SOME RZELATISNSHIPS BEWEEN' ANXIETY, CBGNITEYE STYLE. AND.- .E’RQBLEB! SOLVING T512555 for H16 061ng 0?. M A.- 3%}. Lain“! STATE UN‘VERSITY Leora Dairyi' ’lhomander _ i972 my LIBRARY Michigan State University i l 1 E i ._. “7‘. -l. |‘\ 3/ . O f .. . V! ‘I .3. C 1-. . ABSTRACT SOME RELATIONSHIPS BETWEEN ANXIETY, COGNITIVE STYLE, AND PROBLEM SOLVING BY Leon Darryl Thomander In the present investigation 114 male undergraduates were assessed for academic ability and achievement, the presence of three types of anxiety (test anxiety, trait anxiety, state anxiety), the use of three different cognitive styles (flexible-rigid, impulsive-reflective, ana- lytic-global), and.performance on two types of deductive reasoning prob- lem.solving tasks. The subjects were divided into two groups of 57 and were tested during two sessions spaced one day apart. A completely counterbalanced 2 X 2 Latin Square design was used. One group of sUb- jects experienced.nonstress experimental conditions in the first session and stress conditions in the second session, while the reverse order of treatment conditions was used with the other group. Four general questions were asked. (1) Will the measured anxiety, cognitive styles and strategies, and problem solving efficiency of sub- jects vary systematically with changes in psychologically stressful con- ditions? .Although the treatment conditions were psychologically stress- ful enough to produce differences between groups in state anxiety, they did not significantly affect problem solving efficiency, problem.solving Leon Darryl Thomander strategies, or performance on cognitive ability tests used as measures of cognitive styles. (2) Will subjects who are measured as more anxiety prone tend to exhibit different patterns of cognitive styles and strategies than less anxiety prone subjects? The original set of data was reduced to 26 var- iables by means of several cluster analyses. The intercorrelations of these 26 clusters showed that high anxiety prone subjects tended to score lower on measures of academic ability, be less flexible as measured by a flexibility-rigidity scale, and while working on the problem solving tasks, they tended to lack carefulness, have poor concentration, dislike working on the problems, become frustrated, and make several strategy changes. These findings are consistent with the hypothesis that the worry component of anxiety interferes with performance by taking attention away from the task. Poor concentration, lack of carefulness, and strategy changes may all be indications of inability to maintain attention toward the task. (3) Will subjects be measured as using the same cognitive styles consistently under different experimental conditions and across time? This question was not answered because the two kinds of style measures that were used in both sessions--cognitive ability measures and.problem solving questionaire measures--were f0und to be inadequate fbr use as indicators of styles as they are usually defined. (4) Will the different measures of cognitive styles be shown to have construct validity? Neither the problem solving questionaire measures nor the cognitive ability measures of style used in the present study were fbund to have convergent or discriminant validity as style measures. The findings indicate that if cognitive styles do exist as individual traits, they are very difficult to identify, and that some Leon Darryl Thomander instruments which have been used to assess them are of doubtful utility for that purpose. Although general cognitive processing characteristics that fit the usual definitions of cognitive styles were not fbund, cer- tain.more specific aspects of processing were identified. These were called cognitive strategies and were fbund to have some generality across time and problem types. SOME RELATIONSHIPS BETWEEN ANXIETY, COG‘JITIVE STYLE, AND PROBLEM SOLVING By Leon Darryl Thomander A THESIS Submitted to Michigan State university in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1972 ACKNOWLEDGEMENTS I would like to thank the members of my thesis committee for their help and guidance. In particular, I appreciate Richard Marshall, the committee chairman, fbr his guidance in planning the study, for many helpful suggestions on all phases of the project, and for encouraging me to think and write with greater precision and clarity. I want to express gratitude to John Hunter for the countless hours he spent giving invaluable help with the experimental design, statistical analysis, and interpretation of the results. Thanks also go to Donald Johnson for thoroughly reading the manuscript and for his helpful suggestions. Finally, I thank my wife, Nanci, for typing the final copy and for moral support throughout. ii TABLE OF CONTENTS LIST’OF TABLES .......................... SECTION I. INTRODUCTION ....................... Anxiety ........................ Cognitive Styles and Strategies ............ II. BBKDIDD ...... . . . . ................ III. (DGNITIVE STRATEGIES USED ON THE PROBLEM SOLVING TASKS IV. VI. Method ......................... Results ........................ Discussion ................ . ...... TREATMENT EFFECTS ON STATE ANXIETY, ABILITY MEASURES OF COGNITIVE STYLE, PROBLEM SOLVING STRATEGIES, AND PROBLEM SOLVING EFFICIENCY ............... Discussion ....................... PRACTICE EFFECTS ON STATE ANXIETY, ABILITY MEASURES OF COGNITIVE STYLE, PROBLEM SOLVING STRATEGIES, AND PROBLEM SOLVING EFFICIENCY . . . . . .......... thhod ......................... Results ........................ Discussion ....................... 'DHE CONSTRUCT VALIDITY OF COGNITIVE ABILITY MEASURES OF COGNITIVE STYLES ................... . Method ......................... Results ........................ Discussion ....................... iii Page vi 11 17 17 19 36 37 42 47 49 50 54 S6 57 59 69 SECTION Page VH1. PROFILE DIFFERENCES BETWEEN HIGH AND LOW ANXIOUS SUBJECTS . 72 Method . ................. . ...... 72 Results ........................ 73 Discussion ....................... 75 VIII. SUMMARY AND CONCLUSIONS .................. 78 LIST OF REFERENCES ........................ 82 APPENDICES A. PREMEASURES Reflectiveness Scale .................. 86 Alpert-Haber Achievement Anxiety Test ......... 87 Flexibility Scale ................... 88 STAI-Ttait Scale ................. . . . 89 B. INSTRUCTIONS Nonstress, Group 1, Session 1 ............. 90 Stress, Group I, Session 2 ............... 91 Stress, Group II, Session 1 . ............. 92 Nonstress, Group II, Session 2 ............. 93 C. PRACTICE PmBLEMS Nonstress ....................... 94 Stress ......................... 95 D. COGNITIVE ABILITY MEASURES OF COGNITIVE STYLES verbal Problems Test (Fbrm B) . . .......... 96 Hidden Figures Test (Form B, page 1) .......... 97 Sign Changes Test (Form.A, page 2) ........... 98 Object USes Instructions ................ 99 Object USes (Form A) .................. 100 Matching Familiar Figures Test (Form B, item # 4) . . . 101 E . PROBLEM SOLVING TASKS Malice and Alice (Mysteinroblem, Form A) ....... 102 Murder in the Family (Mystery Problem, Form B) ..... 103 The Hotel Room Problem (Math Problem, Form A) ..... 104 The Horse Trading Problem (Math Problem, Form B) . . . . 105 iv APPENDICES Page F. PROBLEM SOLVING QUESTIONAIRES Mystery Problems Questionaire ............. 106 Math Problems Questionaire .............. 108 G. REVISED STAI-STATE SCALE ....... . ......... 110 H. CLUSTER TABLES ................. . . . . . 111 LIST OF TABLES Table Page 1. Procedures and Design ........ . .......... 13 2. iMeasuring Instruments ............. . ..... 14 3. Cluster 508: Mystery Frustrated .............. 20 4. Comparison of Problem Solving Questionaire A Priori Item Groupings and Item Groupings by Cluster Analysis . . . . 22 5. Correlations Between Problem Solving Questionaire Clusters in each of the A Priori Cognitive Style Groups . . . . . 27 6. Intercorrelations of Math and.Mystery Problem Solving Clusters ............ . . . . . . . . . . . . 29 7. Correlations of Problem Solving Questionaire Strategy and Anxiety Clusters with Measures of Problem Solving Efficiency ....... . . . . . . . . . . . . . . . . 31 8. Correlations of'Math Problem Strategy and Anxiety Clusters With Choice of Answer . . . . . . . . . . . . . . . . 33 9. Cognitive Ability Measures of Style, Problem Solving Efficiency, and State Anxiety: Reliabilities and Treatment Effects . . . . . . . ......... . . . . 43 10. Problem Solving Questionaire Clusters: Reliabilities and Treatment Effects ....... . . . . . . . . . . . . . 46 11. Mean Difference Scores and Z Scores for Cognitive Ability Measures of Style, Problem Solving Efficiency, and State Anxiety . . . . . . . . . . . . . . . . . . . . . . 51 12. Mean Difference Scores fer PrOblem Solving Questionaire Clusters . . . . . . . . . ...... . . . . ..... S3 13. Premeasures: Reliabilities and Group Differences . . . . . 60 14. Correlations Between Cognitive Ability Measures, Personality Scale Measures, and Problem Solving Questionaire Measures of Styles and Anxiety ...... . . . . . . . . . . . . 61 15. Cluster 501: Ability ............ . . . . . . . 63 vi Table 16. 17. 18. 19. 20. Page Cluster 502: Anxiety ................... 64 Cluster 503: Object USes ................. 65 Cluster 504: Time ..................... 66 Cluster 507: Grade-Age .................. 67 Correlations Among the Twenty Six Major Clusters ...... 74 Cluster 509: Mystery Concentrated . . . . ......... 111 Cluster 510: Mystery Enjoyed ............... 112 Cluster 511: hystery Strategy Change ........... 113 Cluster 512: Mystery Careful ............... 114 Cluster 513: Mystery Deliberate .............. 115 Cluster 514: Mystery Global ................ 116 Cluster 515: Mystery Notes ................ 11 7 Cluster 516: Bystery Systematic .............. 118 Cluster 517: Mystery Trial and Error ........... 119 Cluster 518: Mystery New Approach ............. 120 Cluster 519: Mystery Fixated ............... 121 hystery Questionaire Residual Items . ........... 122 Cluster 520: Math Frustrated ............... 123 Cluster 521: Math Concentrated ...... . . ...... 124 Cluster 522: Math Strategy Change ............. 125 Cluster 523: Math Deliberate . . . . . . . ........ 126 Cluster 524: Math Notes .................. 127 Cluster 525: Math Systematic . . . . . . . ........ 128 Cluster 526: Math New Approach .............. 129 Math Questionaire Residual Items .............. 130 vii Table H-21. H-ZZ. H-23. H-24. Cluster 527: Cluster 528: Cluster 529: Cluster 530: STAI-State Scale, Relaxed Items, Session 1 . STAI-State Scale, Tense Items, Session 1 . . STAI-State Scale, Relaxed Items, Session 2 . STAT-State Scale, Tense Items, Session 2 . . viii Page . 131 . 132 . 133 . 134 I. INTRODUCTION In a number of studies reviewed by Sarason (1960) and.more recently by Spielberger and Gaudry (1971) negative relationships have been reported between a variety of intellectual performance measures and scores on anxiety scales. Anxiety is a hypothetical construct that is used to refer to a complex set of human reactions, characterized by feelings of tension and apprehension and by increased autonomic nervous system activity (Levitt, 1967), which frequently occur when people are subjected to var- ious kinds of stress. Anxiety Scales typically consist of items which ask about the presence of certain physical (e.g., heart rate, perspiration level, stomach upset) and mental (e.g., worry, fear, self-doubt) behaviors which are used operationally as indicators of anxiety. When changes occur in these behaviors, inferences are made about changes in levels of anxiety. At the same time changes take place in the commonly used in- dicators of anxiety there may be other important changes taking place that are not.measured by most anxiety scales. For example, there may be cognitive process changes that occur when people are sUbjected to stress which could help to explain the negative relationships consistently found between self-reported anxiety and intellectual or academic performance. In the present study two aspects of mental functioning, cognitive styles and cognitive strategies, are investigated for the role they may play as mediators between psychological stress and overt perfbrmance on intellectual tasks. Four general questions are asked. The first two are concerned with the relationships between measures of anxiety, cognitive 2 styles and strategies, and intellectual performance. (1) Will the measured anxiety, cognitive styles and strategies, and problem solving efficiency of subjects vary systematically with changes in psychologically stressful conditions? (2) Will subjects who are measured as more anxiety prone tend to exhibit different patterns of cognitive styles and strategies than less anxiety prone subjects? The last two questions are concerned with whether the cognitive styles investigated in this study function as stable traits that can be used to differentiate one person from another. (3) Will subjects be measured as using the same cognitive styles consistently under different experimental conditions and across time? (4) Will the different measures of cognitive styles used here be shown to have construct validity, i.e. , will they be found to be measures of the dimensions they have been hypothesized to measure? Before going on to a detailed look at these questions, a ntmber of constructs will be clarified for use in this study and selected research from the literature pertinent to this investigation will be discussed. Anxiety Cattell and Scheier (1958) called attention to the lack of consenus amng behavioral scientists concerning the meaning of the tem "anxiety" when they reported locating more than 300 definitions. Sarbin (1968) has emphasized this lack of clarity and consensus by noting that "anxiety" has becane an Opaque metaphor. He cautioned against at least two mistakes that have comonly been made: (1) thinking of anxiety as a distinct entity that has an existence of its own independent of its behavioral referents, and (2) attributing a causal role to that entity. He has suggested "cognitive strain" as a new, less opaque metaphor to replace the term "anxiety" because it is more patently descriptive, it tends to 3 direct attention to behavioral referents, and it is less likely to be thought of as a cause in and of itself. Stress and strain are constructs which psychology has borrowed.from.the physical sciences. Stress refers to the forces applied to a structure or system.and strain refers to changes in the system that result from the applied ferce. To the extent this analogy is accurate, measures of anxiety may be thought of as indicators of the amount of psychological stress a person is experiencing. As was mentioned earlier, most anxiety scales contain behavioral referents for both the mental and physical aspects of anxiety. The scales may ask if these behaviors are present at the time the scale is admin- istered, if they are present in general, or if they are present in specific kinds of situations. Depending on the context in which these questions are asked, the anxiety measured may be called general trait anxiety or a specific kind of state anxiety whiCh is present only when certain environmental conditions are met. The concept of two quantita- tively different types of anxiety, trait anxiety and state anxiety, came out of the factor analytic studies of Cattell and.SCheier (1958, 1961). Spielberger has been the most active researcher with these two constructs and.has developed scales fer their measurement--the State-Trait Anxiety Inventory (STAI) (Spielberger, Gorsuch, and Lushene, 1970). Trait anxiety refers to relatively stable individual differences in general anxiety or in anxiety proneness while state anxiety is considered to be a temporary condition. Correlations between the STAI-Trait Scale and other widely used measures of general anxiety--the Manifest Anxiety Scale (MAS) (Taylor, 1953), the IPAT Anxiety Scale (Cattell and Scheier, 1963), and the Affect Adjective Checklist (AACL) (Zuckerman, l960)-- range from..52 to .80 fer college students (Spileberger, et. al., 1970). 