment Groups on the Pre Judgmental Task................. 105 6.B Means For the Three Treatment Groups on the POSt JUdgmental TaSKCOCOCOCOOOCOOCCC 105 ix 1' Ill 1 t A A .II 1 (mill. 1.! ff|zf CHAPTER I INTRODUCTION Judgments or decisions require choices which control_‘ the lives of people and involve the efficient use of time. money and effort. So far no methods have been used in . teacher-training programs to facilitate effective decision making by teachers regarding students' instructional _' problems. Any attempts by supervising teachers to enhance the sensitivity to and the diagnoses of possible difficul- ties in students would be impossible to convey to the student-teacher, as the supervising teacher is apt to be vague about the basis of her own judgments. Moreover, the student, relying on her own introspective processes, may disagree with the teacher's observation of her. "Communica- tion under these circumstances is more likely to produce cooperative delusion than an accurate understanding". (Hammond, 1971). The main purposes of this research are first,to characterize the existing judgmental policies of teachers: second, to study the stability of these judgments over time. Third, to assess the modifiability of judgmental policy preferences through training using a form of cognitive feedback; fourthfio predict the performance of teachers in an actual or simulated situation in which judgments have to 2 be made, from their policy preferences: and finally. to study the relative contributions of judgmental policy preferences and personality variables to the prediction of the performance of teachers in an inquiry situation. Rationale Diagnosis and identification of students' problems is a specific instance of the more general process of judgment or decision making that a teacher is involved in within a classroom setting. How a teacher integrates information conveyed by several cues to form the judgment cannot be reduced to a simple and infallible rule that can be taught. Yet. in exercising her judgment the teacher must learn to assign differential weights to the various cues or sources of information. Once having characterized judgmental policies in terms of differential weights, it then becomes becessary to validate the relevance of policy in a simulated real situation. This is the role of the inquiry situation in the present research. Information Processing in Judgment In previous years the difficulties encountered in making decisions were usually blamed on the paucity and inadequacy of available information. Devices were therefore developed to increase the availability of data and to improve the dissemination of information. In Spite of the 3 technological expertise with which this problem has been remedied the effectiveness of decision making has not greatly changed (Slovic and Lichtenstein, 1971). There has therefore been a change in emphasis in recent studies to the process of integration and the interpretation of information itself. Interest now is with the cognitive operations performed on information and the processes and strategies humans employ in order to integrate discrete items of information into a decision. One approach used in the study of judgment is known as the regression approach, so called because of its characteristic use of multiple regression and analysis of variance to study the use of information by a judge. The linear model is one in which judgments are described as a' simple weighted sum of the values of the information available. For a given judge, judging a number of peOple. we let J represent the judgment and consider it as a dependent variable. The dimensions of information are designated by X's which are the independent variables. Given k sources of information the linear additive model can be described as follows: J a f (Xi) 1 = 1. 2. eseeek Since we are interested in a weighted sum of the X1, we may write: J = 80 + B X + ....+ B 1 1 + Ba X2 k xx The basic approach requires the judge to make 4 quantitative evaluations of a number of stimuli. or sources of information, each of which is defined by one or more quantified cue dimensions. For example. a judge might be asked to predict the grade point average for each of a group of college students on the basis of high school grades and aptitude test scores. The regression analysis identi- fies the weights assigned to the information cues available to the judge. The Beta weights reveal the relative degrees that the judgments depend on the various sources of infor- mation available to the judge. Although the linear model is a very powerful tool in predicting the judgments of the judges, it is only a para- morphic representation of the judgmental process, (Hoffman. 1960). The term paramorphic (Hoffman. 1960) borrowed from mineralogy and applied to judgment suggests that the model explains only certain preperties of the judgment but not all its characteristics. Therefore, the mathematical descrip- tion is incomplete and there is no way of knowing how accurately the underlying process has been represented. That is, the linear model adequately predicts the judgments made by the judge or will produce the same results as will the judge himself; however. the actual process by which the judge combines the cues to reach the judgment cannot be directly inferred. Thus, the relationahip between the mathematical model and the underlying judgmental process is not isomorphic but is paramorphic in nature. 5 Brunswik's Lens Mode; The correlational paradigm can describe the judges' characteristic method of processing and weighting informa- tion. In the Brunswikian framework (Hammond, 1955). it can describe the adaptive interrelationship between the organism and its environment. Thus, in addition to studying the degree to which a judge utilizes cues one can analyze the manner in which the judge learns the characteristics of his environment. In Figure l, Brunswik's Lens Model, each cue dimension has a specific degree of relevance to the true state of the - world. This true state. also called the criterion value. is designated Ye. The correlation between cue Xi and Ye indicates the relevance of the ith information source. This value is called the ecological validity of the ith cue. On the subject's side. his response or judgment is Y . and the s correlation of his judgments with the 1th cue is ri,s’ also known as the utilization coefficient for the ith cue. (Slovic and Lichtenstein, 1971). From the following regression equations both the criterion and the judgment can be predicted from linear combination of the cues: Ye aizligé 6 xi (1) k -._. Y8 =i=1>_bi,8 Xi. (2) “‘“J‘ + cow‘- ..x uwo+.:+ uxmmetxrau # 3 WM u omaoamom I\ oumnnam pouowpoum hufiumoaea mmmaommom «we» u em I\ my u unmadmom humanem .AnoomH .hoahmz was machvso so pommmv .mOmnommou m.uooHnsm was .mwuouwnu .moso ecu wnoam afinmnowumaou any mewsonm Home: amen mo amuwmfia .H .mam ReecH assesses m0 w.we u o NwmuGH Ufimam>0fl£0< mWUWH " “H mZOHmZMZHn mDHQEHHm Ammo: mzmq m.MH3mz:Mm xx»«4+.:+ ..x».£+...+mxs~o + .Xe3 u o,» E 0% u osHm> moauouwuu successes susfiseeuoeemum Hmuaoauouweam muses u mm 1‘ 0% u os~m> nowuouwuu 7 Equation (2) provides one possible model of the subject's decision making strategy and has been widely used to capture judgmental policies. That is. the actual weights the subject assigns to each of the cue variables in reaching the judgment are represented by the regression weights in the regression equation computed from the ratings of the judge. By virtue of the experimental control employed in the collection of the data, the only reliable source of judgment variance common across all subjects is the infor- mation supplied. Often these data appear as test scores on a set of protocols being judged. Assuming that a judge combined the information in linear additive fashion, the multiple regression analysis will be quite effective as a tool for describing the judgment process: that is, the set of regression weights when applied to the corresponding predictors can quite prOperly serve as a model for judgment. With certain limitations the regression weights signify the importance attached to each of the predictor variables by the judge. Regression weights could be converted into a set of relative weights in terms of which judges may be compared and contrasted with respect to their characteristic equations: and differences among judges may be related to training and other factors, such as personality, that could conceivably affect the utilization of data. Since in the present study the main interest was in the decision making process of teachers regarding the like- lihood of students having instructional problems in a 8 classroom setting. an attempt was made to relate the judg- mental policies to actual situations in which decisions or judgments had to be made. The In-Basket Technique and Teacher Decision-Making Teaching is an ongoing process of inquiry and decision making. The effective teacher must be sensitive to the problems of the students, be able to formulate hypotheses. use available information effectively to test those hypotheses. and make important decisions. Hopefully, by understanding the manner in which teachers utilize informa- tion at their disposal to arrive at judgments or decisions. predictions can be made regarding their performance on tasks and actual situations where inquiry and decision making are an important and integral part. A teacher must adequately understand the personal and social problems posed by the children in order to effectively guide them in the learning process. The sensitivity to and the identifi- cation of such problems would then be a necessary pre- requisite to the formulation of these problems and the making of the necessary decisions to resolve them. The Teachers' In-Basket. designed by Shulman (1963) is an instrument develOped to study individual decision making and inquiry behavior. It provides an Opportunity to observe inquiry within a setting in which the structural cues are minimal, and yet the potential situational stimuli are essentially the same for all subjects. The instrument 9 maintains the realism of the situation as well as affords experimental control for the observation of inquiry. The subject, who is a female elementary school teacher-in- training. is seated at a simulated teachers desk with its pile of potential problems. Many things have piled up on her desk and have been placed in her in-basket. It is her first day in the school and no pupils are present because of a school holiday. She may begin where she likes and do as she pleases. No time limit is suggested. Subjects are asked to think aloud in order to make their thoughts available to the observer. Thus, it is possible to deter- mine what general information source the subject is utilizing, whether the subject has perceived the situation or information as problematic, and whether the information assisted the subject in resolving the problem. There are three kinds of materials in the situation with which the subject may deal: 1) The contents of the in-basket which consists of telephone messages, test scores, schedules. memoranda and tasks and lists to be completed. 2) Written materials, records, report cards, etc. concern- ing both the school and the pupils in the teacher's class and 3) The human resources consisting of a school secre- tary, a school principal and "reference memory" available for consultation over an intercom with programmed answers to anticipated questions. 10 The Variables: The Problem_§ensitivi§y score is simply the total number of problems sensed by a particular subject. Tige is the number of minutes the subject chooses to Spend in the inquiry situation. Materials Attended is a measure of input, the number of ' pieces of material to which the subject attends in the inquiry period, representing the number of "bits" processed by the subject. Informatign Sources is a count of the number of kinds or categories of information brought to bear by the subject on ten selected problems in the in-basket situation. Competence is a measure of problem resolution. It is an independent judgment of how well each subject comes to understand the nature of the problem situation in the same ten selected problems used to score for Information Sources. General Inguigy is the summed score for problem sensitivity, mean sources and competence. Shifting is the total number of times the subject shifts his search from a bit of information in one source category to information in another. Cognitive Feedback Training and Trangfer "Although learning theorists have long emphasized the distinction between learning and performance. little attention has been given to skill in the application of ll knowledge in tasks which do not involve motor performance. Rather, there is an implicit assumption that once knowledge has been acquired. the application of this knowledge is largely dependent on certain experimental circumstances .... The position taken here, however, is that acquisition and application are independent components of learning in cognitive tasks as well as psychomotor tasks" (Hammond 1972). Psychologists have used the multiple-cue probability learning task in the study of human judgment typically carried out within the Brunswikian framework. In this type of learning the judge learns to integrate differential cues of various degrees of dependability so that the cue weights 8 in his judgmental system match the differential cue weights in the task itself. An example of this kind of learning was demonstrated by Hammond (1971). The learning task required the subject to arrive at a diagnosis or judgment that integrated the information provided by three cues. Cue A was correlated 0.8 with the criterion: and cue B and C were correlated 0.4 and 0.2 with the criterion, respectively. The relationship between the cues and the criterion were curvilinear, and due to the uncertainty built into the task no infallible rule for reaching the judgment would be formulated by the subject. Two hundred trials were used. On each trial the three cue values were presented on a 5-inch by 8-inch card in the form of bar graphs. The height of each bar indicated the value (1 to 10) of that cue. The subject was asked to interpret the three cue values and 12 arrive at his diagnosis on a scale from 1 to 20 for each display. Cognitive feedback was provided pictorially. informing the subjects about the correct weights of the cues, and by informing them verbally about the correct functional relationships between the cues and the criterion. The results showed that the probabilistic learning. utiliz- ing multiple cues in the complex cognitive task was facilitated