4 State anxiety exhibited in academic testing situations has been called "test anxiety" (Mandler and Sarason, 1952; Sarason, 1958; Alpert and Haber, 1960). Test anxiety scales are better predictors of academic performance than general anxiety scales (Alpert and Haber, 1960). Test anxiety has consistently been found to be associated with lower levels of academic achievement in elementary school, high school, and at the university level. Moderate but consistent negative relationships have been reported between various anxiety scales and measures of intelligence (Spielberger and Gaudry, 1971). Anxiety and intelligence appear to have an interactive effect on performance. High anxiety facilitates performance for subjects of high intelligence on simple tasks and nest tasks of moderate difficulty. The combination of high anxiety and low intelligence leads to low performance except on very easy tasks (Spiel- berger and Gaudry, 1971) . The best evidence for the position that intelligence alone cannot account for low academic achievement in anxiety prone students comes from two studies which compare performance on learning tasks of subjects who have the same level of ability but who differ in test anxiety (Sarason, Davidson, Lighthall, Waite, and Ruebush, 1960; Wrightsman, 1962). These studies showed that under stressful conditions, in which the subjects were led to believe that their personal adequacy was being evaluated, the test anxious subjects performed more poorly, while under nonstressful conditions the test anxious and nontest anxious subjects did equally well. Individual differences in anxiety reactions to different kinds of stress are commnly found. After reviewing a nunber of experiments in 5 which the MAS was used, Saltz (1970) concluded that some subjects were more susceptible to pain- induced stress while others were affected more by failure-induced stress. In academic settings, students high in trait amxiety are more likely than those who score low in trait anxiety to react to negative evaluations of their performance and to failure experiences with increased state anxiety (Spielberger and Gaudry, 1971). The performance of test anxious persons is most seriously disturbed under stress conditions which lead them to believe that their personal intellectual adequacy is being evaluated on the basis of their task performance. The performance of people who have low test anxiety scores, on the other hand, is not inhibited in evaluative situations; they may even do better under these conditions than in nonevaluative situations (Wine, 1971). Inverse relationships have been found between measures of anxiety and self-esteem (Rosenberg, 1962; Suinn and Hill, 1964). High anxious subjects have been found to be more self-deprecatory, more self-preoc- cupied, and generally less content with themselves than less anxious subjects (Sarason, 1960). Thus, it is reasonable to assume that high anxious people are more easily threatened by situations in which their personal adequacy is evaluated. But what do high anxious persons do when they are threatened in this way that disrupts their intellectual performance? Liebert and Morris (1967) have suggested that test anxiety is made up of two major components: worry and emtionality. The worry component is considered to be concern over performance and the emotional aspect is thought to be autonomic arousal. Wine (1971) has hypothesized that it is the worry component of anxiety that adversely affects performance. She concludes that the self-focusing tendencies of high test anxious persons are activated by the pressures of testing situations: "the highly test anxious person responds to evaluative testing conditions with ruminative, self-evaluative worry and, thus, cannot direct adequate attention to task-relevant variables (Wine, 1971, p. 99)." This anxiety reaction which includes shifts in attention from the task to worry about self may be accompanied by changes in the cognitive styles and strategies 3 person is using. For example, on a timed test a person who is taking up part of the time with ruminative worry may respond to the task more impulsively, guessing, etc., in order to finish. Because he is hurried, he may also become less analytical in his approaCh. In a problem solv- ing situation he may be less able or feel less free to think of a variety of possible solutions. Thus, he may stick rigidly with the first hypoth- esis that comes to mind. For some kinds of intellectual tasks, these changes in strategies and style could be in a direction opposite that required for their efficient solution, thus, compounding the already deleterious effects of reduced time spent attending to the task because of worry. It was the consideration of possibilities such as these which led to the present investigation of relationships between anxiety, cognitive styles and strategies, and intellectual perfonmance. Qggnitive Styles and Strategies In recent years many investigators of thinking have fecused on the information processing mechanisms hypothesized to underlie intellectual performance. One important processing concept is that of cognitive styles whiCh are used as labels for individual differences in how people think, i.e. , the way in which they handle or process information; rather than what they think, i.e., the content of their thoughts; or the efficiency 7 or capacity of their thinking, i.e., abilities or skills. Cognitive styles are usually given a.meaning similar to personality traits. They are considered to be enduring or persisting characteristics of cognitive processing that can be used to differentiate one person from.another on a wide variety of intellectual tasks. Cognitive strategies, as they are defined.in the present study, differ from cognitive styles in two ways: (1) strategies are regarded as smaller units of mental behavior; one cognitive style may be manifest in several different specific strategies, and (2) strategies vary to meet the demands of each new intellectual task, while the same styles are thought to be used consistently by an individual regardless of situational changes. Cronbach (1970, p. 630) has taken issue with the notion that cog- nitive styles are stable personality traits. He believes that many tests which have been used as measures of cognitive styles are basically tests of mental ability. In order for the trait definition of cognitive styles to be supported it must be shown that they measure something different from that measured.by ability tests, and that the same styles are used consistently across time and on different types of tasks. Several different measures of cognitive styles are used in the present study to facilitate the examination of their validity as individual traits. Three cognitive style dimensions are investigated: flexible- rigid, reflective-impulsive, and analytical-global. The flexibility-rigidity style dimension as conceived in this study is made up of two components: (1) the breaking of sets or Einstellung, and (2) overcoming perseverative behavior, such as trying an unproductive hypothesis or strategy over and over again (Frederiksen, J., 1967). In both of these components flexible style requires the consideration of 8 several alternative ways of approaching the task at hand. Using a task involving the judging of the relative sizes of squares, Eriksen and wechsler (1955) gave one group of subjects random shocks while the stimuli were being presented. The anxiety (shocked) group used their two favorite numbers more times when identifying squares than the non- anxious group when making judgements. This suggests that under these stressful conditions subjects reduced.the variety of their responses, i.e., they responded.more rigidly. The impulsivity-reflectiveness style dimension refers to the tempo and carefulness with which people work on intellectual tasks (Kagan, 1965a). Persons who are classified as reflective tend to be more deliberate and cautious; they do not take many chances, and they spend more time on the task if allowed to do so. Impulsive style is charac- terized.by the opposite approach to intellectual tasks, i.e., lack of carefulness, guessing, and working so quickly that accuracy is affected. Impulsiveness has been measured with a personality scale developed by Barratt (1959), but the most commonly used instrunent is the Matching Familiar Figures Test(MFF) (Kagan, 1965b). On this test the amount of time taken on a series of visual discrimination tasks is used as an index of impulsivity-reflectiveness. conflicting results have been reported in the literature in regard to the direction of the relation- ship between anxiety and.impulsivity-reflectiveness. Barratt (1959), Mes- ser (1970), and Ruebush (1960) reported that anxiety was associated with reflective style, while Jackson (1967) and Shulman, Loupe, and Piper (1968) feund anxiety to be positively related to impulsivity. The analytical-global cognitive style is usually measured with tasks that involve visual perception, but Witkin (1964) believes that it 9 can be identified in other kinds of intellectual activity. A widely used instrument for measuring this dimension is the Embedded Figures Test (Witkin, 1950), which requires the subject to find a simple pattern embedded in a more complex field. People who do this well are classified field-independent. They are viewed as tending to analyze or categorize stimuli based on subelements of a field (analytical) rather than based on the field as a whole (global). Cbgnitive styles are assumed to play an integral role in many kinds of intellectual problem solving. Schroder and.Hunt (1957) found that subjects who avoided a.prob1em solving task on which they were failing 'were more anxious and rigid (not looking for alternate solutions) than were those who persisted in the face of failure stress. Kagan (1964c) has suggested some possible relationships between impulsive style, anxiety and problem.solving. He hypothesized that a child who responds impul- sively is more likely to start on the wrong path to finding a correct solution than is a child who reflects over the probable accuracy of many approaches before committing himself to one. Furthermore, Kagan (1964c) suggested that when the child realizes he is in error, he is likely to become more anxious and, consequently, his selection of a second.solution path is likely to be impaired. Thus, a vicious circle nay be set into action involving impulsiveness, anxiety, and defeat whidh could eventually result in the Child withdrawing from.problem.solving situations. These suggested relationships have not, as yet, been demon- strated. IMore research needs to be done with these variables so that the reasons fer inefficiency in problem.solving may be discovered. In the present study the relationships between measures of three different style dimensions, three types of anxiety, and deductive reasoning problem 10 solving efficiency are examined. In addition, the construct validity of the three cognitive styles is investigated. II . METHOD Subjects were 114 male volunteers who were enrolled in introductory psychology courses at Michigan State University during Fall term, 1971. Data was collected during two sessions spaced one day apart. The number of subjects attending each session varied from about 10 to 25. Sessions were held in a large classroom (200 seat capacity). Subjects sat several seats from each other in order to reduce interpersonal awareness and dis- traction. To facilitate an investigation of the effects of stress on the use of cognitive styles, strategies, and problem solving efficiency, either stress or nonstress conditions were instituted during each session. The details of the stress and nonstress conditions are discussed in Section IV. A completely counterbalanced 2 X 2 Latin Square design was used. One group of 57 subjects (Group I) experienced nonstress conditions in Session 1 and stress conditions in Session 2. The reverse order of treatment conditions was used with the other 57 subjects (Group II) which had the stress treatment in Session 1 and nonstress conditions in Session 2. Two forms of testing materials were prepared and both groups of subjects were given Form A in Session 1 and Form B in Session 2. This arrangement allows between group comparisons to be made within the same session on identical instruments under differing conditions. To the extent the two forms of each measure were parallel, Session 2 was a replication of Session 1 except for the stress manipulation. The design 11 12 is presented diagrammatically in Table 1. Also shown in Table 1 are the procedures used in each session. Session 1 lasted about an hour and a half, and Session 2 took about an hour and fifteen minutes. Ses- sion 1 took longer because two personality scale measures of style and two anxiety scales were administered in that session prior to the time experimental conditions were instituted. Table 2 presents a list of the measuring instruments used. Included are measures for three different methods of style assessment, measures of anxiety, measures of problem solving efficiency, and measures of academic ability. These instrmnents will be discussed in detail in later sections. Samples of each are given in Appendix A through G. In cases where data was not complete for a subject, his group mean for the measure in question was used as a substitute for his score. The analysis was carried out in several phases: (1) cluster analyses of the problem solving questionaires, (2) testing for significant differences in means between groups within sessions, (3) testing for significant differences in means between sessions within groups, (4) cluster analy- sis of the cognitive ability measures of style, and (5) generating cor- relations between the reduced set of measures produced by the cluster analyses. Each phase of the analysis is discussed in detail in subse- quent sections. In this section a brief description will be given of the analyses used to answer each of the questions mentioned in the intro- duction that are asked in the present study. The first question is concerned with whether measured anxiety, cog- nitive styles and strategies, and problem solving efficiency vary system- atically with changes in psychologically stressful conditions. Problem solving efficiency was measured with both time to solution and accuracy 13 Table 1 Procedures and Design Procedures Design Group I Group II 1. Session 1 1. Pre Experimental Condition Measures .A. Personality scale measures of cognitive style B. Achievement Anxiety Test C. STAI-Trait Scale 2. Experimental Conditions A. Instructions Nonstress Stress B. Practice Problems " " C. Cognitive ability measures of cognitive style Form A Form .A D. Problem Solving Tasks " " ” " E. Problem Solving Questionaires F. STAI-State Scale II. Session 2 1. Experimental Conditions .A. Instructions Stress Nonstress B. Practice Problems ” " C. Cognitive ability measures of cognitive style Form B Form B D. Problem Solving Tasks " " " " E. Problem Solving Questionaires F. STAI—State Scale G. Debriefing 14 Table 2 Measuring Instruments Instruments Hypothesized to be a.measure of: II. III. IV. cognitive Style Measures 1. Personality Scale Measures a. Flexibility-Rigidity Scale ...... . Flexible Style b. Reflectiveness-Impulsiveness Scale. . . Reflective Style 2. Cognitive Ability Measures a. Verbal Problems Test. . . . . . . . . . Flexible Style Sign Changes Test. . . . . ...... " ” Object USes: % Categories . . . . . . . " " Matching Familiar Figures Test. . . . . Reflective Style Object USes: Total (reverse scored) . . ” ” Hidden Figures Test . . . . . . . . . . Analytic Style Object Uses: % Part Responses . . . . . ” " roblem Solving Questionaire Clusters New Approach, Strategy Change, Trial and Error, Fixated (reverse scored) . . Flexible Style b. Notes, Systematic, Deliberate, Careful . . . . . . . . . . . . . . . . Reflective Style c. Global (reverse scored) . . . . . . . . Analytic Style PPWF‘PF‘P?‘ Anxiety Measures 1. STAI- Trait Scale. . . . . . ....... Trait Anxiety 2. Achievement Anxiety Test .......... Test Anxiety 3. STAI- State Scale. . . . . . . . . . . . . . State Anxiety 4. Problem Solving Questionaire Clusters: Frustrated, Concentrated (reverse scored), Enjoyed (reverse scored). . . . . ..... " " Problem Solving Efficiency Measures 1. Math Problem. Time ............. Prob. Solv. Efficiency 2. Math Problem: Right/Wrong . . . . . . . . . 3. 'Mystery Problem: Time . . . ,,,,,,,, 'v n n 4. Mystery Problem: Right/Wrong. . . ..... " " H Academic Ability Measures 1. High School Grade Point Average . . . . . . Academic Achievement 2. College Entrance Examination Percentile . . Academic Aptitude 15 of solution for a who-done- it type mystery deductive reasoning problem and a math riddle deductive reasoning problem. A self-report question- aire followed each problem which was designed to assess the use of cog- nitive styles or strategies and the level of state anxiety present while the subjects were working on those tasks. State anxiety present during the entire experimental session was measured with the STAI-State Scale (Spielberger, et. a1. , 1970). A number of cognitive ability type mea- sures of styles were also included to see if stress conditions had any effect on style usage as measured with this method. Treatment effects were investigated with phase 2 of the analysis--the comparison of mean differences between groups within sessions--and with phase 3--compari- sons of mean differences between sessions within groups. These analy- ses are discussed in Sections IV and V. To investigate the second question mentioned in the introduction, whether subjects who are more anxiety prone tend to exhibit different patterns of cognitive styles and strategies than subjects who are less anxiety prone, the data for all the measures listed in Table 2 was pooled across groups and averaged over sessions. A cluster analysis was performed on this data reducing it to a set of 26 variables. The intercorrelation matrix of these 26 clusters was then searched for unique patterns of relationships associated with high and low anxiety proneness. The results of this analysis are discussed in Section VII. The third general question asked in the present study inquires into the consistency with which subjects use the same cognitive styles. Test-retest reliabilities across sessions were computed for the cogni- tive ability measures of style and problem solving questionaire responses and were used as estimates of consistency in responding. This analysis 16 is covered in Sections III, IV and V. The last question mentioned in the introduction asks if the measures used in this study to determine cognitive styles have construct validity. An examination was made of the convergent and discriminant validity of these measures (Campbell and Fiske, 1959). Perfect discriminant validity would be shown if measures of different styles made with the same method of measurement were found to be uncorrelated. Absolute convergent valid- ity would be indicated if measures of the same style made with different methods were perfectly correlated. In testing for discriminant validity, the problem solving questionaire measures of styles and the cognitive ability measures of styles were each subjected to cluster analysis to see if different styles emerged as distinct factors within the same method of measurement. Evidence for convergent validity was sought in the cor- relations between measures of the same styles given by three different methods: the problem solving quest'onaires, the cognitive ability measures, and the personality scale measures. The details of these analyses are presented in Sections III and VI. The problem solving questionaire items were written to measure anxiety and the expression of three style dimensions: flexible-rigid, reflective-impulsive, and analytical-global. As part of the present study the questionaires were cluster analyzed to determine which dimen- sions were represented on them. Since the results of this analysis have a bearing on all the other analyses, it will be discussed first. III. (IDGNITIVE STRATEGIES USED ON THE PROBLEM SOLVING TASKS This section is mainly concerned with the identification of the dimensions measured by the problem solving questionaires. It was hoped that these dimensions included flexible, reflective, and global cognitive styles. If so, these dimensions could be correlated with the cognitive ability and personality scale measures of style to test for convergent validity (see Section VI). Since these dimensions provide measures of the styles or strategies used while working on the problems, they can be correlated with measures of problem solving time and accuracy to see if the styles or strategies used actually had any effect on problem solving efficiency. eggs. Two kinds of deductive reasoning tasks were included in the materials given to each subject, a who-done-it mystery problem and a math riddle word problem. One of each was included in each form of the materials. The mystery problems were adapted from a booklet of deductive reasoning problems by Sumners (1968). The problem on Form A was called "Malice and Alice," and that on Form B was entitled "Murder in the Family." A maxi- nnnn of eight minutes was allowed to reach a solution. The math problem on Form A was called "The Hotel Room Problem." Another version of it is conmonly found in many p0pular quiz and puzzle books. The Horse-Trading Problem (Maier and Burke, 1967) was used on Form B. Four minutes were allowed to complete the math problems. Time spent on each problem was 17 18 recorded by the subjects from a large clock in the front of the room that showed elapsed time. The mystery problems were written out in the materials given to each subject, but the math problems were not. The math problems were read by the experimenter twice to the entire group of subjects present in each session. Each subject had in his materials a list of five answers from which he was to select the correct one. Timing started when the experimenter began reading the problem. Following each problem in the materials was a true-false question- aire designed to measure subjects' reports of the cognitive styles they used asde as the anxiety they experienced while working on the preced- ing problem. The questionaire items were purposefu1ly written to be indicators of flexible-rigid, reflective-impulsive, and global-analytic style dimensions and state anxiety specific to the problem solving task. The items were grouped on the basis of their apparent content according to these a priori dimensions. The mystery questionaire contained 46 items, and the math questionaire 35. The math and.mystery questionaires had 27 items in conmon, i.e., items having the same or almost the same wording. The mystery questionaire had 18 items that were not on the math questionaire, and the math questionaire had seven items that were not represented on the mystery questionaire. The same a priori factors were represented.on both questionaires. The questionaire data was cluster analyzed by using the MGRP sub- routine of the PACKAGE system of computer routines (Hunter and Cohen, 1969). Cluster analyses were performed with the data from both groups of subjects combined on Session 1 (Form A) then replicated.on Session 2 (Form B). The fellowing criteria were used.in identifying unique homo- geneous factors: (1) each cluster must be internally consistent, i.e., 19 the items in a cluster must be relatively highly correlated with each other, (2) each cluster must show external parallelism, i.e., all items in a cluster must have approximately the same level of correlation with the other clusters, and (3) the items in eaCh cluster must be reasonably homogeneous in content. Once the clusters of items representing unique dimensions were identified, correlation coefficients were computed.be- tween the clusters on Session 1 and Session 2 which generated a matrix that included test-retest reliabilities (r Then the data from both 1,2)“ sessions was combined, and using the same cluster structure fbund pre- viously, intercorrelations were computed between all the obtained clus- ters. Correlations were also computed between the clusters and.problem solving time and solution. Results The first step in the analysis was to perform an oblique multiple groups factor analysis on the a priori item groupings. These a priori clusters failed to meet the criteria fer unique homogeneous factors. The item groupings were then revised and factor analyzed several times until homogeneous clusters appeared. ‘Mystery Frustrated (cluster 508) is presented in Table 3 as an example. .All the rest of the clusters (clusters $09-$19 fer the math questionaires and clusters 520-526 for the mystery questionaires) may be referred to in Appendix H. At the top of the Table 2 the items in cluster 508 are written out as they appeared on the questionaire. A.minus sign before an item number indicates that it was scored opposite to the way it reads. The results of the cluster analyses on Session 1 (Form A) and Session 2 (Form B) are shown side by side in the lower portion of the table. Across the top of that portion are the item numbers. Down the left hand side are the item numbers 20 TMfleS Cluster 508: Mystery Frustrated 34. .At times I felt discouraged. 44. After working on the problem fer a while I began to feel frustrated. ~45. This problem was easy for me to solve. 42. At times I worried that I might not be able to get the right answer. 43. Sometimes I wondered.how well the other students were doing on this problem. Form A Form B 34 44 45 42 43 34 44 45 42 43 34 52 51 37 37 24 41 54 29 32 04 44 51 46 42 36 14 54 55 31 32 15 45 37 42 29 22 20 29 31 16 18 05 42 37 36 22 27 20 32 32 18 28 23 43 24 14 20 20 10 04 15 05 23 04 508 73 68 S4 51 32 64 75 39 53 20 Frustrated 509 ~35 ~48 ~37 ~25 ~13 ~37 ~40 08 ~18 ~16 Concentrated 510 ~28 ~31 ~22 ~19 ~19 ~12 ~19 ~13 ~22 ~12 Enjoyed 511 44 32 36 25 26 31 24 ll 35 10 Strategy Change 512 ~32 ~40 ~37 ~16 ~16 ~61 ~55 ~24 ~31 ~13 Careful 513 07 ~04 ~05 19 01 ~06 ~11 ~09 15 12 Deliberate 514 26 10 15 25 17 01 ~14 07 15 06 Global 515 18 09 14 10 21 03 Ol 14 ~12 ~07 Notes 516 ~16 ~23 ~26 ~01 02 ~27 ~33 ~11 ~31 00 Systematic 517 16 26 04 07 03 05 ~03 ~07 ~13 ~13 Trial and Error 518 ~12 ~14' 09 ~26 ~26 ~22 ~18 ~07 ~13 ~13 New Approach 519 05 12 18 19 14 22 07 ~12 ~09 04 Fixated 599 05 09 ~05 ~11 ~11 12 04 ~06 ~05 ~14 r . pb1 21 Efiflb, the numbers of the mystery questionaire clusters (508-519) and point-biserial correlations (599) between Groups I and II, i.e., between treatment conditions. A.point~biserial correlation of .19 is required to meet the .05 level of significance fer treatment effects on any single item. None of the items showed significant treatment effects. Down the right hand side are the cluster names. Above the horizontal line are the item intercorrelation matrices. On the diagonal of these matrices are the communalities for each item. The communality is a measure of the percent of variance in each item which is accounted fer by the factor measured by the cluster to which it belongs. Below the horizontal line are the correlations of each item (corrected for attenuation) with its own cluster (part~whole correlation) and with all the other mystery questionaire clusters. ‘ The three criteria used to define unique homogeneous clusters may be checked as fellows: (1) Internal consistency is indicated in two ways: (a) each item should correlate more highly with its own cluster than with any other cluster, and (b) the coefficient alpha, which indicates homo~ geniety within the cluster (Cronbach, 1951) should.be reasonably large (coefficient alphas for all questionaire clusters are presented in Table 10 in Section IV.) (2) External parallelism may be checked by seaming down all the item: columns simultaneously to see if the direction and magnitude of correlations are approximately the same for all the items ‘with each cluster. (3) Hbmogeniety of content is checked by reading the items to see if they can all be logically connected in a reasonable way to the same dimension or factor. Table 4 presents the distribution of questionaire items in relation to their groupings by a priori style and anxiety dimensions (columns) 22 Table 4 Comparsion of Problem Solving Questionaire A Priori Item Groupings and Item Groupings by Cluster Analysis Item.groupings A priori itemggroupings by 31‘9“” Flexibility Reflectivenes Gfobalhess Anxiety an ys1s Math Wstery Mat}? Mystery Math hystery Math hystery New Approach 3,4 3,4 1,2, 5,6, Strategy Change 34 1,2 17 31 20 Trial and Error 5,6 Fixated 8,10 Notes 9,10 13,14 Systematic 12,13 15,16 Deliberate 14,15 20,21 18,19 25 18,19, 22,23, Carefu1 12 24 26,28, Global 29,30 36 21,22, 34,42, 24,26, 43,44, Frustrated 28,33 45 27,29, 30,31, 37,39, Concentrated 3;, 40,41 Enjoyed 33.35 Residual 8 7,9,11 11.17 25, ;7 23 32.38 Total no. items #10, ll 8 l3 2 6 14 14 23 and.the groupings which resulted from the cluster analyses (rows). Twelve different clusters were fermed from the four a priori item group- ings. Seven were represented on both the math and mystery questionaires and five were present only on the mystery questionaire. These five clus~ ters were, for the most part, made up of items appearing only on that questionaire. Taking the source of the majority of the items in a clus- ter as a criterion, the a priori flexibility items fermed feur clusters: New Approach, Strategy Change, Trial and Error, and Fixated. The reflec- tive items also formed four clusters: Notes, Systematic, Deliberate, and Careful. The global items fbrmed one cluster on the mystery questionaire-- Gldbal, and were an important part of the Deliberate cluster on the math questionaire. The items written to tap anxiety fermed three clusters: Frustrated, Concentrated, and Enjoyed. The names of the clusters were chosen to Smearize the dimension measured by each cluster. Clusters which contained.the same or very similar items on both the math and.mys- tery questionaires were given the same name. The content of the New ApproaCh cluster suggests that it is an indi- cator of the degree to which the first approach used on the problem sol- ving tasks was chosen to fit the particular demands of each problem, rather than using the strategy or approach habitually used on this kind of task. Thus, it may tap the set breaking aspect of flexibility. The Strategy Change cluster is a measure of whether the subjects tried more than one strategy or approach fer solving the problem. That does not necessarily mean that this cluster measures flexibility. A second stra- tegy may be tried.because the first one did not lead to the solution. The high positive correlations of the items in this cluster with the Frustrated cluster tends to support this latter interpretation. The 24 Trial and Error cluster measured the use of a nonlogical trial and