by cognitive feedback. Since such tasks require judgmental learning analogous to that needed to make diagnostic judgments, it could well be used to study the changes in judgment or decision making by teachers when dealing with students' problems. The skill with which humans integrate information based on uncertain data is acquired over a long period of practice and experience. The judge is,however, vague about the basis of his own judgments and it cannot be assumed that experience will increase his awareness of his cognitive processes. Slovic et al,(l968) found that the more experienced the judge was, the less able he was to describe accurately how he arrived at his judgments. If ignorance and uncertainty about one's own cognitive judgmental processes are at the root of the problem of ineffective application of knowledge. then one solution would be to devise procedures to make explicit the characteristics of a person's judgmental system and relate these to the character- istics of the judgmental task. More specifically. a judge should be provided with a picture of the prOperties of the 13 task and a picture of his own cognitive judgmental system in terms that will allow him to compare the two. Hammond has employed computer graphics (Hammond, 1971), to provide these kinds of information for the learner. This research will investigate the possibility of changing teachers' judgmental policy preferences by training involving discussion and reflection on one's policy preference as a form of cognitive feedback. This study will also attempt to predict the perfor- mance of teachers-in-training in an unstructured in-basket situation in which they would be able to utilize informa- tion. sense and formulate problems about students, and make decisions about them. These predictions will be made from the weights subjects assign to various categories of information in the policy preference task and from person- ality variables. CHAPTER II REVIEW OF LITERATURE The literature to be reviewed for this study falls under four domains: 1) literature on the models of the representation of the judgmental process or the way in which information or criteria are combined in order to reach a decision, 2) literature on multiple cue prObability learning or cognitive control, 3) studies on the inquiry process using the Teacher's In-Basket and 4) internal vs. external locus of control as a personality construct. Literature on the Judgmental Process Several studies have been conducted in recent years within the tOpic of information utilization in judgment or decision making. As Slovic and Lichtenstein (1971) point out, there has been a shift in emphasis from studies concerning the problem of the inadequacy of knowledge and the availability of information to studies of the integra- tion process itself. "Their efforts center around two broad questions.... 'What should we be doing with it'? The first is a psycho- logical problem,that of how man uses information. The second problem is a more practical one and involves the 14 15 attempt to make decision making more effective and efficientt (Slovic and Lichtenstein. 1971). Studies attempting to represent the judge's weighting policy by means of the linear regression model include judgments about personality characteristics, (Hoffman, 1960), performances in college, (Dawes, 1970), physical and mental pathology, (Goldberg, 1968: Goldberg, 1970: Hoffman, Slovic and Racer, 1968). There are several studies that could be cited that would be relevant for the model, however, it was decided to limit the review to the works of Hoffman (1960). Dawes, (1970), and Goldberg (1968, 1970). The studies selected for review have one thing in common: namely, they all used the linear regression model ‘to capture the judge's idiosyncratic weighting policy. The paramorphic representation of clinical judgmen . Hoffman (1960), points out that the term"mental process" is often directly equated with subjective experience. The only way such a process can be inferred is through verbal phenomena such as verbal reSponses. It is, however. possible to "describe“ mental activity by means of mathe- matical models. The judgmental process can be studied in a controlled situation wherein the input (information) and the output (judgment) are known or capable of quantifica- tion. The accuracy with which judgment can be predicted would enable one to assess the adequacy of the functional 16 relationships between the input and the output hypothesized. In develOping models that could be used to study the judgmental process. Hoffman describes the restrictions and limitations of information available to the judge. Un- controlled use of clinical data might make judgment an "artistic venture” rather than a subject for scientific study. Controlling the judgment task is necessary in order to ensure objectivity and uniformity of procedure. He suggests that the situation be restricted in the following ways: a) the information available is reduced to a set of variables with respect to which all clients in the sample are evaluated: b) the information is expressed in number or in categorical reaponses: and c) each variable is at least on an ordinal scale. The Linear Model in which judgments are described as a simple weighted sum of the values of the information available has already been discussed in Chapter I. The use of relative weights or standardized weights was then developed by Hoffman following which the configurational models were developed and discussed. The Interaction Model was described as an "appropriately weighted composite of all possible first order interactions of the predictors". As the hypothesized relationships become more complex judg- ments become less dependent upon a simple weighted sum of the categories of information. It may be also true that for some categories of information extreme scores are more l7 decisive in judgment than scores in the middle range. Hoffman then discussed suppressor effects wherein a predictor carries negative weight because it accounts for variance in another predictor that is independent of the criterion. The use of relative weights however. obviates this difficulty since a predictor must correlate signifi- cantly with the judgment in order to obtain a significant relative weight. The emphasis is on the fact that these models or representations of human judgment are paramorphic repre- sentations which describe the judgment process and approaches the chemical description of minerals. The only relationship they are known to bear with the judge's mental processes is that. when employed, they will produce the same results as will the judge himself. This descrip- tion. however, is incomplete for there are other properties of judgment which it does not describe. Illustrations of linear models were given of subjects making judgments of ”intelligence" of 100 persons using nine predictors and judgments on "sociability” of 150 persons on the basis of profiles containing scores on eight selected Edwards Personal Preference Schedule variables. It was found that the judgments of two judges correlated .948 and .829 respectively with the best linear combination of the predictor scores. Thus, for the first judge, the linear model did an accurate job of describing the judgmental 18 policy. For the second judge, however, the linear model was not apprOpriate. Another question was considered. Is a judge able to describe the manner in which he utilizes information in arriving at his judgment? Judges were asked to distribute 100 points among the sources of information available in such a way that the distribution would reflect the relative importance of those variables. Comparison of subjective and relative weights showed that in one case there was a high degree of agreement of relative and subjective weights but great discrepancies for the other. It was found that. given the variables. a computing machine would come closer to producing the subject's judgments than he could himself. Finally, by using a configurational model the R obtained was .88 but by application of the linear model the R was .91. Thus, it was found that the linear model was the better predictor of the judgment. Dawes, (1970) did a study using the linear regression model to see whether 'bootstrapping' would work in the selection of student applicants into graduate programs. Dawes points out that from all the research done in this area the linear combination of criterion variables, which is a simple actuarial method, consistently does better than clinical judgment. In his paper on graduate admissions Dawes examines three principles, namely; a) the simple linear combination 19 of the criteria the admissions committee considers will do a better job of predicting performance in graduate school than will the admissions committee itself: b) behavior of the admissions committee studied can be simulated by a linear combination of the criteria it considers: and 0) under certain circumstances the paramorphic representation of the judge's policy, i.e., the results of the simulation, may be more predictive of the outcome criterion than is the judge himself. Goldberg (1970) terms this latter phenomenon. in which a model of a judge works better than the judge who was the basis for the model, "bootstrapping". The Admissions Committee of four members required all applicants to produce GRE scores, a transcript of past work and letters of recommendation. Each member rated the applicants on a six-point scale ranging from 'reject now' to 'offer a fellowship'. The following Spring faculty ratings of actual performance in graduate school were obtained. It was found that GPA and_QI (quality of under- graduate institution) correlated more highly with later faculty ratings than the ratings of the admissions committee. In order to study the possibility of bootstrapping. three hundred and eighty-four applicants for the fall were studied. The dependent variable of interest was the average rating of the admissions committee. The multiple correlation predicting the admissions committee rating from GPA. GRE and Q1 was .78. A cutoff point based on a linear 20 combination of the three predictor variables was found. 55% of the applicants scoring below this point could be eliminated on the basis of the paramorphic representation of the admissions committee's behavior without a single error being committed. The correlation between the linear combination and later faculty ratings was higher than that between the admissions committee's ratings and later faculty ratings. These results indicate that decisions made by such methods might be more valid than those made by judges relying on their own intuitions. He suggests that a mathematical model is an abstraction- of the mental process being modeled. A decision maker may be distracted by physical. mental and emotional extraneous variables that influence his most recent applications of knowledge. A paramorphic representation of his behavior would not be affected by such extraneous variables. Goldberg (1968) focuses on clinical judgments and the diagnoses of physical and mental pathology, with an emphasis on the process of clinical inference rather than the validity or reliability of such judgments. He suggests that a search for a model should be made which uses information as its "input”, combines the data in some Optimal manner so as to produce as accurate as possible a capy of the responses of the judge regardless of the actual validity of the judgments themselves. The answer according to Goldberg is to start with the 21 simplest linear additive regression model and then to proceed to introduce complications only so far as is necessary to reproduce the inferential responses of a particular judge. If we assume that the judgments can be reproduced by the model J = b1 X1 + b2 X2 ... + bk Xk The b values found on one subset of the judge's reaponses can be cross validated on another subset of judges responses to determine the accuracy of the linear model. The result- ing correlation (Ra) represents the extent of agreement between the linear model and the inferential products of the judge. It is possible to represent the stability of the responses (rtt) or the extent to which one can predict his judgments from his own previous judgments of the same stimuli. This reliability coefficient could be viewed as the upper limit on the predictability of any model. To the extent that the value of Ra approaches the Value of rtt' the model can be seen as representing the cognitive processes of the judge. Since clinicians frequently describe their judgmental processes as complex involving curvilinear, configural and sequential utilization of cues one might expect that the linear additive model would be inadequate in providing a gOOd representation of their judgments. Goldberg suggests that the analysis of variance (ANOVA) could be used 22 alternately if a) the cue values are treated as categor- ical rather than continuous variables and b) the cues are orthogonal. If the number of cues or the number of levels per one is not too large it might be possible to use a completely crossed experimental design, (all possible combinations of each of the cue levels). Thus, a signifi- cant interaction between cues Xland X2 implies that the judge was responding to particular patterns of those cues. In study after study it was found that the accuracy of the linear model was almost always as reliable as the judgments themselves and the introduction of complex terms rarely served to increase the cross-validity of the model significantly. Goldberg postulates three possible reasons for these findings: a) human judges behave like linear data processors, but somehow believe that they are more complex than they really are: b) human judges behave in fact in a rather configural fashion, but the power of the linear regression model is so great that it serves to obscure the real configurational process in judgment: c) human judges usually behave in a decidedly linear fashion on most tasks but on a few tasks they use more complex judgmental processes. Subject matter experts in three different fields were consulted to help select diagnostic decisions of a clearly configural nature and three judgmental tasks. one from each field. were developed for intensive study. Nine judges were 23 asked to make diagnoses for 192 hypothetical patients (two administrations of each of the 96 possible cue combinations). The judges made their diagnoses on a seven-point scale. The inferences of each judge were analysed by ANOVA. The major finding was that the largest of the 57 possible interactions for the most configural judge accounted for only 3% of the variance of his responses. 