error approach. Its items are negatively related to the Systematic cluster on the mystery questionaire which measures the use of a rational systematic plan on the problem solving tasks. The Fixated cluster is a two itemrmeasure that taps the perseverative aspect of rigidity in prob- lem solving, i.e., using the same hypothesis over and over again in attempt- ing to solve the problem instead of trying a new hypothesis. The meaning of the Notes cluster is clear~~making notes while working on the problems versus doing all figuringwithout the aid of paper and pen- cil. The Notes cluster may measure one manifestation of the carefulness aspect of reflectiveness. The Deliberate cluster is somewhat different on each questionaire. The Deliberate cluster on the math questionaire is a combination of items that clustered into both the Deliberate and Global clusters on the mystery questionaire. On the mystery questionaire, the Deliberate cluster indicates whether or not the problem was read.more than .once before the subject decided how to attack it. The mystery Global cluster represents the global-analytic dimension. It emphasizes visualizing the entire problem as a unit and getting an overall picture rather than concentrating on details. In the math Deliberate cluster both trying to get the overall picture of the problem and taking time before deciding how to try solving it were included. This indicates that deliberate and global approaChes are compatible. The different cluster- ing on the two questionaires is probably due to there being only two items on the math questionaire written to represent the global-analytical dimension while there were six on the mystery questionaire. But even on the mystery questionaire, Global was one of the weakest clusters. Since the strongest items in math Deliberate are more similar in con~ 25 tent to mystery Deliberate than they are to mystery Global, this math cluster was given the name "Deliberate." Hewever, the cross correla- tions show that math Deliberate is more closely related statistically to mystery Global (r = .18) than it is to mystery Deliberate (r . .07). The cluster that appears to most closely represent the carefulness aspect of reflectiveness is the Careful cluster. This cluster is made up of items that emphasize not guessing, not writing down an answer until the subject was sure it was correct, and being careful and cau- tious at all times. The Frustrated cluster is an indicator both of discouragement and worry about how well the subject was doing on the prOblem in comparison to others. The Concentrated cluster simply indi- cates whether or not the subject was having difficulty concentrating while working on the problems. The Enjoyed cluster measures whether or not the subject enjoyed working on the problems, taking them as an interesting challenge. These last three clusters-~Frustrated, Concen- trated, and Enjoyed were derived from a priori anxiety items. Their content suggests that they represent subjective states that may accom~ pany or contribute to the level of anxiety experienced rather than be- ing direct measures of the tension aspect of anxiety. The worry ele- ment in anxiety is apparent in the Frustrated cluster. As Will be shown later in Section VI, these three clusters are highly related to the other anxiety measures used in this study, hence, they will be referred to as indicators of state anxiety specific to the problem solving tasks. It is clear that the reduced set of variables produced by the cluster analyses is different from the a priori cognitive style group~ ings of items. Table 4 shows what occurred. The items written to 26 measure flexibility fermed feur clusters, the reflectiveness items formed four clusters, the global items remained one cluster on the mystery questionaire and were absorbed into another cluster on the math questionaire. The anxiety items formed three clusters. If the clusters that were constructed from items written to measure the same dimension are just measuring different aspects of that dimension, they should be correlationally related. If they are not related, then little confidence can be placed in the existence of the a priori dimensions as measured.by these questionaires. Correlations between clusters grouped by a priori style dimensions are presented in Table 5. The clusters derived from the flexibility items show almost no interrelationship. Although there are some signif- icant relationships between clusters in the reflectiveness group, those relationships were different on the math and mystery clusters. Among the mystery clusters, Systematic is positively related to Careful (r = .46) and negatively related to Deliberate (r = ~.22), but Deliberate and Careful are not related nor is Notes correlated with any of the other three. All three of the math clusters in this group, Notes, Sys- tematic, and Deliberate, are moderately positively relatedeut the cor~ relation between Notes and Systematic does not reach significance (r = .17). There were no significant correlations between the three clusters formed.from.the global-analytic items. The anxiety group, in contrast to the a priori style dimensions, shows strong correlations among all its clusters. The anxiety clusters are the only ones that show clear evidence that they are all measuring something in common. Thus, it must be con~ ~cluded that the prOblem solving questionaires did not measure the three 27 Table 5 Correlations Between Problem Solving Questionaire Clusters in each of the A Priori Cognitive Style Groups Flexibility .513: 5.11. 5.12 21.2. 22:) .52; Mystery New Approach 518 100 Mystery Strategy Change 511 O4 100 hystery Trial and Error 517 01 06 100 Bystery Fixated 519 ~06 12 ~14 100 Math New Approach 526 '2§** 02 08 ~08 100 Math Strategy Change 522 10 I21** ~16 ~10 09 100 Reflectiveness .515 .516 .513 .512 524. $231 523 Mystery Notes 515 100 ‘Mystery Systematic 516 11 100 Mystery Deliberate 513 ~09 ~22* 100 ANstery Careful 512 ~14 46*** ~09 100 IMath Notes 524 '§QF** 08 ~04 04 100 Math Systematic 525 ~08 .21* ~12 36*** 17 100 iMath Deliberate 523 ~07 OS .01 00 21* 19* 100 Globalness 514 .513 523 Mystery Global 514 100 Mystery Deliberate 513 13 100 Math Deliberate 523 18 91 100 Arm..__xiet as 30.2 _s_1_9 .529 :21 Nystery Frustrated 508 100 Mystery Concentrated 509 ~31*** 100 hystery Enjoyed 510 ~20* 27** 100 Math Frustrated 520 59*** ~22* ~25** 100 Math Concentrated 521 ~28** 31*** 23* ~57*** 100 Note.~~Cross correlations are underlined. *p<.05 **p < .01 ***‘p < .001 28 hypothesized style dimensions they were designed to measure. The clusters appear to represent smaller units of cognitive behavior than can appropriately be called styles, hence, these clusters will be con- sidered to represent cognitive strategies. The anxiety clusters will be referred to as measures of state anxiety present while subjects worked on the problem solving tasks. Table 6 presents the complete intercorrelation matrix of all the problem solving clusters. There are four that are quite consistently intercorrelated.within and.between both questionaires: Frustrated, Con- centrated, Strategy Change, and Careful. The only exceptions are be- tween math Strategy Change and mystery Concentrated, and.between math Strategy Change and.mystery Careful. Within the math problem, Strategy Change and Concentrated are significantly related. The direction of the correlations shows that frustration and changing strategies were associated with poor concentration and lack of carefudness. Coefficient alphas for eaCh cluster are presented in Table 10 in Section IV. The average size of the coefficient alphas fer the mystery questionaire clusters is .61 and for the math questionaire cluster, .70. These are high enough to indicate that.most of the clusters are quite homogeneous in content. The coefficient alphas for Sessions 1 and 2 are of comparable size for each cluster. This replication of cluster homogeniety lends confidence to the belief that the factors identified are not spureous. Since identical questionaires were used in both sessions, correla- tions between the same clusters on each session are estimates of coeffi~ cients of stability. Test-retest reliabilities (r1,2) are shown in Table 10, Section IV: They are low when compared with the coefficient 29 eee.ve n ew.~.e .8.venefinw gave-weep .UQEHHQUGD Ohm memo-muwHOHHOU mmOHU:.-.®HOZ eee ww- ee we we we- ee we- ww we .MH- we ee ee- ee- we we- ee- ee- eww eueoeee<.3oz ww- eee we we we ww we- ee- ee- we- ew e- ee we- ew ee- we ee we www oeweeeewww ee we eee ew we we. we we we- we- we ww we .we- we we we ee- we eww weeoz we we ew eee ew we- ew we- ww we- we we- we we ee .ww we we- we ”www eeeweeeewe we we we ew eee ww- ww ee- ee we. we- ee we- we- we- ew we-.wp- ww www eweeeu wweweeew we- ww ee- ee- ww- eee ww- ee we ee we ee- ww- we ew ww- ww ew mm» eww eeeeweeeueoe ee we- we ew ww ww- eee ee- we- we ww- we we we ew- ew ww- ww- ew eww eeeewewsew 3398 new: men ee- we we- ee- ee ee- eee ee- ee- we we we ee ee we- ee we- ee wew eeewxew ww ee- we- ww ee we we- we. eee ee we- ew- ee ew we we we- we- we. wew eueowee< 382 we .mw- we- we- we. ee we we. ee eee ew- ee- we- we ww- ee we- we- we. wew woewm eee eeeew we- ew .mp we we- we ww- we we- ew- eee ee we ww- we ew- ww ww ww- eew ueeeeeewww we we- ww we- ee ee- we ee ew- ee- ee eee we we- we- ee we ee we wew wewoz ee ee we .mw we- ww- we we ee we- we we eee we ee ew we we- ee eew eee0eo we- we- we. we we- we we we ew we ww- we- we eee we- ww we- we. we wew eweeeeeeee ee- ew ee ee .mp- ew ew- ee we ww- we we. ee we- eee ww- we ww ew- wew eswewwe we ee- we we ew ww- ew we- we we ew- ee ew ww ww- eee we- ww- ww eew eweeeu wweweeew we- we we we we-.w~ ww- ee we- we- ww we we we- we we- eee ww ew- eew eewewem ee- ee ee- we- ee- ew .mm- we- we- we- ww ee we- we. ww ww ww eee ew- eew eeweweeeueou ee- we we we ww ww- ew ee we- ee- ww- we ee we ew- ww ew- ew- eee wew eeeeeewsww meoumeeeu beam? eww www eww www www eww eww eew wew wew eew wew eew wew wew eew eew wew wew $335 mageom 2030.5 topaz one new: mo 803223335 e 033. 30 alphas. The average r1,2f6r both the mystery and math clusters is .23. This low average test-retest reliability indicates a lack of consistent responding among many subjects across sessions. This could be partially due to differences between the problems or inconsistencies in practice effects. The cross correlations between same name clusters from the mystery and.math questionaires are underlined in Table 6. The average cross correlation is .26. The ratio of test-retest correlation to coefficient alpha (average .35) or the ratio of cross correlation to coefficient alpha (average .40) may be used as an index of generality of the clusters across time or problem types. That this index is about the same across time or problem types indicates that the strategies and anxiety factors measured by the clusters are not specific to the type of problem. This is further evidence for the stable existence of these factors. Although the strategies show some generality across time and.problem.type, the low test-retest reliabilities indicate that they were probably not used consistently by all subjects. The last set of results in this section are concerned with the relationships between the cognitive strategies used and state anxiety experienced while working on the problems and problem solving efficiency. These correlations are shown in Table 7. Werking carefully and maintain~ ing good.concentration were helpful in solving the mystery problems, ‘while trying the same hypothesis over and over hindered correct solution. Taking notes and getting frustrated took time but did not necessarily affect finding correct solution. On the math problems, a logical sys- tematic approach was helpful in solving them. WOrking the problems systematically took more time, but it paid off in leading to right answers. As with the mystery problems, getting frustrated and taking 31 Table 7 Correlations of Problem Solving Questionaire Strategy and Anxiety Clusters with Measures of Problem Solving Efficiency Mieaam hystery Problem: correct 417 100 Mystery Problem: time 418 00 100 iMath Problem: correct 419 03 ~02 100 Math Problem: time 420 16 34*** 01 100 ‘Mystery Clusters Strategy Change 02 04 ~13 03 Fixated ~19* 11 ~02 ~16 Trial and Error ~12 ~07 ~08 ~14 Notes ~06 34*** 03 01 Systematic 07 05 14 ~04 Careful 27** ~06 25** 09 Deliberate 05 ~09 ~22* ~10 New Approach 11 ~09 ~19* 21* Enjoyed 06 06 17 ~09 Frustrated ~06 29** ~06 08 Concentrated 23* 07 02 06 Global 03 ~07 ~06 ~09 Math Clusters Frustrated ~07 ~04 ~02 20* Concentrated 08 04 12 ~02 Strategy Change 02 12 ~03 42*** Deliberate 03 ~09 05 23* Notes 12 10 13 19* Systematic 06 09 24** 23* New Approach 12 ~13 ~05 15 *p < .05 **p <.01 ***p < .001 32 notes added time but did not affect the solution. This was also the case for changing strategies and fer deliberating before beginning on the math problems. There were five answers to each math problem, only one of which was correct. Correlations between the math strategies and state anxiety and answer choice are presented in Table 8. Only six out of 70 correla- tions are significant. It was thought that some strategies might consis~ tently lead to certain answers, but no meaningful pattern of relation- ships isevident between strategy and choice of answer. Discussion The use of questionaires that are factor analyzed then correlated with measures of problem solving efficiency was found to be an effective method for identifying differences between problems in terms of the cog- nitive processes used in accurately and efficiently solving them. Simi- lar use of the questionaire technique has been made by Marshall (1972) investigating problem solving strategies and.by C. Frederiksen (1969) who used them to help identify memory retrieval strategies for verbal learning tasks. The results of the cluster analyses of the question- aires showed that the hypothesized cognitive style dimensions were not identified on the prdblem solving tasks. Instead, smaller, more specific units of reported.problem solving behavior were isolated. These were labeled cognitive strategies rather than styles because they represent very specific subject characteristics that do not fit the content defin~ itions of the styles under study. These strategies exhibited some gen- erality across time and problem type. However, the low test-retest reliabilities obtained suggest that many subjects did not use them con~ sistently. 33 Table 8 Correlations of'Math Problem Strategy and Anxiety Clusters With Choice of Answer Frustrated Concentrated Strategy Change Deliberate New Approach Hetel Room Problem |._1 U1 O (N 0 ON I I l . HOOOHSmwmuc NVQJVU‘I O O l. 30 ~‘l = 27’+ 2 03 ~06 2. 30 ~ 2 = 27 + 1 ~10 06 ~09 00 ~11 ~11 3. l = 3 ~ 2 a 01 ~04 07 ~05 ~14 10 4. 25 + 2 = 27 03 06 ~05 02 07 ~07 5. 