0n the average. roughly 90% of a judge's reliable variation of response could be predicted by the simple linear additive regression model. The results of the other two studies were remarkably similar. There- fore. he concluded that the hypothesis that judges can process information in a configural fashion, but that the general linear model is powerful enough to reproduce most of those judgments with very small error was the most plausible one. That is, the configurational model showed no demon- stratable gain over the linear model. Goldberg (1970) considers the question of whether the accuracy of prediction from the linear model can be im- proved when the criterion information is not available to the judge. This he points out could be done if the clinician's judgmental strategy can be separated from his judgmental unreliability. This is what the linear model can do. Since a mathematical model is an abstraction of the process it models it is free from the influences of extraneous variables such as boredom, fatigue and other physical. mental and emotional distractions. By modifying 24 the lens model, he goes on to develOp a model which specifies the conditions under which the model outperforms the performance of the judge. In his study the judgmental problems of differentiating psychotic from neurotic patients on the basis of MMPI profiles was used (Meehl. 1959). The profiles of 861 . psychiatric patients who had been diagnosed as either psychotic or neurotic were used as predictors. The validity coefficients in Meehl's study were used as an index of diagnostic accuracy. Since criterion information was avail- able for this task it was possible to compare the validity coefficients of each judge's model with that achieved by the judge himself. In each case it was found that the model was more valid than the judge himself. When a composite judgment of all 29 clinicians was used it was seen to be more accurate than the typical individual judge and was not improved by using the"modeling" procedure. In situations where criterion information is lacking the most accurate predictions may come from the composite judgments of the total group. This concludes the discussion of the literature pertaining to the judgmental process and the models that describe it. In the present study the linear model is used to capture the judgmental policies of subjects and the relative or standardized beta weights are used to predict inquiry performance. 25 Empipical sppgigs of multiple cue probability lgarning or cognitive control. Multiple cue probability learning is represented schematically by the lens model in which the judge learns to integrate differentially weighted cues of various degrees of dependability. The differential cue weights in the task must be matched by the cue weights in the judgmental system of the subject. An example of multi- ple cue probability learning has been cited in Chapter I. Hammond (1971) and his colleagues have focused on the learning aspect of judgment. They have contended that specific feedback derived from the lens model (i.e., feed- back about the weight the subject gives to each cue, and the weight the environment gives to each cue) is more effective than non-Specific or outcome feedback. Hammond has taken the position that acquisition and application are independent components of learning in cognitive tasks and demonstrates that even when knowledge is complete imperfect cognitive control can prevent high achievement in judgmental tasks. Previous studies relating to the learning of clinical inference have shown no improvement in the predictive accuracy of clinical judgments. Goldberg and Rorer (1965) did a study in which judges were given immediate feedback on a task requiring the differential diagnosis of psychosis versus neurosis from MMPI profiles by three groups of judges: expert, middle and naive. Only the naive group showed any 26 transfer on the testing profiles. In spite of introducing a number of experimental variations in an attempt to increase judgmental accuracy none of the experimental group showed any substantial learning. Hammond, (1971) discusses the inadequacy of this approach and contends that providing the correct answer after having made a judgment is virtually useless since the outcomes are related to the cues in complex and uncertain ways. An approach to the solution involves making explicit the characteristics of a person's judgmental system and relating these to the judgmental task. In general the sub- . ject should be able to compare what should be done with what he is doing. Information about the task properties enables the subject to perceive not only that his judgment was in error but why it was in error. More specifically. the subject should be provided with the differential weights he actually assigns to the cues and allowed to compare them with the weights required by the task. Studies were done using traditional outcome feedback as a control. verbal and pictorial feedback, and computer graphics to provide cognitive feedback. The results clearly showed the superiority of the cognitive feedback group. Hammond and Summers (1972) point out that multiple-cue probability learning tasks can be varied in three ways: a) the number of cues related to a criterion can be varied: b) the uncertainty associated with each cue can be varied 27 by creating differential cue validitiessand c) the form of relationship between cue and criterion can be varied. They use the"1ens model“ equation to explain the relation- ship between the subject and the task. Two subjects might have identical achievement indexes either because of perfect knowledge or because of perfect cognitive contro. Thus, poor performance in complex inference tasks can be attribu- ted to difficulties in cognitive control, as well as to difficulties in acquiring knowledge about the task. When the criterion is a simple linear function of the cues subjects achieve a high level of performance with little difficulty. They conclude that cognitive feedback can facilitate performance on judgmental tasks as there is evidence that computer technology can be used to produce this kind of feedback. In the present study cognitive feedback was provided to the subjects verbally and from computer printouts. Since computers were not available this could be done only once. In addition, however. the relationships between the cues and the task were discussed and reflected upon. Since no criterion was available the interest in this research was in the changes in policies per se. Research on the inquiry procgpe using the Teachep;§ Ip-Basket. The teacher's In-basket designed by Shulman (1963) is a situational simulation in which the subject is asked to 28 role-play an inexperienced teacher beginning her first teaching job. The in-basket technique has been described in Chapter I: and in Chapter III a more detailed description of the instrument and the variables of interest are fully discussed. Therefore. in this chapter a brief review of the findings of the following studies will be made: , Shulman (1963): Shulman. Loupe and Piper (1968): Piper (1969): and Loupe (1969). Shulman (1963) was interested in investigating seeking style. as a determinant of inquiry behavior. Seeking style. according to Shulman, is a continuum, the two extreme poles of which are the dialectical and didactic seekers. The basic prediction was that subjects identified as dialectical would surpass those identified as didactic in their ability to inquire effectively. It was hypothesized that the personality characteristics that typified the dialectical seeker would predispose her to be a more effective inquirer than her didactic counterpart since she preferred the complex, was willing to risk, and was more Open to her environment she would be more willing to engage in inquiry. In both studies Of inquiry (Shulman. 1963: Shulman et a1” 1968) the results indicated that dialectical seekers exceeded didactic seekers in all measures of inquiry. The dialectical seekers spent more time in inquiry, attended to more"bits" of information. consulted more sources of infor- mation. sensed more problems and reached more competent 29 solutions. In the study by Shulman et al.(l968), each subject participated in the teacher's in-basket twice, once before and once after student teaching. The results showed that b as the influence of seeking style increased from the first to second administration. the influence of GPA on inquiry decreased.h It was also found that shifting was highly correlated with inquiry variables and relatively uncorre- lated with seeking style. This indicated that inquiry may be a function of two very different factors: namely, seeking style and a learned strategy. Based on the above finding, Piper (1969) and Loupe (1969) used different training experiences to facilitate inquiry performance. Loupe focused on changing the learned strategy component of inquiry performance, whereas Piper studied the effect of changes in the affective component. The results showed that seeking style was not a significant predictor of inquiry performance; however, it did predict performance on the problem solving test used at the end of the training program. A possible explanation was that seeking style is a meaningful determinant of inquiry only when intellectual prerequisites have been met by all subjects. Openness training did increase time spent and informa- tion used in the inquiry situation but there were no significant changes in problem sensitivity or inquiry 30 competence due to the training. Problem solving training increased the variety of information used in solving a problem. The differential results of the two training techniques tended to confirm the distinction drawn by Shulman et a1. (1968) between commitment to inquire and learned problem solving strategies. The present experiment attempted to sift out the relative contributions Of personality variables and judgmental beta weights in the prediction of inquiry performance and to see whether the independence of the affective and the cognitive components could be replicated, treating judgmental policies as cognitive component pre- dictors and personality variables as predictors of the affective component. Lgterature reggped to internal versus external locus of pontrol as a pepponality construct. In studying the personality determinants of inquiry a new personality dimension was included in the present experiment. Locus Of control is a personality construct which refers to a person's perceptions of the agency of control of the reinforcements he receives. If a person perceives that an event is contingent upon his own behavior he is said to have internal locus Of control. That is. he feels that the reinforcements which he receives occur primarily because of his own purposeful behavior. On the other hand if he feels that the reinforcements he receives 31 occur primarily because Of forces beyond his control such as luck or chance, then he is said to have an external locus of control. Rotter (1966) developed scales to measure this dimension and related these measures to a wide variety of behaviors. He explains the theory behind his hypotheses. ”In social learning theory, a reinforcement acts to strengthen an eXpectancy that a particular behavior or event will be followed by that reinforcement in the future.... It follows as a general hypothesis that when the reinforce- ment is seen as not contingent upon the subjecth own behavior that its occurance will not increase an expectancy as much as when it is seen as contingent.... It seems likely that, depending upon the individual's history of reinforcements. individuals would differ in the degree to which they attributed reinforcements to their own actions". Rotter goes on to explain how a generalized expectancy is developed for a class of related events which might affect a wide variety of behaviors in a variety Of situations. Such generalized expectancies can be measured and are predictive of behavior under different circumstances. The scale developed by Rotter to measure these expectancies for control is a 29 item scale including 6 filler items. He reviews a series of studies in which support was found for the following hypotheses. A person high in internal locus of control is likely to:l) place greater value on skill or 32 achievement reinforcements and is more concerned with his ability and failures: 2) he is likely to take steps to improve his environmental condition: and 3) he is resistant to subtle attempts to influence him. Butterfield (1964) studied the relationship between locus of control and frustration; and locus of control and anxiety responses. He also investigated the relationships between locus of control and students' academic aspirations and expectations. The results revealed that frustration reactions become less constructive as locus of control becomes more external. It was also found that debilitating. anxiety reaction scores increased and facilitating anxiety reaction scores decreased as locus Of control became more external. Regarding the relationships between locus of control and achievement, interesting results were reported: as locus of control became more external the range of expected grades increased and the grades which subjects earned increased. This was interpreted to show that inner directed students study those things which they regard interesting while externals are more other-directed, thus mostly study what their professors regard as important. Mirels (1970) did a factor analysis on the I-E Scale to study its factor structure. He found that the scale was not unidimensional but found two factors: one concerning the mastery over the course of one's life and the other concern- ing the belief of the extent to which a citizen can effect 33 political change. Schnieder (1972) studied the relationship between locus of control and activity preferences. The preference for skill versus chance activities was studied. He found that 'internals' tend to prefer skill activities over chance to a greater extent than do 'externals'. He also found that _ the correlations of skill-chance preferences and locus of control varied as a function of the sex of the sjubect and with masculinity or femininity of the skill items. He concludes from this study that the construct of locus of control is multidimensional. The use of this scale in the present study was purely exploratory. No relationships were hypothesized between locus of control and inquiry performance except that an individual high on the internal end of the scale would tend to sense and solve more problems in an instructional situation. He would tend to fall into the personality type that Shulman calls dialectical. This study will attempt to research the following questions: 1) Are judgmental policy preferences stable over time? 2) Can policy preferences be modified by training involving cognitive feedback and discussion and reflection on one's policy preference? 