5 ~ 3 = 2 05 ~12 05 02 ~05 19* HOrse Trading Problem 1. Lost $10 ~07 07 ~14 ~05 00 00 ~10 2. Broke even ~14 07 ~15 02 ~23* ~01 ~20* 3. Made $10b 13 ~11 10 01 00 ~17 22* 4. .Made $20 ~07 ll 01 O6 11 28** 00 5. Made $30 07 ~10 02 ~11 06 ~19* ~11 a Correct answer to the Hotel Room Problem. b Correct answer to the Horse Trading Problem. *p <.05 *‘p <.01 34 The strategies were not related to each other in the way that would be expected if groups of them were measuring different aspects of the same cognitive styles. Thus, it is concluded that the questionaires do not have discriminant validity as measuring instruments for the styles in question. Of course, this may be due, in part, to inadequacies in the questionaires themselves. The highest cross correlation feund was .50 for the Frustrated cluster. If this value was corrected for atten- uation with the Spearman-Brown fbrmula, it would be raised to .67 which demonstrates that reasonably good reliabilities can be obtained with this questionaire method. Reliabilities could be improved if identical questionaires were used on the two types of problems. Better items could also be writtene Nbre items could be constructed for some of the strategies; the global-analytic dimension in particular was inadequately represented. The results also showed that the strategies which were associated ‘with time taken to solve one type of problem and correctly solving it were not always associated in the same way with another type of problem. Since this was the case, it may be inferred that the ability to adjust strategies to match the requirements of different types of problems could.be an aid to general problem solving efficiency. Finally, positive interrelationships were feund between changing strategies, poor concentration, lack of carefulness, and the Frustrated cluster which tapped discouragement and the worry component of anxiety. These relationships are consistent with the thesis proposed by Wine (1971) that worry detracts attention away from the task (poor concentra~ tion). In addition, these findings suggest that when worry is present, subjects are less careful and change strategies more. The Strategy 35 Change cluster does not appear to be a measure of flexibility, and it is possible that subjects who changed strategies more may have done so because they chose ineffectual strategies more often. Further investigations of the construct validity of cognitive styles will be examined in later sections. Discriminant validity of the cogni- tive ability measures and the convergent validity shown by the personal- ity scale measures, problem solving measures, and cognitive ability measures will be covered in Section VI. IV. TREATMENT'EFFECTS ON STATE ANXIETY, ABILITY MEASURES OF COGNITIVE STYLE, PROBLEM SOLVING STRATEGIES, AND PROBLEM SOLVING EFFICIENCY The main question asked in this section is whether cognitive styles, cognitive strategies, and problem solving efficiency vary systematically with changes in psychologically stressful conditions. This involves four specific experimental questions: Are group means under stress ver- sus nonstress conditions significantly different for: (1) measures of state anxiety, (2) ability measures of cognitive style, (3) reported problem solving strategies, and (4) time used and correct solution of the prOblem.solving tasks? In this section comparisons are made between groups within the same session to see if stress or nonstress conditions produced differ- ences in strategy Choice and style usage. In Section V strategy choice and.style usage changes will be examined within groups across sessions. For the time being it will be assumed that the cognitive ability measures of style have construct validity. Their validity will be examined in Section VI. If cognitive styles are stable personality traits, they should not be seriously affected.by changes in external conditions. So one reason for trying to find out whether or not style usage differences are a550ci~ ated with differences in stress ' conditions is because of the bearing this has on the definition of cognitive styles. The other reason is to see if stress induced changes in styles (and in strategies) can help ac- 36 37 count for stress induced changes in problem solving efficiency. m As was mentioned in Section II, a completely counterbalanced Latin Squares design was used fer this phase of the study. Table 1 outlines the design and procedures used in eaCh experimental session. One group of subjects (Group I) experienced nonstress conditions in experimental Session 1 and stress conditions in Session 2. The reverse order of treatment conditions was used with Group II which had the stress treat- nent in Session 1 and nonstress conditions in Session 2. Two forms of testing materials were prepared and both groups of subjects were given Form A in Session land Form B in Session 2. This arrangement allowed between group comparisons to be made within the same session on identi~ cal measures under differing conditions. Each subject was given a packet of test materials which had in- structions on the first page. This page was read aloud by the experimen- ter, while the subjects silently read their own copy. For nonstress conditions the instructions were titled "Problems" and for stress con- ditions the heading was "Measuring Intellectual Abilities." Under stress conditions the instructions led the subjects to believe that they were taking an IQ test and encouraged them to do their very best. Then they were subjected to a failure experience by being given five minutes to "warm up" on some unsolvable practice problems. This was intended to be a combination of ego threat and failure stress. Nonstress conditions consisted of instructing the sUbjects that they would.be 'working on a variety of problems so that the experimenter could learn more about differences in the way people like to work on these kind of tasks. They were told that achievement was not important and were en- 38 couraged to relax. These instructions were followed by five minutes to work on some easily solvable practice problems which was intended to be a success experience for the subjects. The practice problems were adapted from problems used on a collection of short tests of IQ (Eysenck, 1966). The actual instructions and practice problems as well as samples of all the other instruments used may be referred to in the Appendices. At the end of Session 2 all subjects were given a debriefing which was intended to alleviate possible detrimental effects that may have occurred if any of them mistakingly assumed that they had done poorly on an IQ test. Questions were invited from the group, the reasons for the use of deception were made clear, and the true nature of the study was explained. Two forms of cognitive ability and problem solving measures were prepared. The problem solving tasks were described in Section III. The cognitive ability measures used in this study were designed in the fermat usually feund on tests of intellectual abilities, but they have been used extensively as indicators of cognitive styles. The tasks were pre- sented in the sane order on both forms and will be discussed in that order below. The first task, the verbal Problems Test (wand, 1958) has been used as a measure of cognitive flexibility. Here it is used as an indicator of flexible style. On this task subjects must be able to ignore the con~ textual meanings of words to get a high score. Ten minutes were allowed to work on the verbal Problems Test. Next came the Hidden Figures Test (HFT) (French, Ekstrom, and Price, 1963). This test measures the cognitive factor known as Flexibility of Closure. The HFT is an adaptation, as is Witkin's Embedded Figures Test, 39 of the Gottschaldt Figures Test popularized by Thurstone. Subjects are required to find which one of five sample figures are embedded in each of 16 different complex patterns. In this study the HFI‘ is used as a measure of field-independence which, according to Witkin (1964), is the perceptual task manifestation of analytical style. The HFI‘ has correlated .62 with the individual form of the Embedded Figures Test and .44 with a group administered form (Jackson, Messick, and Myers, 1964). The time limit on the HFI‘ was 10 minutes. The Sign Changes Test (French, et. a1., 1963) was next. It is another measure of cognitive flexibility that is less dependent on ver- balability than the Verbal Problems Test- It was used in this study to indicate flexible style. On this task, subjects are required to do simple mathematical calculations that are opposite those indicated by the symbols shown. For example 8 % 4 is read 8 multiplied by 4, and the correct answer is 32. This task had a two minute time limit. The Sign Changes Test was followed by the Object Uses task which requires productive thinking. Goldner (1957) used this task to study individual differences in whole-part approach and flexibility-rigidity in problem solving. Subjects were given five minutes to name all the ways they could think of to use a common object. On Form A the object was "a cardboard box," and on Form B, it was "a sheet of newspaper." Responses were classified in three ways. Total number of uses listed in the five minute period allowed was used as an indicator of the impulsive-reflective style dinension. The percentage of total uses given that were in differ- ent categories was used as a measure of flexibility. The percentage of total uses given that required the object to be broken up into parts or pieces was used as a measure of analytical versus global style. 40 Four items from Shulman's adult form of the Matching Familiar Figures Test (Shulman, Loupe, and Piper, 1968) made up the next task. It was used as a measure of impulsivity. On each page there are seven drawings of a familiar object. One figure is the correct sample. Some- thing is changed in five of the other six drawings. The subjects' task is to find.the drawing that is exactly like the sample. Short solution times and high errors are the characteristics of impulsive responding. A maximun of three minutes was allowed on each item. They recorded their own elapsed time to solution for each item from a large clock placed in the front of the room. Next followed the problem solving tasks and questionaires. The last of the materials in each experimental session was a slightly mod- ified version of the state anxiety form of the STAI (Spielberger, et. al., 1970). The modification was to put the items in the past tense-~"I felt calm," instead of "I feel calm." Subjects were instructed to fill out the 20 item, four point scale according to how they felt during the entire experimental session. Four sets of means and standard deviations were computed on all of these measures, one set fer each group under each experimental condition. Point-biserial correlations were computed on Session 1 between Group I (nonstress) and Group II (stress), and on Session 2 between Group I (stress) and Group II (nonstress). When one variable is continuous, as are the measures described above, and the other variable dichotomous, as are the treatment conditions, computer programs for giving Pearson r's automatically yield point-biserial r's. Another fermula fer the point-biserial r (Guilfbrd, 1965, p. 322) is: 41 (M 'M) = p g M— rpbi at W mean of one dichotomous group on the continuous variable M P Md mean of the other dichotomous group on the continuous variable P proportion of cases in one group q = proportion of cases in the other group at = standard deviation of the total sample on the continuous variable If Mb is larger than MH’ rpbi will be positive, otherwise it will be neg- ative. Since rpbi depends directly on the difference between the means of the two groups, a significant correlation indicates a significant dif- ference between the means. Therefore, the usual t_test of difference between the means can be used to test the significance of the departure of the correlation from zero and vice versa. The formula given by GuilfOrd (1965, p. 163) was used in the present study: T‘ Vii " 2 ’VI ~ r5 Split-half reliabilities (rxx) and coefficients of similarity,(0) 3: were computed fOr all the measures that had more than one item per form. The similarity coefficient is an index of the degree to which the two split halves correlate in the same way with all the other variables in an intercorrelation matrix. A coefficient of stability and equivalence (r1,2) was computed for all the measures by combining groups within sessions and correlating their scores on Session 1 (Form A) with their scores on Session 2 (Form B). .A cluster analysis was performed on the STAT-State Scale. The same procedures and criteria were used as in the cluster analyses dis- cussed previously in Section III. The analysis was performed first on 42 the data from both groups combined on Session 1, then replicated on Ses- sion 2 data. Results The reliabilities, means, and point-biserial correlations for the cognitive ability style measures, problem solving efficiency measures, and STAI neasures of state anxiety are presented in Table 9. Measures, for which rxx and ID reliabilities were not calculated had only one item on each form. The Hidden Figures and Sign Changes Tests were shown to be very homogeneous, the Verbal Problems Test slightly less so, and the Matching Familiar Figures Test considerably less homogeneous. Test- retest reliability is indicated by r1,2 to the extent that Form A and B measures are parallel. A low r1,2 indicates either nonparallelism be- tween forms or instability in subject performance over time. Table 9 shows that most of the test-retest reliabilites are high enough, rela~ tive to the size of r xx and D, to indicate fairly consistent subject performance on the two forms. If the neasures are rank ordered by the size of their internal consistency reliabilities and test-retest reli- abilities, the sane order is obtained. Thus, assuming that subject per~ formance was stable, the single item measures with low test-retest reli- abilities (Object Uses: 3 Categories, Math Problem, and Mystery Problem) probably would have had low parallel form reliabilities had it been pos~ sible to measure them. Reliabilities for the STAI-State Scale are also shown in Table 9. Since identical scales were used on Forms A and B, r1,2‘1 is an estimate of a coefficient of stability. Relative to the high internal consistency reliabilites obtained, the test-retest reliabilities are quite low. Cal- culated separately, the test-retest coefficient for Group I was .54 while 43 Table 9 Cognitive Ability Measures of Style, Problem Solving Efficiency, and State Anxiety: Reliabilities and Treatnent Effects . ,Reli'abilitya Measunng p Tn r1 , 2 Group I Group I I rpbi Instmnt sass. 1 2 1 2 1 2 1 2 1 2 Cond. NS 8 8 NS.- Verbal Problems .65 .79 .58 .69 .52 8.68 8.86 8.28 7.541 .07 .20* Hidden Figures .94 .94 .90 .89 .64 4.60 7.56 5.11 7.32 ~.08 .03 Sign Changes .80 .83 .70 .76 .55 20.40 22.30 20.00 22.26 .05 .00 O.U. total .65 15.04 17.09 13.74 15.93 .10 .10 O.U. % categ. .13 55.02 68.53 55.56 75.75 ~.01 ~.Z7** O.U. % parts .28 13.72 17.16 14.91 15.23 ~.05 .1l PETE correct .46 .43 .39 .35 .20 2.60 2.89 2.67 2.47 ~.03 .19* NET: time .76 .69 .67 .68 .48 4.32 5.78 3.93 5.59 .14 .05 Myst. P. correct ~.02 .49 .35 .51 .44 ~.02 ~.09 Math P. correct .02 .46 .46 .44 .39 .02 .07 Myst. P. time .29 5.66 5.83 5.54 6.34 .03 ~.l3 Math P. time .13 2.37 1.36 2.67 1.38 ~.15 ~.01 STAI-State Scale .97 .97 .94 .95 .35 2.21 2.18 2.39 1.89H~.10 .29** STAI-State: .89 .94 .40 2.36 2.32 2.08 2.58 .297 ~.16 relaxed cluster STAI-State: .89 .87 .34 1.78 1.69H 1.85 1.35 ~.06 .37*** tense cluster a 1' ficient alfi‘fias. All other r relations corrected for attehfia *p <.05 **p <.01 ***p <.001 tion with the Spearman-Brown formula. coefficients for the STAI relaxed and tense clusters are coef- reliabilities are odd-even split half cor- 44 that for Group II was .31. The lower value for Group II may reflect a greater variety of reactions to the treatnent conditions among subjects in that group. As will be discussed shortly, Group 11 showed greater changes in state anxiety across sessions than did Group I. The full scale values for the STAI-State scale presented in Table 9 show that Group I (stress) was significantly more anxious than Group II (nonstress) during Session 2. During Session 1 the mean for Group II (stress) was higher than that for Group I (nonstress) but the difference between means did not reach significance (p < .10) . The cluster analysis of the STAI-State Scale showed that it contains two distinct factors, one for relaxed items (e.g., "I felt content," "I was relaxed") and one for the tense items (e.g., "I was tense," "I felt nervous"). The clusters may be referred to in Appendix H. Coefficient alphas, test-retest reliabilities, means and point-biserial correlations are presented for these clusters in Table 9. The tense and relaxed item clusters behaved sonewhat differently for each group under the treatment conditions. In Session 1 the stress group was significantly less relaxed than the nonstress group, but there was no difference between them in tension. In Session 2 the stress group was significantly more tense than the nonstress group and was also less relaxed but not significantly so (p < .10) . Looking again at the full scale values for state anxiety, means are expressed in terms of single item values where l = not at all, 2 = sone- what, 3 = moderately so, and 4 = very much so. There were 20 items on the scale so means expressed in total score values are as follows: Group I nonstress = 44.20, Group I stress = 43.60, Group II stress = 47.80, and Group II nonstress = 37.80. These measures may be compared with those 45 found by other researchers. USing college males, Spielberger, et. al., (1970) report means of 32.70 under relaxed condifions, 36.99 under normal conditions, and 43.01 for exam conditions in which students were led to believe that they were taking an IQ test. Thus, it may be concluded that subjects in Group I were about as anxious in both sessions as stud- ents usually are when taking IQ exams, and that subjects in Group II were even more anxious than that under stress conditions and returned to nor- mal levels of state anxiety under nonstress conditions. From the point-biserial correlations presented in Table 9 it can be seen that only three of 24 between groups comparisons of means on the cognitive ability style measures and.problem solving tasks showed a sig- nificant treatment effect. Under stress conditions Group I did better on the verbal Problems Test and.on the Matching Familiar Figures Test: Cor~ rect. Under nonstress conditions, Group II produced a higher percentage of object uses of different categories. Pure treatment effects would have given point-biserial correlations of equal magnitude and opposite sign for each session. The three effects obtained were all in Session 2, so it is likely that they are the product of interactions between treat- ‘ments and.practice effects. Table 10 presents the reliabilities, means and point-biserial cor- relations for the prOblem solving strategies and anxiety clusters. The reliability of these measures was discussed in Section III. Group ‘means were not significantly different for any of the problem solving strategies or anxiety clusters on either Session 1 or Session 2, i.e., there were no treatment effects on these measures. Thus, stress con- ditions did not cause Changes in strategy or the approach used on either prOblem solving task nor did stress conditions significantly affect the 46 Table 10 Reliabilities and Treatment Effects Problem.Solving Questionaire Clusters: Reliabili I‘ Cluster 1 2 Mystery Questionaire Frustrated Concentrated Enjoyed Strategy Change Careful Deliberate Global Notes Systematic Trial and Error New'Approach Fixated Math Questionaire Frustrated Concentrated Strategy Change Deliberate Notes Systematic New'Approach .69 .55 .66 .54 .59 .66 .42 .62 .85 .43 .50 .48 .74 .71 .79 .36 .83 .73 .70 .62 .67 .79 .58 .61 .73 .54 .68 .84 .53 .61 .48 .69 .64 .84 .54 .85 .61 .72 .22 .27 .27 .31 .04 .31 .41 .22 .15 .13 .16 .31 .19 .31 .03 .22 .19 .31 .36 4.02 6.77 3.63 6.32 9.30 3.16 7.81 2.81 3.02 3.09 3.42 3.16 7.96 10.95 10.19 8.26 2.65 3.30 3.16 3.17 7.33 3.46 5.75 9.68 2.86 7.18 3.33 3.14 2.98 2.81 2.81 6.58 11.63 8.58 7.02 3.32 3.65 3.16 4.12 6.90 3.39 6.18 9.51 3.10 7.77 2.80 3.11 2.91 3.30 2.98 8.19 11.16 9.90 8.16 2.58 3.28 3.39 3.25 7.21 3.46 5.80 9.86 2.90 7.09 3.37 3.07 2.80 2.93 2.70 6.77 11.70 8.47 6.63 3.32 3.67 3.23 .04 .17 .05 .06 .03 .01 .01 .05 .11 .08 .11 .07 .04 .04 ~.13 .03 .00 .02 .05 .02 .03 .02 .04 .12 .07 .06 .06 .04 .02 .14 .00 .00 .04 47 degree to which subjects reported enjoying working on the problems, the amount of frustration they experienced, or the degree to which they were able to concentrate. Discussion The stress conditions were effective in elevating state anxiety to levels as high or higher than those reported by at least one group of researchers using similar treatment conditions. Although treatment con~. ditions were psychologically stressful enough to produce differences be- tween groups in state anxiety, they did not affect most cognitive ability measures of styles, use of problem solving strategies, or problem solving efficiency to a significant degree. It would appear that the intellectual performance of most college students is relatively unaffected by moderate increases in level of anxiety. These conclusions are based on comparisons between group means, and it is possible that some individuals did change strategies and styles as a result of the treatment conditions, while group means remained unaffected. Some styles and strategies are related to problem solving efficiency (as shown in Section 111), but the lack of treatnent effects in this phase of the study makes it difficult to con'~ clude anything at this point about how systematic changes in these vari~ ables may cause changes in problem solving efficiency as conditions vary between stress and nonstress. Correlational data bearing on the relation- ships between anxiety proneness and use of problem solving strategies will be examined in Section VII. The cluster analyses of the STAI-State Scale and the differential effects of the items that measured tension versus those that measured relaxation suggest that these may not be opposite ends of the same continuum, but separate aspects of the anxiety response. The relaxed items emphasize state of mind (e. g., calm, secure, 48 self-confident) and most of the tense items refer to body states (e.g. , tense, jittery, nervous). Hence, the results of the cluster analyses support a two component model of anxiety reactions. Style usage as measured by the cognitive ability tests appears to be unaffected by the kinds of stress conditions used in this study and the test-retest reliabilities show that performance on them was fairly consistent. This may be taken as evidence for the stable usage of styles so long as these measures are considered to be valid measures of cognitive styles. The use of styles and strategies within groups across time and their relation to changes in problem solving efficiency across time will be taken up in Section V. ‘V. PRACTICE EFFECTS ON STATE ANXIETY, ABILITY MEASURES OF COGNITIVE STYLE, PROBLEM SOLVING STRATEGIES, AND PROBLEM SOLVING EFFICIENCY In this section practice effects will be considered using the same data and design discussed in Section Iv in regard to treatment effects. Changes in state anxiety, style scores on the cognitive ability tests, changes in problem solving strategies, and changes in problem solving efficiency will be examined within each group of subjects between Ses~ sion 1 and Session 2. Although the cognitive ability measures of style were mostly Lmaffected by differences in treatment conditions, there is still the possibility that they may change as a result of practice. If they do, this will have a bearing on the definition of cognitive styles as stable individual traits. Data from this phase of the analysis will also be examined to see if changes in strategies over time occur which can be tied to changes in problem solving efficiency. e229. This analysis is much more straight forward than it would have been if there were generalized treatment effects. Separating treatment from practice effects would have been a complicated matter. As it is, except for the few measures for which treatment effects were indicated in Section Iv, changes in means will be considered to be pure practice effects. Group means for all the measures were computed and the Session 1 means were subtracted from the Session 2 means, giving mean difference 49 50 scores. Standard deviations of the difference scores were computed, and the mean difference scores were converted to Z scores for testing the sig- nificance of changes between sessions. Results The mean difference scores and 2 scores for the cognitive ability measures of style, measures of problem solving efficiency, state anxiety measured by the STAI—State Scale are presented in Table 11. Since Ses- sion 1 means were subtracted from Session 2 means, a negative difference score indicates that scores went down-on that measure across sessions and a positive difference score indicates increasing of scores across sessions. Group I went from nonstress to stress conditions across ses- sions, and Group II went from stress to nonstress so pure treatment effects would be indicated.by mean difference scores for each group that are equal in size but opposite in sign. Pure practice effects ‘would.be indicated by mean difference scores of the same size and in the same direction for both groups. Group I did not change significantly in state anxiety going from non~ stress to stress conditions. Group II, however, shows a large reduction in anxiety as they went from stress to nonstress conditions. A possible explanation for the difference between the anxiety reactions between the two groups will be suggested in the discussion. Two (verbal Problems Test, Matching Familiar Figures: Correct) of the other three measures for which treatment effects were found in between groups comparisons (discussed.in Section IV) also have mean difference scores of opposite sign. The difference scores for the third measure, Object USes: % part responses are in the same direction, but one is significantly positive 'while that for the other group is barely above zero change. For all 51 Table 11 Mean Difference Scores and 2 Scores for Cognitive Ability Measures of Style, Problem Solving Efficiency, and State Anxiety Measuring Mean Difference Z Score 13‘1““th Group I [Group II Group I LGroup II verbal Problems 0.18 ~0.74 0.45 ~l.96* Hidden Figures 2.97 2.21 7.85*** 5.29*** Sign Changes 1.90 2.26 3.85*** 4.53*** O.U. total 2.05 2.18 2.72** 3.78*** O.U. % categ. 13.51 20.19 4.98*** 6.42*** O.U. % parts 3.46 0.32 2.04* 0.23 IWFF: correct 0.28 ~0.19 1.74 ~1.06 MFF: time 1.45 1.66 6.34*** 7.32*** Myst. P. correct ~0.14 ~0.07 ~l.51 ~0.76 Nhth P. correct 0.00 ~0.05 0.00 ~0.53 Myst. P. time 0.17 0.81 0.53 2.94 ‘Math P. time ~l.02 ~l.28 ~6.72*** ~8.23*** STAI-State Scale ~0.02 0.50 ~0.38 6.72*** * p <.05 at p <.01 «as p <.001 52 three of these measures the mean difference scores for each group are of unequal size indicating an interaction between treatment and practice effects. Treatment effects probably played a minor role in the changes that occurred across time on the other measures shown in Table 11. Subjects in both groups improved significantly with practice on the Hidden Fig- ures Test, Sign Changes Test, Object USes: total and % categories. Both groups took longer in Session 2 on the Matching Familiar Figures Test and less time on the math problem. Group II took more time on the mystery problem.in Session 2. Table 12 presents the group mean difference scores for the problem solving strategies and.problem solving anxiety clusters. 'Virtually every significant change in one group is accompanied by a change in the same direction for the other group. Hence, treatment conditions do not appear to have played a significant part in the changes that took place across sessions on these measures. Going from Session 1 to Session 2, both groups of subjects became less frustrated (less discouraged and worried), and were better able to concentrate. On Session 2 both groups used fewer strategies, but were less likely to use the same hypothesis over and over. On Session 2 both groups were also less deliberate (took less time before beginning), were less global in their approach (more analytic), and more members of each group took notes rather than doing the problems in their heads. ‘More subjects were sys- tematic on the math problems in Session 2 but did not change signifi- cantly on this variable for the mystery problem. Fewer used.a new approach on the mystery problem in the second Session, but this measure did.not change significantly on the math problem. Nor were there signifi- 53 Table 12 iMean Difference Scores for Problem.Solving Questionaire Clusters Cluster Name ,Mystery Problems Math Problems Group I [GroupII Group I (LGroup II Frustrated ~0.85*** ~0.87*** ~l.38*** ~1.42*** Concentrated 0.56** 0.31 0.68*** 0.54** Enjoyed ~0.17 0.07 Strategy Change ~0.57** -0.38* ~l.6l*** -1.43*** Careful 0.38 0.35 Deliberate ~0.30* ~0.20 ~l.24*** ~1.53*** Global ~0.63** -0.68*** Notes 0.52*** 0.57*** 0.67*** 0.74*** Systematic 0.12 ~0.04 0.35** 0.39*** Trial and Error ~0.11 ~0.11 New Approach ~0.61** ~0.37** 0.00 ~0.16 Fixated ~0.35** ~0.28* *p <.05 **p'<.01 ***p <.001 54 cant changes in the number of subjects who enjoyed working on the mys~ tery problems, who used a trial and error approach, or who were careful and cautious in their approaCh. Discussion The level of state anxiety in Group I did not show any increase from nonstress to stress conditions. Group II had a marked reduction in anxi- ety going from stress to nonstress conditions. One possible reason for this discrepancy between the groups is that a certain amount of anxiety may have been caused in Session 1 just because of the novelty of the sit- uation. Group II had this effect plus stress conditions in Session 1 and, hence, a higher'mean level of anxiety in that session. In Session 2, the ‘portion of anxiety due to novelty would.be predicted to decrease for all subjects since they had been in the situation before and knew more or less what to expect. The finding that bOth groups were less frustrated and.better able to concentrate in the second session suggests that the novelty effects were wearing off. In Session 2 Group I was given a stress treatment that was strong enough to counteract the reduction in anxiety due to the familiarity of the situation. Apparently the stress instruct- ions were less believeable in Session 2 or Group I probably would have increased in their level of anxiety over What they experienced in Session 1. Group II received the anxiety reduction due to familiarity with the situation in Session 2 plus the bonus of nonstress conditions, which brought their level of state anxiety even lower. Practice effects were found for most of the problem solving strate- gies. This corrOborates the suggestion based on low test-retest relia- bilities discussed in Section III that they represent cognitive behav- iors that are quite changeable. If the cognitive ability measures are 55 taken as valid indicators of style, then it would appear that with prac~ tice on those tests, subjects became more analytical (Hidden Figures Test), more flexible (Sign Changes Test, Object USes: % categories), and more im- pulsive (Object Uses: % part responses). If these tests are primarily measuring abilities, however, the improvements in scores could be due to increases in skill with practice. The question of the construct validity of these measures will be taken up in Section VI. VG. THE CONSTRUCT VALIDITY OF COGNITIVE ABILITY MEASURES OF COGNITIVE STYLES Up to this point it has been assumed that the cognitive ability neasures were valid indicators of cognitive styles. In this section both the convergent and discriminant validity of these measures will be investigated. In Section III it was shown that when the problem solving clusters were grouped together as measures of the styles indi- cated by their content, only the group of anxiety clusters had inter- correlations indicative of a common trait being shared between them. Those comparisons were made within one method of measurement, the problem solving questionaire measures. Another important comparison to be made is between the problem solving questionaire clusters and style measurenents made by different methods. The relationships between the problem solving neasures, cognitive ability measures, and personality scale measures that have been hypothesized to be indicators of the same styles will be examined in this section in order to check for their convergent validity. In addition, the discriminant validity of the cognitive ability measures of style will be investigated via cluster analysis to see if this method of neasurement discriminates between dif- ferent styles. If discriminant validity is present, cognitive ability neasures of different styles should emerge as separate factors. 