3) Can inquiry performance be predicted by the weights subjects assign to different cues in reaching a judgment? 5) What are the relative contribu- tions of personality variables and judgmental beta weights 34 in the prediction of inquiry performance? In the following chapter the design of this study, including the training and testing conditions will be discussed in detail. CHAPTER III METHOD The research design for this study involved the following steps: (1) The develOpment of cardexes having information on hypothetical children in the classroom, and the development of rating scales on which all the subjects would rate the likelihood Of these hypothetical students being instructional problems in a classroom setting. (2) The capturing Of the judgmental policies of all subjects from the rating of the cardexes using a multiple linear regression analysis. (3) The prediction of inquiry performance of a randomly selected subgroup of subjects measured by the teacher's in-basket from the judgmental beta weights and from selected personality tests. (4) The development of training procedures using dis- cussion Of the variables on the cardexes and the reflection of subjects' policy preferences. (5) The administration of the in-basket and training sessions. (6) The administration of the post judgmental task or rating of the cardexes again on the same questiOn. (7) The testing of the hypotheses based on the theories of judgment or decision-making, cognitive feedback and inquiry. 35 36 Subjects Fifty-four female students in elementary and secondary education were randomly selected from the Educational Psychology classes at Saginaw Valley College. The subjects were contacted personally in their classes and were asked to participate in the present study for a total of seven hours for credit toward their grade in Educational Psychology. They were allowed to drop their lowest test score out of six tests given during the term. Subjects were randomly assigned to one of three treatment groups. Each subject was given an identical description of the task. and was contacted individually by telephone to set up appointments for each training and testing session. Sub- jects were told that they would learn things relevant to teaching and that their work was valuable only if they completed all phases of the study and would thus receive credit only if they completed all the work. Pre Tests All the fifty-four subjects were given a group of selected tests and inventories which took approximately thirty-five to forty-five minutes to complete. The measures were comprised of Complexity (Barron, 1967): Lecture- Discussion (Shulman, Loupe and Piper, 1968): Political Position (Shulman, Loupe and Piper,(l968): and Internal vs. External Locus of Control (I-E Scale, Rotter, 1966). 37 The Complexity Scale consists of thirty items designed to elicit a statement Of preference for either simple or complex situations. The subject either agrees or disagrees with each item. A high score indicates preference for complexity, ambiguity. etc. The Lecture-Discussion Scale consists of six items which relate to a student's preference for lecture or dis- cuSsion in a classroom situation. The subject responds to the items in an identical fashion as the items in the com- plexity scale. The items were therefore presented inter- spersed with the complexity items. A high score on lecture- discussion indicated a preference for the discussion approach in the classroom. The Politics Scale is a simple four-item self-report from which subjects' political positions could be elicited. The scoring of the items was in the direction Of liberalism. This Political Scale was a revision of the one used by Shulman. Loupe and Piper, (1968). The Internal versus External Locus of Control Scale is a twenty-nine item forced choice test including six filler items intended to make the purpose of the test somewhat more ambiguous. Subjects were told to select those items which they actually believed to be more true as far as they were concerned: that this was a measure of personal belief and that there were Obviously no right or wrong answers. These four personality tests were used to predict 38 inquiry performance in the in-basket situation. Those high on these variables were classified as dialectical seekers by Shulman (1963) and were shown to be highly successful in inquiry. In the present study the contribution of these personality variables in the prediction of inquiry perfor- mance was examined. The pre judgmental task was administered to all the subjects individually. They were asked to rate one hundred and eight hypothetical students on five different variables. The ratings were to be made on a seven-point-scale on the following question: ”What is the likelihood that this student will be an instructional problem in a classroom setting?" The develOpment of the cardexes and their use in the judgmental tasks are fully described later in this chapter. Desigp This research was comprised of two related studies. The first study or phase was concerned with the modifica- tion of judgmental policy preferences of subjects through training. The second phase investigated the prediction of inquiry performance from judgmental policies and selected personality tests. In Study I the experimental design consisted of three levels of training: discussion-reflection training. in-basket followed by discussion-reflection training,and 39 control. This study had a repeated measures design. The major dependent variables were the judgmental policies or beta weights of the subjects on the post judgmental task. A linear regression analysis was used to capture the judgmental policies for both the pre and the post tests of all the 53 subjects. This phase of the experiment involved . all 53 subjects in different kinds of training and the post judgmental task. In Study II or the second phase Of the experiment the performance of a subgroup of 21 subjects was predicted in the inquiry situation from their policy preferences and from selected personality tests. The measures of perfor- mance were: the number of problems sensed within various information sources in the in-basket, the competence measure of problem solving, the time spent in inquiry. the amount of shifting, and the number of "bits" of information attended to. An attempt was made to determine how far a teacher's policy preference is predictive of inquiry performance. Figure 2 presents the experimental procedures in the form of a chart. Development of Cardexes Five variables or categories Of information sources were selected which would correspond. with a little modifi- cation. with the information sources categories in the Teacher's In-Basket. The five variables were: 1) SES. 4O A. AA!- .I.I.( llliul 13' lulil . . 1 .N madman mums wsflnfimua huwHQGOmno nowuooammm :1 no mmm:sH 1 1 I. 1 1 uses umom :noummsomfin x oosmsuomuom HHH howaom noxmmm:sH oosouomeum AHN a zv mo sowuowpoum SOHHom momma e wagswmuH muHHmGOmuom l I: 1|: 1...: .l dbwuooamom .l i llnll: 1' II it . umoa umom :QOfimesOmHn Ana zv HH howaom monouomoum . .338 I I. i, lilAAl: mumoa I. huwamaOmuom sees seem I 1 1: 31H .1. ziv 3380 H senses mosouomoum howaom umOH umom mewsfimua masouw.mumoa mum unmanmoomm AaHHmZHmmmxm Md quflu 41 2) IQ. 3) grades, 4) sex, and 5) comments. The first three variables had three levels each of high. medium. and low. The remaining two variables had two levels each, namely male and female. and either favorable or unfavorable. ' TheSe variables or cues were treated as categorical variables and a completely crossed experimental design was used where all possible combinations of each of the cue levels were presented on the cardexes. -To avoid any misinterpretations of scores or values presented within the categories of high. medium and low, special procedures were adopted. A pilot test was develOped consisting of grades ranging from A+ to E, IQ scores ranging from 130 to 85 and a list of occupations selected from published lists of vocations and professions in different socioeconomic classes. All these values were presented in a random order to a group of teachers-in- training in an elementary Science Methods class. Comments were selected from cards presently being used in various school systems and were also presented to the subjects in a random order to be rated as favorable or unfavorable. (See Appendixgg). The subjects were instructed to rate the scores and occupations as high, medium or low and the comments as favorable or unfavorable according to their own subjective perceptions. Only those values were used on the cardexes where a two-thirds majority of the subjects agreed upon the rating. 42 A completely crossed experimental design was used where all possible combinations of each of the cue levels were presented on the cardexes. This procedure was necessary to Obtain the independent contributions Of the cue dimensions in the prediction of the judgment and to ensure obtaining reliable beta weights. In order to have all possible combinations of the levels Of information sources presented on the cardexes, it was necessary to develOp one hundred and eight hypothetical students having different combinations of those variables (3 x 3 x 3 x 2 x 2 = 108). To control for order effects. a b x 5 latin square procedure was used to present the information in all five positions of sequence an equal number of times. The presentation of the cardexes was randomized and the random order was kept constant across subjects. The names of the hypothetical students were selected at random from the telephone directory, the occupation of the father was used as an indicator of SES and comments were used to describe the personality of the hypothetical students. Development of the Cardex Rating Scale In order to quantify the judgments made on the basis of the five specific cues on the cardexes a seven-point rating scale was developed. Each cardex was to be rated on one such scale on the following question: "What is the chance that this student will be an instructional problem in a classroom setting?” If the subjects felt that the 43 hypothetical student showed a 95% or greater chance of becoming an instructional problem they were asked to place a checkmark at 7. If they felt that the student had a 50% chance of becoming an instructional problem they were asked to place the checkmark at 4: and if they felt that the student had a 5% or less chance of developing instruc- tional problems, they were to place the checkmark at 1. Appendix C has the completed directions that were used when administering the judgmengal task. To prevent a subject from being influenced by her prior ratings, each rating scale was presented on a separate page. The scales were numbered from one to one hundred and eight. Corres- ponding numbers were used on the back of the cardexes to ensure that the individual ratings could be identified if necessary. This also kept the order of the presentation of cardexes constant across all subjects. Computer cards were punched including the five cues in a certain combination of levels as independent variables and the correSponding rating as the dependent measure for one hundred and eight ratings per subject. By doing a multiple linear regression analysis judgmental weighting policies were derived, thus each subject's policy prefer- ence was represented in the form of a regression equation. Training All training took place in small groups of four to five subjects for approximately an hour and one half. As 44 stated in Chapter I, the central problem Of this document was to modify judgmental policy preferences through discussion and reflection training as a form of cognitive feedback. A second training procedure was the in-basket experience followed by discussion-reflection training. The effect of these two forms Of training was studied by measuring the differences in policy preferences of the three training groups on the post judgmental task. Discussion-Reflection_Trainipg The discussion-reflection training consisted of feedback and reflection on subjects' policy preferences presented to them in the form of a standardized regression equation and discussion Of categories of information sources used in the cardexes. The objectives Of the training were to make explicit the judgmental policy preferences of each subject by providing feedback regarding the beta weights that they had assigned to the various cues on the pre judgmental task and to sensitize them to the effect of these variables on student behavior in classroom settings. The instructional materials were taken primarily from ngppipg_and Human Abilities (Klausmeier and Ripple. 1971) and were discussed under the five headings correSponding to the five variables used in the development of the cardexes. (See Appendix E ). There were two distinct learning situations employed to attain the objectives: 45 l) Fgedback and reflection_g;spussion Of pgligy preferenceg. This phase was conducted on an individual basis. The subject was made aware of her policy preference and by being allowed to study the relative weights she actually assigned to the five cues. A discussion and reflection on why she had assigned those weights and whether her subjective weights coincided with those were discussed. 2) Discussion_gf cue variables as related to student behavior. During this phase a ten minute didactic presenta- tion was given to subjects in small groups followed by a discussion of each variable. Prior to the discussion of each section in the training booklet (see Appendix E), sub- jects were asked to read the section aloud to the rest of the group in turn. This procedure seemed to encourage the active participation of each member in the group. The variables were discussed in the following order: 1) what SES means, the plight of the child from low SES families, environmental factors contributing to the various character- istics of the low SES child: 2) what IQ means, range of 10's in school children, limitations of IQ scores in organizing instructional programs: 3) what grades mean. sources of information about achievement, precautions regarding the use of grades as valid measures of achievement: 4) sex roles, sex differences, precautions: 5) comments on student records. personality characteristics related to school achievement. At the end Of the session, subjects were asked to take 46 the booklet with them and were given instruction to read the information, noting the important points and reflecting on their policy preferences before the post testing. There was a least a week's time lapse between the training and the post judgmental task. The In-Basket Experience A major purpose of this study was to study the rela- tionship between judgmental policy preferences and inquiry performance: or to test whether policy preferences could predict teacher decision making and problem solving in a realistic, teacher relevant situation. Further, if the underlying processes in the two situations were similar, the in-basket experience would have the effect of sensitiz- ing subjects to the influence Of certain variables on problem behavior in the classroom. Therefore, the experi- ence would have the effect of changing the differential weights one assigned to those sources of information on the post judgmental task. The Teacher's In-Basket was employed since it provided the realism necessary to serve as an externally valid test of teacher decision-making. It is composed Of specific sets of material into which potential problems are embedded to stimulate maximum inquiry. Subjects were asked to role play the part of a sixth grade teacher and were asked to think alOud in order to make their thoughts available to the Observer. Thus, it was possible to determine the sources Of 1.7 information the subject was utilizing, whether the subject perceived the situation or information as problematic and whether the information assisted the subject in resolving a problem. The subject was brought into a one-way Observation room and was told that this was her new classroom in Ridge Forest Elementary School. She was seated behind a desk on which was an intercom, the in-basket materials. a folder containing current report cards, cardexes, attendance book, anecdotal records and paper and pencil. The in-basket materials included telephone messages, tasks and lists to be. completed, test scores. schedules, etc. The subject was told that she was a new sixth grade teacher taking over the class in December after a succession of substitute teachers. She could call out for additional information to the school secretary and"reference memory"over the intercom. She was next told that the success of the research depended on her ability to think out loud so it was necessary to make all her thoughts verbal whether she deemed them trivial or not. Prior to being left alone to proceed with her inquiry the subject was given training to think aloud. She was given five or six different Objects and was asked to group them in all possible ways giveing reasons for why they belonged together as she went along. Before leaving the room the eXperimenter explained the situation to the subject. The subject was informed that she 48 could use materials from the cumulative records and medical records only one at a time. They were placed on a table across the room, all other materials were for her unrestric- ted use but that she should only write on the pieces Of paper provided. The subject was then asked to proceed by reading the written description of the situation out loud. - Embedded within the in-basket were tasks, letters and memos designed to structure the situation to the degree that all subjects would at least attempt to undertake the same basic set of tasks for example. identify those students who must see the school psychologist. The choice of whether to continue inquiring or not was up to the subject. The sub- ject was observed in the above situation by a single Observer who also functioned as reference memory and school principal. Everything the subject did could be viewed through the one-way mirror and everything the subject said was heard by the observer. The observer dictated a complete log of what the subject did during her stay in the situation on a tape recorder and simultaneously checked Off lists of information attended to within categories of information sources. Appendixf‘ gives a typical scoring sheet used to keep count of the bits within information sources. Detailed scoring and interpretations Of the subject's inquiry performance were done after all the observations were made. 49 Dependent-Vgpigbles A number of scores were abstracted from the observa- tions of inquiry behavior of the subject. The potential problems, embedded in the in-basket and students' records were of two basic types, the first being a simple isolation of expectancy and the second kind was in the form of con- flict between sources of information. There were approxi- mately 250 such potential problems identified in the problems manual. The Problem Sensitivity Score was simply the number of problems sensed by a particular subject. Since use Of information is important to the process of. inquiry two measures of information usage were used. pipp was simply the number of times the subject consulted any source materials including her own written notes, for information: and Information Sources was a measure of diver- sity of information usage. For scoring purposes the Teacher's In-Basket was divided into ten basic problem areas. A record was kept of the different sources consulted within each area. Thus, if a subject looked at a cardex, the sociogram. a cumulative record. the attendance book and the same cardex again, the number of sources consulted would be only four if they all concerned a particular problem. Total sources were all the sources used across the ten problem categories: whereas, the mean sources was calculated by dividing the total sources by the number Of problem areas into which the subject inquired. 50 Competence was a qualitative measure of problem resolution. For each of the ten problem areas, model resolutions were written at varying levels of complexity. the lowest being a simple recognition of the problem and the highest represented the fullest understanding of the problem achieved with the use Of all available materials (see Appendix G). The model solutions were rated 1-5. according to their complexity. The competence score was derived by comparing the subjects' problem resulution with this standard. Tip; was another variable of interest since a subject remained in the situation until she called on the intercom and said she was done. A cognitive process variable was shifting which was the total number of times the subject shifted his search from one bit of information in one source category to infor- mation in another. Two other variables of interest were the problems senseg within categories pertaining to SES, IQ. grades, sex and personality variables. These were labeled Ppoblems Info. Bits Info on the other hand, were bits of information attended to involving SES, IQ, grades, sex and personality variables. Thus, the two measures were not entirely independent. Finally. the dependent measures Of major interest were the beta weights for each subject on the post judgmental task. The cardexes used on the pre test were used again on 51 the post test. [Mpgpures of;§pp§ormanceL:§gperimental Questions and Hypotheses. Much research effort has been directed toward in- creasing our understanding of the relationship between different personality types and the general manner in which 4 individuals mediate the world around them cognitively. But so far no effort has been made to relate the judgmental process and active inquiry or problem-solving even though both processes involve the utilization of information in the making of a decision. Therefore, it will be of interest to determine how far a teacher's policy preference is A predictive of inquiry performance and whether an individual when faced with a situation which must be categorized and acted upon will tend to pick up cues that will allow him to classify it in his most commonly exercised schemata. Questions regarding modifiabigity Of_judgpgntal policy preferences due to tpgining: In order to develop sensitivity to the problems of students in a classroom a teacher-in-training must be aware of possible difficulties a student can encounter due to factors such as low aptitude, emotional disturbance, poor economic conditions at home, poor past achievement and advantages and disadvantages of being male or female. Aware- ness of one's own biases and weights that one gives to these variables in making a judgment about student problems is also a critical variable in understanding how one integrates 52 information to reach a decision. Previous studies (Hammond, 1971) have used cognitive feedback to improve the accuracy and the effective application of knowledge in judgmental , tasks. The possibility of the modification Of one's judge mental policy by such procedures could raise the following questions: 1. Can the judgmental policies of teachers-in-training be modified by giving them training using discussion and reflection in which the characteristics of their judgmental system are made explicit and are related tO the judgmental task? 2. Will the in-basket experience serve to modify the judgmental policies of teachers-in-training on subsequent judgmental tasks? 3. Are there significant differences between teachers- in-training who do not receive discussion training and those who receive discussion training in the way they perform on subsequent judgmental tasks? 4. Are there significant pre test, post test differ- ences in the judgmental policies of teachers-in-training who have had both the discussion training and the in-basket experience and those who have had no training at all and those who have had the discussion training only? 5. Are the contributions of personality variables and policy preferences as predictors Of inquiry performance independent of each other? Which makes the largest 53 contribution to the prediction of inquiry performance? Questions Regarding Prediction of Inquiry Performance: 6. Is there a positive relationship between the number of problems sensed within categories of information sources used by a teacher-in-training in the in-basket and the weights she gives to those sources of information in her judgmental policy? The embedded problems in the teachers' in-basket are generally of two kinds: one, which could be sensed by attending to information within one category of information sources and the other, which involves a discrepancy between two bits of information within two different categories of information sources. Different kinds of questions might be asked regarding the second type of potential problem. 7. Is there a positive relationship between the amount of shifting in the in-basket situation and the assigning of equal weights to those categories of informa- tion sources in the subjects' judgmental policies? Since there are several "bits" of information within the various categories of information sources. one might also ask the following question: 8. Is there a positive relationship between the number of bits of information attended to within each category of information sources and the weights used in the subjects' judgmental policies? 5n However. since the total number of bits available within each information source is different a correction factor will have to be used. The foregoing questions have dealt with the relation- ships between different levels of training and subsequent judgmental tasks. They were also concerned with the validation of certain predictions of inquiry performance. Hypotheses: 1. The judgmental policies for the discussion and reflection training group will be different from the policies for the control group on subsequent policy making as determined by the post judgmental task. The policies of the in-basket plus discussion- reflection training group will be different from the policies for the control group on the judg- mental post test. The in-basket plus discussion-reflection training group will differ from the discussion-reflection only training group in their policy preferences on the post judgmental task. The two training groups, that is. the discussion- reflection training group and the in-basket followed by the discussion-reflection training group will differ on the post judgmental task from the pre test measure. 55 To determine the differences between groups due to the three treatments an analysis of covariance will be used. Thus. after controlling for pre test differences the differences on the post judgmental task due to training will be examined. This analysis will be followed by planned comparisons to study the effects of training in each individual group. Confidence intervals for each least square estimate of treatment effects will be computed. These intervals would show for which cues the differences were significant among the groups. 5. Judgmental policy variables will significantly predict inquiry performance as measured by the Teacher's In-Basket. The measures are: problem sensitivity, competence. information sources, bits and shifting. Time is a measure of commitment to inquire therefore would not be predicted by the cognitive policy variables. 6. There is a positive relationship between the number of problems sensed pertaining to SES, IQ. grades, sex and personality variables in the teacher's in-basket and the beta weights assigned to those cues in the pre judgmental task. 7. There is a positive relationship between the number of bits of information subjects attend to in the teacher's in-basket related to SES. IQ. grades. sex and personality factors and the weights 56 assigned to these variables in their judgmental policies. 8. There is a positive relationship between the amount of shifting in the in-basket and the equal weighting of the cue variables in the judgmental policies. That is, a subject who tends to take all information sources into account before reaching a judgment will tend to assign equal weights to all categories of information sources in the judgmental task. Such a person would tend to look for information in all categories of infor- mation sources in the in-basket and would therefore. tend to shift from one source to the other more often than a subject who just considered one or two cues to be important in making a judgment. 