56 57 m Table 2 in Section 11 presents all the measuring instruments used in this study. Several of them have not been involved in the analysis until now because the data.from themnwas not gathered under experimental condi- tions. Before the subjects were given any instructions as to the nature of the experiment, they were asked to provide some general information about themselves, such as their age and year in college. They were also given four scales to fill out which may be referred to in Appendix A. The first was a 24 item true—false reflectiveness-impulsiveness scale which was made up of items selected from the Impulsiveness Scale (Bar- ratt, 1959), the RI Rigidity Scale (Rehfisch, 1958), the RAPH Scale I (Meresko, Rubin, Shontz, and.MOrrow, 1954), the wesley Rigidity Scale (wesley, 1953), and also included a few items written by the experimen- ter. Second was the Achievement Anxiety Test (AAT) (Alpert and Haber, 1960); the name is misleading, as this is a measure of academic test anxiety. Subjects responded on a five point scale of anxiety reactions in test taking situations that went from "never" to "always". The third scale was a 20 item true-false flexibility-rigidity scale made up of items taken from the same scales as those used.by the experimenter in constructing the reflectiveness-impulsiveness scale. The fact that items concerning flexibility and impulsiveness were found on the same scales indicates that there is a good deal of confusion among the authors of these scales as to the difference between these two constructs. On the same page as the flexibility-rigidity scale were two items ‘written by the experimenter which were not included in that scale. "I get very tense and anxious when I think other people are disapproving of 58 ne" was used as a measure of disapproval anxiety. The other item, "I tend to lack self-confidence in my academic ability" was included so that relationships between this aspect of self-concept and performance could be investigated. The last scale was the trait anxiety form of the State-Trait Anxiety Inventory (STAI) (Spielberger, et. al., 1970). This scale is intended to measure individual differences in general anxiety proneness. It is a four point scale that goes from "almost never" to "almost always" experience certain kinds of anxiety reactions. The other preneasures used were high school grade point average (GPA) and college entrance examination percentile which were obtained from the University's Evaluation Services. Both are used as indicators of academic ability but GPA is more a measure of academic achievement and college entrance exam percentile is more a measure of academic aptitude. These percentiles were calculated for each class of new students at Michigan State University and are based on scores from the SAT, ACT, and (III. The experimental design and procedures used in collecting the other measures of cognitive styles and strategies, the problem solving effici- ency measures and the measures of state anxiety, have been discussed in previous sections. Reliabilities, means and point-biserial correlations between groups were computed for the preneasures. Then data from all the measures, pre and experimental, was pooled across groups and sessions and correlations were computed between all measures. Finally, the cor- relation matrix of all the measures but the problem solving questionaire clusters was reordered using the ORDER and ARRANGE routines provided in PACKAGE (Hunter and Cohen, 1969). Based on the resulting correlation matrices and similarity coefficients, a primary set of clusters was 59 formed. Then an oblique multiple groups factor analysis was performed on these clusters. They were revised and factor analyzed several times until they met the three criteria discussed in Section III for unique homogeneous clusters. Results Table 13 shows reliabilities, group means, and.point~biserial cor- relations for the premeasures. The split-half and similarity coeffi- cients obtained indicate that the STAI-Trait Scale is very homogeneous, ‘while the Flexibility Scale, AAT, and the Reflectiveness Scale are mod- erately so. Although subjects were assigned to groups on a random.basis, the group means and.point~biserial correlations show that in some ways the groups were different. Group I was significantly younger, had spent less time in college, was less reflective, and.more flexible as measured by the personality scale measures of style._ These group differences did not show up in any discernable way on the other measures in the study. Correlations between different kinds of measures of the same cogni- tive styles are presented in Table 14. For flexibility, only the verbal Problems Test and the Flexibility Scale were significantly related. For reflectiveness, significant correlations were obtained between the Re- flectiveness Scale and the Careful cluster and math Systematic. There *was also a significant correlation between.math Systematic and time on the Matching Familiar Figures Test. The matrix of global style measures did not produce any significant correlations. In summary, there is an Obvious lack of generality across the different kinds of measures for each style. These different methods of measurement appear to have lit- tle in conrnon as measures of style, hence, they may be said to lack con- 60 Table 13 Premeasures: Reliabilities and Group Differences Reliability* Mean Premeasure rpbi rxx 0 Group I Group II Reflectiveness Scale .55 .57 70.26 71.77 ~.20* AAT (Test Anxiety) .57 .64 53.60 53.60 .00 Flexibility Scale .64 .68 59.32 57.47 .28** STAI-Trait Scale .88 .92 47.93 45.63 .13 Academic Self-confidence 1.70 1.53 .18 Disapproval Anxiety 1.72 1.58 .14 High School GPA 3.16 3.02 .16 College Entrance Exam 52.39 53.14 ~.01 Percentile Score Year in College 1.28 1.83 ~.32*** Age 18.25 18.98 ~.27** Note.~~r coefficients are odd-even split half correlations for the Reflectiveness Scale, Flexibility Scale and STAI-Trait Scale. tenuation with the Spearman-Brown formula. *p'<.05 **p <.Ol ***p <.001 On the AAT the correlation is between the facilitative and debilitative test anxiety items. All are corrected for at- 61 Table 14 Correlations Between Cognitive Ability Measures, Personality Scale Measures, and Problem Solving Questionaire Measures of Styles and Anxiety Flex1b1l1ty 19}- 392 ill fl; Flexibility Scale 403 100 verbal Problems 409 19* 100 Sign Changes 411 ~01 17 100 O. U. % Categories 413 16 ~05 ~16 100 Myst. New Approach 518 ~07 ~13 08 ~09 " Strategy Chg. Sll ~01 ~09 ~06 05 " Trial 8 Error 517 ~11 ~02 09 ~08 " Fixated 519 ~02 ~03 ~06 ~06 Nhth New Approach 526 ~15 ~11 ~07 i10 " Strategy Chg, 522 ~12 ~18 ~07 03 Reflectiveness 591- fllé. 319. £12- Reflective Scale 401 100 MFF: correct 415 06 100 MFF: time 416 10 32*** 100 O.U. Total 412 04 ~09 ~12 100 Byst. Notes 515 ~11 l7 17 07 Systematic 516 02 06 13 05 " Deliberate 513 ~04 02 ~14 01 " Careful 512 21* 01 04 07 Math Notes 524 ~08 01 08 18 " Systematic 525 27** 08 21* 15 " Deliberate 523 08 01 09 10 Gldbalness 410 313- Hidden Figures 410 I00 O.U. % Parts 414 O4 100 Myst. Global 515 ~10 04 " Deliberate 513 01 10 Math Deliberate 523 ~10 04 551—651 402 .103. 191 32s 9.99 AAT (Test Anxiety) 402 I00' STAI-Trait Scale 404 43*** 100 STAI-State Scale 407 25** 48*** 100 Disapproval Anx. 405 10 36*** 16 100 Lack Aca. Self-Conf. 406 32*** 41*** 20* 13 100 ‘Myst. Frustrated 508 23* 27** 42*** 19* 22* " Concentrated 509 ~23* -21* -29** ~12 ~10 " Enjoyed 510 ~01 ~02 ~28** ~09 ~08 IMath Frustrated 520 29** 28** 47*** 08 27** " Concentrated 521 ~17 ~12 ~23* ~06 -15 *p <.05, **p< .01, ***p< .001 62 vergent validity for that purpose. The results of comparing the different measures of anxiety are sharply contrasting to those just reviewed. A large amount of generality was fomd among these measures. Here 60 percent of the intercorrelations are statistically significant. These measures show a good deal of evi- dence for convergent validity, i.e. , most appear to be measuring the same trait. The results of the cluster analysis of the cognitive ability mea- sures of style, problem solving efficiency measures and preneasures are presented by individual clusters in Tables 15 through 19. Each cluster will be discussed briefly. Table 15 presents the eight measures which made up the Ability cluster, so named because its strongest member is a measure of academic ability-~College entrance exam percentile. This cluster is considered to be a general measure of academic ability. The three weakest items are the math and mystery problems: correct, and high school GPA. GPA correlates about as well with the Time cluster as with the Ability cluster, but it was placed in the Ability cluster to preserve content homogeniety. The mystery problem correlates as highly with the Object Uses cluster, but it was also kept in the Ability clus~ ter for reasons of content. Three of the cognitive ability style nea- sures, Verbal Problems, Sign Changes, and Hidden Figues, show more var- iance in conmon with measures of academic ability than with any other measures. If there is some variance in each that is due to cognitive style, it is probably less than that of the ability component and was not enough to separate them in this analysis. The six neasures that make up the Anxiety cluster are shown in Table 16. All Correlate more highly with this cluster than with any other. 63 Table 15 Cluster 501: Ability Coefficient alpha = .65 426. College entrance exam percentile 409. verbal Problems Test 422. Solvable practice problems answered correctly 411. Sign Changes Test 410. Hidden Figures Test 419. Bath problems: correct 424. High school grade point average 417. Mystery problems: correct 426 409 422 411 410 419 424 417 426 55 57 35 21 31 19 21 17 409 57 49 27 17 24 24 28 13 422 35 27 35 23 3O 30 13 07 411 21 17 25 14 16 16 08 11 410 31 24 3O 16 14 ~08 11 O9 419 19 24 30 16 ~08 06 ~07 O3 424 21 28 .13 O8 11 ~07 05 ~01 501 75 70 59 38 37 24 23 18 Ability 502 ~31 ~16 ~18 ~21 ~25 05 ~12 ~09 Anxiety 503 ~09 01 04 ~15 14 ~08 ~02 l8 Object Uses 504 20 02 17 07 20 02 24 16 Time 505 05 19 03 ~01 20 07 ~03 01 Flexibility Scale 506 07 08 04 00 14 ~02 19 03 Reflectiveness Scale 507 ~05 ~07 02 06 10 ~04 ~06 ~03 Grade-Age Coefficient alpha 64 Table 16 Cluster 502: Anxiety = .67 404. 402. 407. 406. 408. 405. Trait Anxiety Scale Test Anxiety Scale State Anxiety Scale: Session 1 I tend to lack self-confidence in my academic ability. State Anxiety Scale: Session 2 I get very tense and anxious when I think other people are disap- proving of me. 404 402 407 406 408 405 501 502 503 504 505 506 507 404 80 43 41 41 36 36 ~18 90 08 ~17 ~20 ~11 ~04 402 43 24 26 32 13 10 ~14 49 01 01 ~20 ~16 ~03 407 41 26 23 08 28 20 ~17 48 18 03 ~02 ~08 ~12 406 41 32 08 20 14 13 ~21 45 04 ~20 ~10 ~13 23 408 36 13 28 24 17 06 ~20 41 ~23 05 ~22 03 ~06 405 36 10 20 13 06 10 05 31 ~07 ~17 ~22 ~~02 ~02 Ability Anxiety Object USes Time Flexibility Scale Reflectiveness Scale Grade-Age 65 Table 17 Cluster 503: Object USes Coefficient alpha = .62 413. 412. 414. Object USes Test: % of total uses in different categories. Object USes Test: Total number of uses listed. Object USes Test: % of total uses which required breaking the object into parts. 413 412 414 501 502 503 504 505 506 507 413 412 69 50 50 33 41 22 00 11 11 ~05 86 56 ~26 ~14 16 28 ~04 04 01 07 Ability Anxiety Object Uses Time Flexibility Scale Reflectiveness Scale Grade-Age 66 Table 18 Cluster 504: Time Coefficient alpha = .62 418. iMystery PrOblems: Time 416. Matching Familiar Figures Test: Time 420. Math Problems: Time 421. Unsolvable practice problems answered 415. Matching Familiar Figures Test: Correct 418 416 420 421 415 418 57 40 34 38 23 416 40 31 26 11 32 420 34 26 21 21 13 421 38 11 21 14 09 415 23 32 13 09 13 501 18 l4 14 15 19 Ability 502 ~16 ~02 ~06 ~12 ~01 Anxiety 503 ~24 ~01 ~09 ~14 ~13 Object Uses 504 77 56 46 37 35 Time 505 01 00 ~08 ~12 11 Flexibility Scale 506 02 10 19 13 O6 Reflectiveness Scale 507 ~05 05 ~20 ~02 ~06 Grade-Age 67 Table 19 Cluster 507: Grade-Age Coefficient alpha = .85 423. Class in college 425. Age 423 425 501 502 503 504 505 506 507 423 75 74 ~02 ~04 14 ~05 ~01 01 86 425 74 75 ~01 02 06 ~15 ~05 05 86 Ability Anxiety Object Uses Time Flexibility Scale Reflectiveness Scale Grade-Age 68 This cluster can safely be considered to be a measure of the general ten- dency to experience anxiety reactions or as a measure of anxiety prone- ness. Note that all but one of the members of this cluster (#405) corre- late negatively with the Ability cluster. The Object Uses cluster, made up of three closely related measures, is shown in Table 17. Although these three measures are closely related. both in the way they are mathematically derived and in content, they show lower intercorrelations than might be expected. This indicates that, to a certain extent, each is measuring a different trait. All are uncorre- lated with the Ability cluster and are positively correlated with the Flexibility Scale. Exactly what trait the Object Uses Task is measuring is not clear from this data. It may well be an indicator of semantic spontaneous flexibility as suggested by French, et. a1. (1963). Table 18 presents the five measures that make up the Time cluster. They are all positively correlated with the Ability cluster. The content of Unsolvable Practice Problem Answered and Matching Familiar Figures Test: Correct are not obviously homogeneous in content with the other three more direct measures of time. Correctly solving the Matching Familiar Figures Test items is more highly correlated with Time spent than with Ability. This indicates that doing well on this task is more a matter of taking enough time than it is having a special cognitive skill. Since both time and correct for the Matching Familiar Figures Test were most highly related to direct measures of time, there is good reason to believe that it measures a careful or reflective approach. A possible reason why the number of unsolvable practice problems that were answered joined the Time cluster is that subjects who put down an improb- able answer to a difficult problem just in case they may be right and 69 get credit may be more careful and cautious as they work on the kind of tasks used in this study, and being careful and cautious takes more tine. Nunber of unsolvable practice problems answered was correlated .22 with the Careful cluster which measures a careful, cautious approach. The Time cluster is considered as general measure of reflectiveness. The Reflectiveness Scale and Flexibility Scale each formed a sep- arate cluster. Their correlations with other clusters are shown in