9. The contributions of personality variables and policy preferences are independent of each other in the prediction of inquiry. To validate the predictions made in the above hypo- theses regarding inquiry performance. correlational and regression analyses will be conducted. Predictions from judgmental beta weights and personality tests on inquiry performance will be analyzed using multiple linear regression analyses. The results of the experiment including tests of the above hypotheses and questions are presented in Chapter IV: and the interpretations of the results follows in Chapter V. CHAPTER IV RESULTS This chapter will report the basic findings of the present experiment. In order to aid the reader in inter- pretation. the following scheme will be used to inspect. organize and report the data. First, the effectiveness of training in achieving changes in judgmental policy prefer- ences will be explored in terms of the judgmental policy post test. The effects of discussion-reflection training and the in-basket experience on subsequent decision making will be carefully examined. Second, the hypothesis regarding the relationships between beta weights represent- ing judgmental policies and the inquiry processes in the in-basket will be examined in the light of the research findings applicable to them. Third, the relationships between the personality variables and the inquiry variables will be eXplored. Fourth. the relative contributions of personality variables and judgmental policy variables (beta weights) as predictors of inquiry performance will be inspected. Finally. individual policy profiles will be plotted and examined in order to describe the weighting patterns characteristic of the experimental and control groups on the post-judgmental task. 57 58 The Judgmental Policy Variableg The reader is reminded that all the subjects were requested to rate one hundred and eight cardexes which reported information about one hundred and eight hypothe- tical students regarding the SES, IQ, grades, sex and personality of the students. The ratings were to be made on a seven-point-scale on the following question: "What is the likelihood that this student will be an instructional problem in a classroom setting?" The information was presented in all possible combinations of the cue levels. The three levels of high, medium and low for SES, IQ and grades were coded 3, 2 and 1 respectively: and the two levels of sex and comments were coded 2 and 1 on the com- puter cards. That is. males were coded 2 and females were coded 1: and the favorable comments were coded 2, whereas. un- favorable comments were coded 1. This procedure was necessary in order to obtain a judgmental policy preference for each subject in the form of a regression equation. The regression weights in the equation indicated the amount of weight the subject gave to each of the cues in reaching a judgment. The same procedure was adopted in capturing judgmental policy preferences on the subjects at the end of the training period. Thus, pre and post policy preferences were obtained for each subject. 59 It was necessary to convert the raw regression co- efficients in the regression equations, obtained for each subject, into standardized beta weights in order to make comparisons among subjects, and between the pre and post judgmental policies of subjects in the three experimental groups. The pre judgmental beta weights were used as the co- variables in determining post judgmental policy differences in the three groups. The pre judgmental beta weights were also used as predictors of inquiry performance. The post judgmental beta weights were the dependent variables used to study the effectiveness of training in achieving changes in judgmental policy preferences. Due to the nature of the question on which the ratings were made; ahat is, the higher the rating, the greater the likelihood of the hypothetical student being an instructional problem): and the coding of the cue levels, (that is, the higher the IQ, etc. the less the probability of a student being an instructional problem), the beta weights were usually negative. Therefore, to eliminate the need to put negative signs repeatedly before the beta weights, the signs are reversed in all tables and graphs. In conducting the various analyses, however, the original signs of the beta weights were maintained on the computer cards. Finally, since the standardized beta weights were always below 1.0, they were multiplied by 100 for the sake of 6O convenience and were used in that form in all the tables, graphs and analyses. Reliability o§;the Instrgment To assess how well the linear model was representing 2'8 for the judgments the judgments of the judges the mean R was calculated and frequency distributions of the Ra's of the individual regression equations were plotted in the form of histograms for both the pre and the post judgmental tasks. The Rz's are the coefficients of determination which inform us of the amount of the variance accounted for in i the judgments by the five cue variables. Figure 3 shows 2's on the the frequency distributions and means of the R two administrations of the judgmental task. Except for a few extreme cases the instrument seems to be fairly reliable. The mean of the pre test Rz's is 60.26 (62.49 after ex- cluding the two extreme cases): whereas the mean R2 for the post test is 65.30. Judgmental Policies and Stability of Beta Weights Since the judgmental task was given again to all the subjects after a period of eight weeks, stability or reli- ability across time (for the same subject making the same decisions) were calculated. Table 4.l reports means, stan- dard deviations and coefficient of stability or reliabil- ities of beta weights of each of the five cues on the pre and post test measures. 61 Histogram of Rz's for 53 subjects on pre judgemental task whose mean is 60.26; and 62.49 after excluding : two extreme subjects. 1% ”J 3“’ .1 r MI Ho IL" h—‘fil :0 I“... 9“ L s 1' 1 * *—1 .——I— 0 ’3 g3 35 m ,5,» (,9 cm W 70 H30 Fig . 3A L Histogram of Rz's for 53 subjects on post ‘ judgemental task whose mean is 65.30. l .L H 0 7 Aio 3'0 40 50 £0 ’76 {0 9'0 ’20 Fig. 3B 62 TABLE 4.1 MEANS AND STANDARD DEVIATIONS OF STANDARDIZED BETA WEIGHTS X 100! AND STABILITY COEFFICIENTS OF BETA WEIGHTS OF FIVE CUES ON PRE AND POST TEST MEASURES (N=53) Variable 2 so Stability Pre SES ' .23 6073 $.05 Post SES 6.86 13.56 Pre IQ 5.27 18.07 Post IQ 13.41 18.70 .37 Pre Grades 48.35 37.59 Post Grades 60.41 25.70 .21 Pre Sex 1.56 7.66 Post Sex .90 9.61 .44 Pre Comment 32.65 24.54 Post Comment 27.70 21.77 ~35 For N = 53 the probability of a correlation of .27 occurring by chance = . occurring by chance = .01. 05: the probability of a correlation of .23 Figure 4 shows the average weights assigned to the five variables on the pre and post judgmental tasks. 63 Pre - a 10*L Post - X :0 ‘L 0'0 " .yo 4' . . Figure 4. Means of Beta Weights of five cues on pre and post test measures. (N=53) 64 Table 4.2 reports prediction of the five post test beta weights with the five pre beta weights as covariates using a regression analysis. Only the F's for post comment, post sex and post SES are significant. TABLE 4.2 REGRESSION ANALYSIS FOR THE PREDICTION OF POST BETA WEIGHTS FROM FIVE PRE BETA WEIGHTS AS PREDICTORS Variable Multiple Multiple Overall P Less R2 R F Than Post SES 0.24 0.48 2.76 0.029 Post IQ 0.17 0.41 1.84 0.124 Post Grades 0.11 0.33 1.07 0.386 Post Sex 0.32 0.57 4.31 0.002 Post Comment 0.26 0.51 3.23 0.014 Chi square = 52.458 DF = 25 P<:.0011 Although the correlations between the pre and post beta weights as shown in Table 4.1 are not very high, individual weighting policies as shown in the individual weighting profiles, pre and post, seemed to be highly consistent. To explore this further, each individual's weighting profile on the pre test was compared to the same subject's weighting profile on the post test. There appeared to be a high degree of stability between the two sets of beta weights. 65 It was therefore thought prOper to compute product-moment correlations for the pre and post training beta weights for each subject using the pre beta weights and the post beta weights as the two variables to be correlated with five observations each. The 53 correlations were then converted into Fisher Z's using the apprOpriate statistical tables. After averaging the Fisher Z's over all the subjects, the overall correlation between pre and post beta weights was found by converting the average value of the 2's back to a correlation coefficient by using the tables. This was found to be quite high (r = .89). This clearly shows that individual judgmental policies are stable over time and resistant to change. That is. the relative pattern 2; weights assigned across the five variables remains quite stable, although the magnitudes of partigular weights may not be stable. Judgmental Policies and Training One of the major interests of this study was to inves- tigate the effects of discussion and feedback training and the combined effect of the above training with the in-basket experience on subsequent decision making. fiypgthesi§:l: The judgmental poligig§,0f_tgachers;ig; training can be modified by giving them training using discussion and reflection in which the characterigtics of their judgmental system are made explicit and are related to 66 the judgmental tagk. One phase of the present study involved training subjects by giving them cognitive feedback, i.e., by showing them their weighting policies on the pre judgmental task. The actual weights each subject had assigned to the five cues in reaching a decision as to the likelihood of a ' hypothetical student being an instructional problem in a classroom setting were taken from the computer printout and given to the subjects. Thus feedback was provided to the subjects individually. The rest of the training was devoted to a discussion in small groups of the five variables and their effect on student behavior in the classroom. Hypothesis 2: The in-basket experience will serve to modify_judgmental_policies of teachepg-in-trainipg_on subse- quent judgmentgl_tasks. ThgggfpreL the policigs of the in- basketgplus discussion-reflgption training wil;_be different from the_policies for the control group on the post judg- mental task. Experimental Group III had the in-basket experience prior to the feedback and discussion training. Bypothesis 3: The in-basketpplus discussion-reflection training grogp_will differ from the_discussion-reflgctig_ gply training gropp‘in their policy ppeferences on the judgmental post test. To test the above hypotheses an analysis of covariance was done, followed by planned comparisons. The analysis of 67 covariance revealed that there were significant differences in all three groups in their weighting policies from pre to post tests (F = 19.079 P<:.0001). The post beta weights were significantly different from the pre beta weights on SES, IQ, grades and comments but the difference between the pre sex and post sex beta weights was not significant. The ' univariate F's for post SES, post IQ, post grades, post sex and post comments were 20.01, P<:.0001: 12.46, P<:.0010: 68.51, P9 I 137 Rate the following comments about student progress and growth in school as being favorable or unfavorable by writing the apprOpriate abbreviation (F for favorable. U for unfavorable) in the blanks provided after each of them: 1. Memorizes where reasoning should be used 2. Very fine student _____ 3. Not always dependable _____ a. Does work neatly 5. Low score on tests 6. Should develop power to concentrate 7. Shows interest and eagerness to improve 8. Lacks persistence 9. Listens carefully _____ 10. Finds worthwhile things to do _____ ll. Wastes time 12. Irregular attendance 13. Hard worker 1h. Capable of doing better work _____ 15. Follows directions and responds promptly l6. Prepares only part of work 17. Too many outside activities 18. Good attitude towards private and public prOperty l9. Able to meet new situations 20. Lacks foundation 21. Thoroughly reliable _____ 22. Fails to do home assignments 23. Works independently _____ 2h. 25. 26. 27. 28. 29. 30. 138 Careless or inaccurate work Required work late or incomplete Uses time and materials wisely Fails to follow directions Completes work on time Subject is difficult for student Poor attitude APPENDIX B Selected Pergonality Tests Student Opinion Survey Attitude Inventory Political Position 139 140 STUDENT OPINION SURVEY Name Age Date Education Instructions Below are a number of statements about various topics. ‘ They have been collected from different groups of people and represent a variety of opinions. There are no right or wrong answers to this questionnaire: for every statement there are large numbers of peOple who agree and disagree. Please indicate whether you agree or disagree by circling the statement to which you agree. Please read each item carefully and be sure that you indi- cate the response which most closely corresponds to the way which you personally feel. 1. A. Children get into trouble because their parents punish them too much. B. The trouble with most children nowadays is that their parents are too easy with them. 2. A. Many of the unhappy things in people's lives are partly due to bad luck. B. PeOple's misfortunes result from the mistakes they make. 3. A. One of the major reasons why we have wars is because peOple don't take enough interest in politics. B. There will always be wars. no matter how hard people try to prevent them. h. A. In the long run. peOple get the respect they deserve in this world. B. Unfortunately. an individual's worth often passes unrecognized no matter how hard he tries. 5. A. The idea that teachers are unfair to students is nonsense. B. Most students don't realize the extent to which their grades are influenced by accidental happenings. 10. 11. 12. 13. A. 141 Without the right breaks one cannot be an effec- tive leader. Capable peOple who fail to become leaders have not taken advantage of their opportunities. No matter how hard you try some people just don't like you. People who can't get others to like them don't understand how to get along with others. Heredity plays the major role in determining one' s personality. It is one's experiences in life which determine what they're like. I have often found that what is going to happen will happen. ' Trusting to fate has never turned out as well for me as making a decision to take a definite course of action. In the case of the well prepared student. there is rarely if ever such a thing as an unfair test. Many times exam questions tend to be so unrelated to course work that studying is really useless. Becoming a success is a matter of hard work. luck has little or nothing to do with it. Getting a good job depends mainly on being in the right place at the right time. The average citizen can have an influence on government decisions. This world is run by the few people in power. and there is not much the little guy can do about it. When I make plans. I am almost certain that I can make them work. It is not always wise to plan too far ahead because many things turn out to be a matter of good or bad fortune anyhow. 1h. 15. 16. 170 18. 19. 20. 21. 22. B. A. A. B. 142 There are certain peOple who are just no good. There is some good in everybody. In my case getting what I want has little or nothing to do with luck. Many times we might just as well decide what to do by flipping a coin. Who gets to be the boss often depends on who was lucky enough to be in the right place first. Getting peOple to do the right thing depends upon ability. luck has little or nothing to do with it. As far as world affairs are concerned. most of us are the victims of forces we can neither under- stand. nor control. By taking an active part in political and social affairs the peOple can control world events. Most people don't realize the extent to which their lives are controlled by accidental happenings. There really is no such thing as "luck". One should always be willing to admit mistakes. It is usually best to cover up one's mistakes. It is hard to know whether or not a person really likes you. How many friends you have depends upon how nice a person you are. In the long run. the bad things that happen to us are balanced by the good ones. Most misfortunes are the result of lack of ability ignorance. laziness. or all three. With enough effort. we can wipe out political corruption. It is difficult for people to have much control over the things politicians do in office. 23. 25. 26. 27. 28. 29. A. B. B. A. A. B. ...-:4 143 Sometimes I can't understand how teachers arrive at the grades they give. There is a direct connection between how hard I study and the grades I get. A good leader expects people to decide for them- selves what they should do. A good leader makes it clear to everybody what their jobs are. ~Many times I feel that I have little influence over the things that happen to me. It is impossible for me to believe that chance or luck plays an important role in my life. PeOple are lonely because they don't try to be friendly. There is not much use in trying too hard to please pe0ple. if they like you. they like You. There is too much emphasis on athletics in high school. Team sports are an excellent way to build character. What happens to me is my own doing. Sometimes I feel that I don't have enough control over the direction my life is taking. Most of the time I can't understand why politicians behave the way they do. In the long run. the people are responsible for bad government on a national as well as on a local level. 11m ATTITUDE INVENTORY This questionnaire is composed of 36 statements with which you will be asked to agree or disagree. For each state- ment. respond according to the following key: (1) True (2) False Please proceed through the inventory quickly. and respond to every item. ‘1. 10. 11. 12. I like to have a place for everything and everything in its place. Some of my friends think that my ideas are impracti- cal. if not a bit wild. I don't like to undertake any project unless I have a pretty good idea how it will turn out. For most questions there is just one right answer. once a person is able to get all the facts. Politically I am probably something of a radical. Perfect balance is the essence of all good composi- tion. I prefer to engage in activities from which I can see definite results rather than those from which no tangible or objective results are apparent. I find that a well-ordered mode of life with regular hours is not congenial to my temperament. The unfinished and the imperfect often have greater appeal for me than the completed and the polished. I like to listen to primitive music. I have always had goals and ambitions that were im- practical or that seemed impossible for me to realize. When a teacher lectures on something other than what he originally announced. I feel uneasy. 13. 1a. 15. 16. 17. 18. 19. 20. 21. 22. 23. 2“. 25¢ 26. 27. 28. 29. 30. 145 Trends toward abstractionism and the distortion of reality have corrupted much art of recent years. It bothers me to have different news commentators give different interpretations of the news. The sign of a good teacher is the ability to teach a class spontaneously. without careful preparation. I like to fool around with new ideas. even if they turn out later to have been a total waste of time. I don't like to work on a problem unless there isa possibility of coming out with a clear-cut unambig- uous answer. I have always hated regulations. The give-and-take of a class discussion is usually much more rewarding than a lecture. Many of my friends would probably be considered unconventional by other peOple. I like classes in which notes can be easily taken. It doesn't bother me when things are uncertain and unpredictable. Nothing is more infuriating than an instructor who jumps around among topics and never sticks to the point. My way of doing things is apt to be misunderstood by others. I value courses that provide an abundance of meaning- ful factual material. Facts appeal to me more than ideas. Small discussion groups often leave me with a feeling of dissatisfaction concerning the way time was spent. I have had strange and peculiar thoughts. I don't like things to be uncertain and unpredictable. The worst thing an instructor can do is to make very specific plans for each lesson. 31. 32. 33- 34. 35. 36. 146 It is a good rule to accept nothing as certain or proved. I dislike following a set schedule. Usually. I prefer known ways of doing things rather than trying out new ways. I like to go alone to visit new and strange places. I much prefer friends who are pleasant to have around to those who are always involved in some difficult problem. I have had very peculiar and strange experiences. 147 POLITICAL POSITION To the best of your knowledge. what are (were) the predominant political leanings of your parents? Please circle the letter corresponding to your answer. a. Democratic b. Republican c. Independent d. Other (specify) Politically speaking. would you consider yourself: (circle) a. Quite conservative b. Somewhat conservative c. Middle-of-the-road d. Somewhat liberal e. Quite liberal Rank your own personal preference for the following political figures. were they all to be candidates for :he presidency in the same election. Rank from 1 "' 0 George McGovern Hubert Humphrey Richard Nixon Barry Goldwater APPENDIX C Cardex - Rating Scale 148 149 DIRECTIONS Please read the instructions carefully. If you have any questions ask them before you start. Once you have begun no questions can be answered. Check the yellow cards with the booklet to make sure the numbers correspond. Once you have made your choice and marked the rating scale on the first page of the booklet turn both the card and the page over. Do not look back at what you have done. but continue to progress forward. **__w CARDEX RATING SCALE From the information provided on the Cardexes. using a seven point scale. rate each of the students on the following question: "What is the chance that this student will be an instructional problem in a classroom setting"? SAMPLE SCALE 5% 50% 95% Low High 1 2 3 4 5 6 7 If you feel that the student shows a 95% or greater chance of becoming an instructional problem. place a check- mark at 7. If you feel that the student has a 50% chance of becoming an instructional problem. place the checkmark at 4. If you feel that the student will have 5% or less chance of develOping instructional problems. place the checkmark at 1. APPENDIX D Sample Cardexes 150 151 PERMANENT ELEMENTARY SCHOOL RECORD Guardian (:Z(zf 7L 71;? flair-t’c/VL ix // JV ” ' ' ”/'-\ Name a ,L,/L,Q/gr/,.\ \y';,;fi (2 “f7. N,” Sex 1/ (14ch / Average Grade _ C <, ,- , , Father's Occupation £111" 1’ 7L,ZL',:¢1/é IQ .MQ _‘ ,. 1 , '_ J/ , , , Comments - -.. ' f7 - ' 2 .1 PERMANENT ELEMENTARY SCHOOL RECORD Guardian_ __E//J [@2721 14 (fix/11, Name (1)17; 7:72 c z 1.2 /C(1¢{l黀,c my '77 .fi/mu Average Grade [L Father's Occupation K 7414 WW (2 152 PERMANENT ELEMENTARY SCHOOL RECORD Guardian 57 Average Grade [8” - -' .. .4 ‘ ' Father's Occupation ZW’ IQ A»? / '~...-,."$/‘ ‘ Z— ,. Comments #71442 (>ng PERMANENT ELEMENTARY SCHOOL RECORD Guardian fl? 74,1 (4'6 fl/Z'Iégi/ IQ / 41" Name fl 7'1, )1. [ML] t" ’7‘LJ S ex ’ 'jazfltlc Ala/c. Average Grade (C " Father's Occupation p; "422:2 c.4224.) Comment ”flit/t [z'XQZI/Mfd jb: 1:41:4{11 APPENDIX E Discussion — Reflection Training 153 154 DISCUSSION-REFLECTION TRAINING Name Date The purpose of this training is to increase our under- standing of the sources of information that can help teachers to recognize or solve student-problems in the classroom.. . ’ ‘ Can anyone tell why it is important to understand what such factors as SES. IQ. Grades. Sex of the child and Comments by teachers mean on student's records? (Ask individuals). Are these bits of information about students necessary in order to understand students' problems? Why? (Ask individuals). Before we begin the formal training I'd like you to look at your regression equation which tells you how you actually weighted these sources of information when you made judgments regarding the likelihood that the students. presented to you on the yellow cardexes, would have problems in a classroom setting. (Pass out sheets with S's beta weights). You weighted SES . IQ , Grades Sex . and Comments . 155 I. First let us see what SES means: Socioeconomic status and social class are closely re- lated. The lower the income the lower the social class. Other factors contributing to the definition of social class would be the educational level and the occupa- tion of the individual. Although SES can be divided into several divisions. for our purposes three categor- ies would be sufficient. That is. children in schools might come from high. medium or low socioeconomic status families. There is evidence to show that opportunities and re- wards in life are unequally distributed - more good things go to the children of higher - status families. A child who is from a low socio-economic status has many barriers to hurdle in order to achieve happiness and progress in school. The plight of the child who has both low academic ability and low SES is extreme. Children from low SES families usually: have 1. Poor and unsafe housing. 2. Lack of books. magazines and educational materials at home. live 3. Poor neighborhoods - no opportunities for in recreational and creative activities. have 4. Poor attitudes in the home towards education. 5. Lack distinct and complex verbal stimulation and good models. 6. Peer group pressures encouraging delinquency. go to 7. Unresponsive schools. The child learns he is a social outcast. Text books and instruction in general lack meaning and relevance. and 8. Discrimination faces These environmental factors contribute to the various charac- teristics of the low SES child in the cognitive. affective and psychomotor domains which are as follows: They have: II. 156 l. Perceptual and language deficit. Their speech and thought processes are restricted. simple repetitive and disconnected. They have poor vo- cabularies. 2. Depressed intellectual development as represented by IQ scores. . Low achievement scores. . Inadequately developed prosocial values. 3 u 5. Poor self concept. 6. Emotional problems due to neglect and broken homes. 7 . Lack of saleable skills. (Discuss each point with group) Sppond. let us see whgt Igymeans: Scholars do not agree on the nature of intelligence. however. IQ scores as measured by tests of general intelligence have been used for years in schools to predict academic achievement. that is. levels of achievement in Mathematics. English and other academic subjects. An IQ can therefore be referred to as academic aptitude. In the standardization of the Revised Stanford-Binet Scale. one of the two individual intellignece tests most widely used in America. a range in IQ from 35 to 170 was reported. The majority of the standardization group had IQ'S between 84 and 116 (approximately 68 percent). About 14 percent had IQs between 116 and 132. and approximately 14 percent had IQs between 84 and 68. 2 percent above 132 and 2 percent below 68. At the top end of the scale. 145 and higher. we expect superior performance in all types of academic work. Those in the 130 to 145 range also are predicted to do very well. Occasionally. however. students in this 130-145 group do produce relatively mediocre perfor- mances. Among the 68 percent in the range of 84 to 116. we expect much variation in achievement: for example. those with 103 of 100 would vary widely in the level of reading and arithmetic achievement. We would expect children with 103 below 70 who come from en- riched homes and neighborhood environments not to perform well in the academic subjects. Further. those below 55 will probably not do well in any school tasks 157 including those heavily based on psychomotor abilities. We would anticipate that children with IQs of 35 and below would require help throughout life with the simplest tasks of eating. dressing. keeping clean and the like. Children showing a high need for achievement. competi- tive striving and curiousity have shown gains in IQ scores. Limitations of IQ scores in Organizing Instructional Programs. 1. IQ scores will not be useful in assessing readiness of individuals for typing. art. music and physical education. 2. Predictions made from IQ scores do not apply to children from impoverished environments. IQ scores are not equally accurate and valid for measuring intelligence of students from all types of homes and neighborhood backgrounds. 3. IQ scores are not fixed. They can change 15 points or more for more than half the students. 4. Children having the same IQ may vary widely in levels of achievement in academic subjects. 5. Different IQ tests result in somewhat different IQ scores for the same individuals. 6. IQ scores do not correlate highly enough with a- chievement in various school subjects. therefore should not be used as the sole basis for predicting or categorizing children as slow learners. dis- advantaged. or unsuited for later college work. III. Third. let us see what grgdes mega: Accurate information about a student's present level of achievement in any curriculum area. for example. in reading or science. is useful in predicting how he will do in the future in the same curriculum area. It is more useful actually than an IQ score. Increasingly. instructional programs are being organized for partic- ular students. not on the basis of IQ scores or other specific ability tests. but on the basis of the student's present level of achievement. The two main sources of information about level of achievement that are available to teachers are results of standardized educational achievement tests and teacher-made tests or other teacher-develOped procedures. IV. 158 Precautions: 1. Achievement for one student varies widely from one subject to another. 2. Grades vary from one test to another and depend on the difficulty level of the tests. 3. Grades are not always reliable measures of a student's achievement. Unreliability can result from physical and motivational conditions of the student. Unreliability can also result from sub- jectivity in scoring on the part of the teacher. 4. Grades are not always valid. They are a measure of something other than what the teacher says they measure. 5. Grades of Ds and Es produce a sense of failure and frustration in the student that may result in maladaptive behavior in the classroom. Fourth. let us see how the sex of the child can affect his performance in the classropm. Although sex roles of males and females in American society are not crystallized certain sex differences are observed. Some differences are due to biological differences in physical growth and maturation. Other sex differences are due to sex-typing and the identi- fication of children with adults of the same sex. Further, the reinforcement of behaviors that are acceptable within a culture such as aggression in males and dependency in females might encourage the differen- ces between the two sexes. Sopgysex differences: 1. Girls mature faster than boys in physical growth. however. the average height and weight of boys is greater than girls except at ages 11 to 15. 2. Girls typically score higher on verbal items and boys on quantitative and spatial items in both intelligence and achievement tests. 3. Sex differences in preference for play. games and other activities become apparent in early child- hood and increase with successive age levels. Girls prefer to read where as boys engage in active games. 159 4. Males show a higher degree of aggressive and dominant behavior than females. Delinquency is higher in boys. Females are more co-operative than males. Girls comply with the teachers wishes more frequently than boys. Precaution: There are large individual differences among individuals of the same sex. V. Fifth. let usggee‘whgthpments on students' records mean: Comments about students. are written by the teacher to describe the affective characteristics and work habits of the students. Comments are usually descriptions of the motives. attitudes. interests and values of the student. Many studies and firsthand reports of teachers could be cited to show wide variability among students of the same age in affective characteristics. 1. Individuals vary widely in the need to achieve success and related need to avoid failure. This variability markedly influences their tendency to undertake or avoid activities. Other motives are the need for love and belonging. for self esteem and for self-actualization. 2. Honesty is one of the prosocial values that students differ in. Others are punctuality. dependability. sincerity. orderliness. conformity to group norms. inhibition of aggressive impulses. enjoyment of study. interest. respect and desire for freedom of self and others. Other differences are: 3. Attitudes towards school. towards others and them- selves. 4. Identification with values and practices of the older generation. 5. Interest in work and study. Positive affective characteristics greatly influence the intellectual development of students. APPENDIX F Scoring Sheet for IN-Basket 160 161 A.) c (1) m E P E . m 0 'Q o IEQOI QSHSBH-UI. dew seoueaegag Kaeqeaoes paooeg teoIpew eIIg~queuemaeg meaBoIoog SGIBOS‘QDS dLO dIO eqopoeuv x003 eouepuequ xepaeo paequaodaH <1) E H E—l $4 (1) +> m 5:: E (D Q) 4.) E H o 'r-1 ,0 (1) $4 $40 "-3 (D 0L: .0 Cl. "—30-: :3 >< (U U) LL] 2 APPENDIX G Competence Scoring Key 162 Cooper 1. 2. Hoffman 1. 2. LOpez 163 COMPETENCE SCORING KEY Birthday. only 3 signed up. Won games day. RC A's in Physical Education. Sociogram: isolate. Youngest in class. Sociogram and CTP discrepant. ill Fridays. has all brothers. 1 t Misses dancing regularly. not popular. socially immature. not too feminine in her role. ’ r Low RC. low achievement scores. low CTP. not pOpular. To psychologist or remediation. Absent and tardy. dissatisfied with school. transferred from Detroit. Unstable family. divorce. remarriage. Stepsister in same class. Frustrated can't compete with stepsister. emotional and family problems. Poor RC. low CTP and low subscore. Grades discrepant with IQ (underachiever). sociogram isolate. Emotional problems. low CTP. isolate. sister coming. Stuttering. connected with emotional problems. Sister is a star. can't compete. Home pressure getting worse because of sister's visit. Low RC. low cardex, low CTP (to psychologist). almost isolate. Maloney 1. 2. Rosen l. 2. 3. Sieminsky 1. 2. 3. 164 Migrant. transfer student. absent and tardy. Intelligent. IQ 151 but an underachiever. may have language problem. Illiterate parents. she's so smart. low educated parents. School not valued in her home so lacks motiva- tion. Low RC. except reading. same for achievement. reading high. low CTP. overweight. Overweight. low CTP. D in Phys. Ed.. Sociogram dyad. Emotional problems connected with overweight. High reading grade. maybe reads alot alone. Give extra help in math from anecdote about math. Other grades o.k.. "D" in math. Band meets same time as math. RC excellent. CTP high. President of class. Bored. causes disorder. Sociogram: pOpular. enrichment needed. Excellent in everything. not accelerated. strange. small. thin boy. Low RC. absent for CTP. no field trip slip in. popular. Behavior problem and yet popular. disrupts class. Unstable family. father deserted and returned. parents quarrel. 165 4. Discrepancy between parent's education. Home problems in relation to RC and good art grade. 5. Parents compete. father aggressive with Stu. Stu identifies with mother. fears father. beaten by father. Fagen-Moore 1. Either one because other isn't seen. Negro in class, "crossburned. Nigger lover". 2. Two William's. 3. Graves and Fagen are brother and sister (either result o.k.). 4. Lives in integrated neighborhood. Moore is Negro. has lived here a longer time. APPENDIX H Analysis of Covariance 166 167 ANALYSIS OF COVARIANCE DIFFERENCES IN POST JUDGMENTAL TASKS IN THE THREE TREATMENT GROUPS DUE TO TRAINING F-RATIO FOR MULTIVARIATE TEST OF EQUALITY OF MEAN VECTORS = 19.079 D.F = 5 and 41.00 P 0.0001 Variable Univariate F P 1. Post SES 20.01 0.0001 2. Post IQ 12.46 0.0010 3. Post Grades 68.51 0.0001 4. Post Sex 0.74 0.3956 5. Post Comment 22.09 0.0001 APPENDIX I Sample Imbedded Problems In Teacher's IN-Basket. 168 VI. 169 SAMPLE OF IMBEDDED PROBLEMS Juanita Lopez 0. General as b. Lives in migrant cottages Remedial? c. Psychologist? 1. ROC. as *b. Ce d. *e. f. Poor student with varied grades Often absent and tardy Poor attitudes toward school D in phys. ed. 12 yrs. old -- oldest in class Underachiever (high IQ) 2. Cardex a. b. *c. d. e. f. IQ is 151 (mistake) WISC is different test than other kids had Overaged Transferred from El Paso Low achievement scores Father born in Mexico 3. Cum File a. b. Cs d. e. f. *g. h. 1. 4. Rec. *a. 5. CTP a. b. WISC should be 121 Social promotion in 3rd grade Flunked 4th grade K.A. in grade 4 is discrepant with WISC Did better in 5th grade Large family Parents are migrants Parents are illiterate Possibly trouble with English language Book Many absences Scored 8O Subscored are low 6. Sociogram a0 b. Mutual choice with Maloney Picks boy for second choice 7. Medical a. First immunization shots when she entered school 170 William Fagen 0. General a. Related to Mary Beth *b. Is he a Negro? c. Stepsister doing well by comparison d. Remedial? e. Psychologist? 1. R.C. a. Very poor grades b. Unsatisfactory deportment c. Signed by Graves *d. Absent and tardy a great deal 2. Cardex a. CTMM test given in 6/54 b. He transferred from Detroit c. Both parents work d. Achievement scores are low e. Mother remarried f. Low IQ (85) 3. Cum File a. Never been a good student b. Good attendance in the past 4. Rec. Book a. Numerous absences and tardies 5. CTP a. Scored 82 b. Low subscores 6. Sociogram *a. Mutual choice with Terry *b. Chosen by Mary Beth 7. Medical a. Underweight and short b. No big weight gain since 1961 8. Anecdotes *a. Who is B.H. b. Misspelled his name *0. William beat up Terry--who is his best friend d. Neither would say why it happened e. What happened at principal's office APPENDIX J Judgmental Policy Summary Sheets. 171 1mflrn’p (.1-‘1; ' ‘V 172 JUDGMENTAL POLICY PROFILES FOR INDIVIDUAL SUBJECTS Judgmental Tasks Subject_Npmber_l USING STANDARDIZED BETA WEIGHTS Judgmental Tasks Subject Number 2 Pre Post Pre Post SES -.02 .03 SES -.02 -.08 IQ -.01 -.05 IQ -.14 0 Grades -.40 -.76 Grades -.65 -.68 Sex -.06 -.03 Sex .02 .02 Comments -.57 -.35 Comments -.52 -.25 Subjgct Number_3 Subject Number 4 Pre Post Pre Post SES .08 -.05 SES .02 -.04 IQ .02 0 IQ .03 -.30 Grades -.77 -.82 Grades -.64 -.75 Sex .05 -.06 Sex -.04 0 Comments -.28 -.30 Comments -.49 -.15 Sgbject Number_§ Subject Number_é Pre Post Pre Post SES -.O3 -.08 SES -.O3 0 IQ -.08 -.06 IQ .10 .24 Grades -.81 -.69 Grades -.34 -.40 Sex -.03 -.06 Sex -.13 -.05 Comments ~.O9 -.02 Comments -.61 -.62 Juggmental Tasks Subject Number 2 173 Juggpental Tasks Subject Number 8 ... 11.1.“! .— A‘l' .r" Pre Post Pre Post SES -.05 -.03 SES 0 -.05 IQ .29 .09 IQ -.21 -.19 Grades -.76 -.83 Grades -.89 -.90 Sex -.08 -.05 Sex -.06 -.05 Comments -.23 -.25 Comments -.06 .01 Subject Number 2 Spbjgct Number 10 Pre Post Pre Post SES .21 .16 SES .01 O IQ -.27 -.03 IQ .05 .26 Grades -.43 -.16 Grades -.03 -.48 Sex .18 .42 Sex -.05 -.04 Comments -.36 -.41 Comments -.21 -.54 Subject Nggperyll Spbjgct Numbery__ Pre Post Pre Post SES .01 -.10 SES -.01 .05 IQ .21 -.22 IQ -.11 -.06 Grades .73 -.73 Grades -.67 -.75 Sex .10 -.02 Sex -.01 .01 Comments .10 -.30 Comments -.48 -.44 Subject Nppber_13 Subject Number_14 Pre Post Pre Post SES -.08 -.02 SES -.01 -.Ol IQ -.27 .23 IQ o -.24 Grades -.34 -.50 Grades 4.61 -.54 Sex .25 0 Sex -.02 -.11 Comments -.35 -.30 Comments -.44 -.35 * Juggpental Tgsks Subject Number 15 Judgmental Tasks Subject Number 16 Pre Post Pre Post SES -.O6 -.21 SES .Ol -.17 IQ -.28 -.30 IQ -.13 -.23 Grades -.66 -.60 Grades -.77 -.81 Sex -.04 .08 Sex -.03 .07 Comments -.32 -.27 Comments -.35 -.23 Subjgct Ngmber_lz Subject Number 18 Pre Post Pre Post SES .03 -.03 SES .01 0 IQ -.07 -.O7 IQ -.12 -.23 Grades -.65 -.82 Grades -.75 -.65 Sex .04 -.01 Sex -.03 -.07 Comments -.47 -.28 Comments -.25 «.18 Subject Npmpertlg Subject Number 20 Pre Post Pre Post SES .0? -.08 SES -.11 -.O3 IQ .03 .05 IQ -.02 -.05 Grades -.65 -.82 Grades -.10 -.45 Sex -.08 0 Sex .04 .11 Comments -.39 -.24 Comments .02 -.53 Subject Number 21 Sgpject Number 22 Pre Post Pre Post SES -.O3 .12 SES .06 -.33 IQ .30 -.15 IQ -.O7 .04 Grades -.47 -.72 Grades -.08 -.29 Sex -.02 -.07 Sex 0 -.03 Comments -.38 -.30 Comments .-.62 -.53 175 Ju ental T sks Subject Number 23 Jugggental Tasks Subject Number 24 Pre Post Pre Post SES -.01 -.16 SES -.Ol .07 IQ .08 .02 IQ -.07 -.30 Grades -.61 -.57 Grades -.67 -.57 Sex -.03 -.13 Sex -.07 -.03 Comments -.36 -.23 Comments -.45 -.31 Subject Number 25 Subject Number 26 Pre Post Pre Post SES -.02 .04 SES -.04 .01 IQ -.02 -.05 IQ .04 -.40 Grades -.11 -.81 Grades -.67 -.72 Sex 0 -.14 Sex -.09 -.05 Comments -.80 -.10 Comments -.29 -.07 Subject Numbgr 27 Subject Number 28 Pre Post Pre Post SES .13 -.29 SES 0 -.60 IQ -.12 -.16 IQ -.11 -.04 Grades -.06 -.52 Grades .71 -.55 Sex .09 .19 Sex .04 .01 Comments -.59 -.58 Comments .30 -.06 Subject Number 29 Subject Number 39 Pre Post Pre Post SES 0 .02 SES .03 -.O7 IQ -.16 -.27 IQ -.11 -.09 Grades -.82 -.82 Grades -.70 --47 Sex -.12 -.07 Sex -.09 --09 Comments -.01 -.05 Comments -.52 -.44 176 Judgmental Tasks Judgpental Tasks Subject Npmber3l Sub ect Number 2 Pre Post Pre Post . SES .02 -.3O SES .03 -.01 IQ -.63 -.43 IQ .25 .22 Grades -.35 -.59 Grades -.50 -.83‘ Sex -.09 .01 Sex -.18 -.11 Comments -.18 -.12 Comments -.28 -.17 Subject Number 33 Subject Numbery34 Pre Post Pre Post SES .04 -.33 SES. .07 .03 IQ .17 -.40 IQ .02 .03 Grades -.74 -.45 Grades -.28 -.88 Sex .03 .23 Sex -.05 -.01 Comments -.31 -.42 Comments -.68 -.05 Sgbject Numbery35 Subject Number 36 Pre Post Pre Post SES -.08 -.08 SES -.02 -.28 IQ -.11 -.48 IQ -.O6 -.37 Grades -.65 -.68 Grades -.73 -.59 Sex -.02 -.01 Sex 0 .06 Comments -.42 -.13 Comments -.40 -.20 Subject Number 32 Subject Number 3S Pre Post Pre Post SES 0 -.10 SES -.09 -.O3 IQ -.11 -.18 IQ -.15 -.33 Grades -.58 “~63 Grades -.80 -.82 Sex -.06 -.11 Sex -.06 .03 Comments -.53 -.41 Comments -.24 .05 Judgpental Tgsks Subject Number 32 Judgpental Tasks Subject Number 40 Pre Post Pre Post SES .09 .05 SES .12 -.21 IQ .21 .03 IQ .15 -.43 Grades -.39 -.38 Grades .66 -.60 Sex -.02 -.15 Sex .07 -.03 Comment -.46 -.54 Comments .48 -.33 Subject Numbgr 41 Subject Number 42 Pre Post Pre Post SES .0? -.27 SES .0 .02 IQ .02 -.44 IQ -.09 -.29 Grades -.23 -.42 Grades -.63 -.44 Sex .16 -.03 Sex -.08 -.03 ‘ Comments -.62 —.46 Comments ‘-.55 -.61 Subject Number 43 Subject Numbp£_44 Pre Post Pre Post SES .0 -.06 SES -.05 -.23 IQ -.05 -.10 IQ -.44 -.31 Grades -.70 -.53 Grades —.74 -.80 Sex 0 -.06 Sex -.04 .05 Comments -.51 -.40 Comments -.l3 -.16 Subject Number 45 Sppjggt Numbgp_4§ Pre Post ‘ Pre Post SES -.18 0 SES -.09 -.18 IQ .05 -.29 IQ -.10 -.16 Grades -.27 -.29 Grades -.80 -.86 Sex 0 -.01 Sex .04 .03 Comments -.39 -.47 Comments -.16 -.09 178 Juggmental Tasks Subject Number 42 Judgmentgl Tasks Subject Number 48 Pre Post Pre Post SES .0? .05 SES .01 .04 IQ .06 .05 IQ -.13 -.26 Grades -.66 -.51 Grades -.85 -.88 Sex -.12 -.01 Sex -.06 -.01 Comments -.39 -.66 Comments .07 .04 Subject Numbgr 49 Subject Number_50 Pre Post Pre Post SES 0 -.02 SES .03 .02 IQ -.06 -.02 IQ -.14 -.37 Grades -.89 -.84 Grades -.61 -.39 Sex -.O9 -.14 Sex -.03 0 Comments -.14 -.23 Comments -.49 -.46 Sgbject Number 51 Subject Number 52 Pre Post Pre Post SES .03 -.O6 SES .06 O IQ .04 -.13 IQ -.49 -.59 Grades -.80 -.56 Grades -.49 -.50 Sex -.02 -.01 Sex -.04 -.02 Comments -.23 -.54 Comments -.28 -.20 Subject Number 53 Pre Post SES -.12 0 IQ -.27 -.40 Grades -.58 -.53 Sex -.03 -.09 Comments -.34 -.37 APPENDIXZK Raw Data 179 180 3.0 H.3 3HH mHH mm Hm. 0.N m.m mm mm 3m 00. 0.m m.m mm mHH mMH mm. 0.N H.m 0mH m0H 00H 05. 0.0 m.H ow m0 0H mm. m0 H.m um 0mH mm mm. H.0 0.m m3 00 mm Hm. 3.H 0.5 mmH 00H H0 H5. m0 0H t mmH 3H mm. 0.0 m.N mm hm 3H mm. 3.0 H.3 mm on 33 30. m.H 0.N 30H m0H mm 0». 3.0 3.3 me 0NH Hm mm. 3.0 H.m H0 MMH mm pm. H.N 3.m mm 0NH mm H3. H.0 m.m MHH 00H 03 om. 0.H m.H 30 >0 mm Hm. 0.m m6 N3H mm mmH mm. 0.3 m.0 0MH 0MH um 00. 5.0 5.3 mm mHH mm 30. N.0 0.N 03 0HH mm NH. mOCmpmqeoo mwohdom UHm MEHB hPH>HDHmC®m OHpmm DMHEW Gmmz deué EmH DOLHnm UQPOGHFOO HI mqm