Table 20 in Section VII. The Flexibility Scale was most highly corre- lated with the Object Uses Cluster (r = .36) which may well be a mea- sure of cognitive flexibility. The Reflectiveness Scale was correlated nest highly with the Time cluster (r = .20). These significant correla~ tions are evidence for convergent validity of flexibility and reflec~ tiveness not found previously between unclustered separate measures. Table 19 presents the Grade-Age cluster whose meaning may be taken at face value. This cluster was not significantly correlated with any others. Discussion Intercorrelations between different methods of measurement which were used as indicators of the same cognitive styles were generally fomrd to be very low. When scores were summed across several neasures, however, by the formation of a cluster, significant correlations were found be- tween the Reflectiveness Scale and the Time cluster and between the Flexibility Scale and the Object Uses cluster. These correlations may be taken as convergent validity evidence for the existence of these traits. Individual measures. did not show strong convergent validity, hence, their utility as single measures of flexibility or reflectiveness is question- able. Evidence for the convergent validity of the global-analytical dimen- 70 sion was not found. This may be because fewer measures were used to assess this trait than were employed with the others. The results of the cluster analysis showed that the Verbal Problems Test and the Sign Oranges Test that have been used as cognitive flexi- bility measures and the Hidden Figures Test which has been used as a measure of global-analytic style are more similar to measures of academic ability than measures of cognitive styles. If cognitive styles do play a role in performance on these tasks, it is probably less important than that played by ability. It was hypothesized that the total number of responses on the Object Uses Task might be a measure of impulsivity in the sense of quick nonreflective thinking, that the percent of total uses in different categories was a measure of flexibility, and that the percent of total uses which required the breaking of objects into parts was a measure of analytical thinking. It was no surprise that the three neasures combined into one cluster, indicating that there was not enough variance unique to each measure for them to be used separately as indi~ cators of different styles. As one measure, the Object Uses Task is probably an indicator of some aspect of cognitive flexibility. The Matching Familiar Figures Test: time and correct was shown by the cluster analysis to be nere closely related to working speed than to any other neasure. Although it was a member of the Time cluster and the Time clus- ter was significantly correlated with the Reflectiveness Scale, neither the Matching Familiar Figures: time or correct scores alone were signif- icantly related to the Reflectiveness Scale. The Matching Familiar Figures Test showed more evidence for construct validity than the other cognitive ability measures of style, but there were still some inconsis- tencies shown which weaken the confidence that can be placed in it as an 71 adequate measure of reflectiveness. .After an analysis of the convergent and discriminant validity of the cognitive ability measures, it is concluded that, in general, they cannot be relied upon when used individually as adequate measures of specific cognitive styles. There is some evidence that the Matching Familiar Figures Test measured reflectiveness and that the Object USes Task measured flexibility. The Hidden Figures Test, Sign Changes Test, and verbal Problem Test have questionable construct validity as measures of cognitive style. VII. PROFILE DIFFERENCES BETWEEN HIGH AND LOW ANXIOUS SUBJECTS When this study was undertaken it was planned that the relationships between anxiety, cognitive styles and strategies, and intellectual per~ formance would be investigated with both experimental and correlational nethods. The effects of stressful treatment conditions and practice ef- fects on these variables were examined in Section Iv and V. In this sec- tion, correlational data is used in looking at the relationships between these variables. The specific question that was proposed for analysis was whether subjects who are more anxiety prone tend to exhibit differ- ent patterns of cognitive styles and strategies than subjects who are less anxiety prone. The results of the analyses covered in Sections III and VI have shown that although cognitive styles may exist as indi- vidual traits, they were not identified with the methods used in this study. Hence, the relationships discussed here are restricted to mea~ sures of anxiety, academic ability, cognitive strategies used on the problem solving tasks, and the personality scale measures of flexibility and reflectiveness. Method The data for all subjects was pooled across groups and sessions. Intercorrelations were computed between the 26 major clusters. These included the 12 mystery problem clusters and the seven math problem clusters described in Section III and the seven clusters just described in Section VI. The resulting matrix was investigated for patterns of 72 73 relationships which involved large numbers of clusters. Results The intercorrelation matrix of the 26 major clusters was divided into two tables. Correlations among the problem solving questionaire clusters were presented in Table 6; Table 20 shows the remaining por- tion of the matrix. The two clusters that had the largest number of significant correlations with other clusters were the Anxiety and Ability clusters. The Anxiety cluster was positively related to Frus- trated, Strategy Change, and Notes on the math problem only, and neg~ atively correlated with Ability, the Flexibility Scale, Concentrated, Enjoyed and Careful. The Ability cluster was positively associated with Time, Concentrated, Careful and math Systematic, and was negatively related to Anxiety, Frustrated, and Fixated. It was unexpected that Fixated, i.e. , sticking with the sane hypothesis, would be a function of ability rather than anxiety and that Strategy Change would be a function of anxiety rather than ability. Two distinct profiles emerged from the above relationships in regard to the Anxiety cluster. One profile shows low anxiety proneness with high academic ability, taking ample time on cognitive tasks, and the following problem solving characteristics: carefulness, good concentration, enjoying the problem, not getting frustrated, and not using the sane hypothesis over and over again. The other profile shows high tendency toward anxiety, poor per~ formance on measures of academic ability, rigidity as measured by the flexibility-rigidity scale, and the following problem solving character- istics: lack of carefulness, poor concentration, not enjoying working on the problem, getting frustrated, and making several strategy changes. Several of the clusters were not highly correlated with either Abil~ 74 Table 20 Correlations Among the Twenty Six Major Clusters flflfléi‘flififliéfl Ability 501 100 ~37*** 01 32*** 15 16 ~02 Anxiety 502 ~37*** 100 01 ~15 ~31*** ~18 ~01 Object USes 503 01 01 100 ~25** 36*** 05 12 Time 504 32*** ~15 ~25** 100 ~03 20* ~11 Flexibility Scale 505 15 ~31*** 36*** ~03 100 ~16 ~03 Reflectiveness Scale 506 16 ~18 05 20* ~16 100 O3 Grade-Age 507 ~02 ~01 12 ~11 ~03 O3 100 IMystegy Clusters Frustrated ~20* 52*** ~08 14 ~21* 00 ~09 Concentrated 26** ~37*** 02 21* 01 23* ~01 Enjoyed 16 ~22* 15 05 14 12 ~09 Strategy Change ~07 33*** 07 ~07 ~01 ~12 ~11 Careful 28** ~31*** 10 09 06 21* 03 Deliberate ~17 04 01 ~17 04 ~04 07 Global ~14 01 20* ~08 ~03 ~12 ~08 Notes 08 06 04 35*** 07 ~11 12 Systematic 06 ~16 l4 13 15 02 04 'Trial and Error ~15 02 ~18 ~20* ~11 ~06 04 New Approach ~03 ~07 ~07 05 ~07 02 14 Fixated ~21* 03 ~09 00 02 ~07 ~03 Math Clusters Frustrated' ~20* 55*** ~08 19* ~28** ~02 O4 Concentrated 21* -29** ~22* ~01 05 22* ~02 Strategy Change ~06 25** O4 28** ~12 05 00 Deliberate ~13 05 08 08 ~07 08 01 Notes 14 19* 16 24** ~08 ~08 06 Systematic 23* 04 07 23* ~02 27** ~01 New Approach ~08 ~05 02 ~08 ~15 ~04 01 *p <.05 **p <.01 ***p <.001 75 ity or Anxiety. Notes was most highly correlated with Time. Trial and Error was negatively related to the Flexibility Scale. New Approach was most strongly correlated with Deliberate. Object USes was positively related to the Flexibility Scale and Global and negatively related to Time. The Reflectiveness Scale was just below significance in negative correlation with Anxiety. It was positively correlated with Time, Con- centrated, Careful, and math Systematic. Grade-Age was not significantly correlated with any other cluster. Two items that were part of the Anxiety cluster have not been cov- ered elsewhere and will be mentioned briefly here. These correlations are not presented in any of the tables. Lacking self-confidence in academic ability was negatively correlated with Ability (r = r.21), Time (r = ~.20), and Systematic on the mystery problem (r = ~.20) and was positively related to Anxiety (r = .45), Frustrated (mystery r = .22, math r = .27), and Grade-Age (r = .23). Anxiety over disapproval from others was negatively related to Deliberate on the mystery problem (r = ~.21) and the Flexibility Scale (r = ~.22) and was correlated.pos~ itively with Anxiety (r = .31), Fixated (r .20), and Notes on the math problem (r = .20). Discussion The relationships found here between high anxiety proneness and low academic ability have a number of possible explanations. One pos~ sibility is that people who actually have low academic ability find their ego or self-esteem threatened every time they are faced with failure experiences in academic situations. There may be external pun- ishments like flunking out of school, not being able to pursue the ca- reer of one's choice, etc., associated with poor academic performance 76 that elicit anxiety reactions in people of low ability. Both of these explanations are based on the assumption that the ability (i.e. , poten- tial) of these people is low and, therefore, anxiety reactions to aca- demically stressful situations are the natural and probably unavoidable result. Another possibility is that some peOple who have high potential for academic achievement become so anxious when they work on academic tasks that their performance is inhibited and consequently they get lower scores on measures of academic ability than they would if they were not so anxious. That is, although ability plays the largest role in deter- mining academic performance, whatever the level of a person's ability, his performance can be hindered by high levels of anxiety. But how does this happen? The results of this section suggest some possibilities that are consistent with the hypothesis proposed by Wine (1971), that it is the worry component of anxiety that interferes with performance by taking attention away from the task. Subjects who lacked self-confidence in their academic ability tended to spend less time on the tasks in this study and were nere easily discouraged and worried (Frustrated cluster). They scored lower on the measures of academic ability and were more prone toward anxiety reactions. This confirms the frequently found re- lationship between low self-opinions and anxiety proneness (Sarason, 1960) and suggests the possibility that poor performance in these sub- jects may be due, in part, to a lack of persistence. Also, the strong relationships found here between anxiety proneness, changing strategies, poor concentration, lack of carefulness, and poor academic performance suggests that anxiety reactions may be accompanied by alterations in cognitive functioning that could conceivably be contributing causes to 77 poor performance. Poor concentration, lack of carefulness, and strategy changes may all be indications of inability to maintain attention toward the task. VIII. SUMMARY AND CONCLUSIONS Four general questions were asked in this study: (1) Will the neasured anxiety, cognitive styles and strategies, and problem solving efficiency of subjects vary systematically with changes in psychological- ly stressful conditions? This was asked for two reasons: (a) because cognitive styles have been defined as relatively stable individual traits that should be found consistently across time and in different situations, and (b) because of the possibility that problem solving ef- ficiency may be affected by changes in styles and strategies and these, in turn, may be related to anxiety reactions. Although the treatment conditions used in the experimental phase of this study were psycholog- ically stressful enough to produce differences between groups in state anxiety, they did not affect performance on congnitive ability tests thought to be measures of cognitive styles, problem solving strategies, or problem solving efficiency to a significant degree. Hence, no dif- ferences between groups in strategy or style usage or in problem solving efficiency caused by differences in stress conditions were measured. The results of the cluster analysis of the problem solving question- aires reported in Section III showed that working carefully and main- taining good concentration were helpful in solving the mystery problem. In Section VII lack of carefulness and poor concentration were found to be characteristics of anxiety prone subjects, and anxiety proneness showed a weak negative relationship (r = ~.09) with accurately solving 78 79 the mystery problem. The size of this correlation is precisely what would be expected if it represents an indirect relationship, thus being a function of the correlation between carefulness or concentration and anx- iety proneness (avg. r = ~.34), times the correlation between carefulness or concentration and accurately solving the problem (avg. r = .25). These re- sults suggests that an indirect causal chain was present between anxiety proneness, carefulness and concentration, and problem solving performance. (2) Will subjects who are measured as nere anxiety prone tend to exhibit different patterns of cognitive styles and strategies than less anxiety prone subjects? This question was asked to investigate the possibility thatganxiety prone students tend to do poorly on neasures of academic ability because they adopt maladaptive styles or strategies when they become anxious which interfere with their performance. The results presented in Section VII showed that high anxiety prone subjects tended to score lower on measures of academic ability, be less flexible as measured by the flexibility-rigidity scale, and exhibited the follow~ ing problem solving characteristics: lack of carefulness, poor concen- tration, disliked working on the problem, getting frustrated, and mak- ing several strategy changes. Some hypotheses were suggested as to how these characteristics could adversely affect academic performance. (3) Will subjects be measured as using the same cognitive styles consistently under different experimental conditions and across time? Performance on the cognitive ability tests thought to be measures of style was generally unaffected by changes in treatment conditions. On the other hand, there were changes across time that looked like practice affects on mental ability test. Later it was found that nest of these tests nere closely resembled measures of ability than measures of styles. The use of problem solving strategies changed over time, but these stra- 80 tegies do not represent styles as they are comnenly defined. So the two kinds of style measures that were used to answer this question about the stability of style usage were found to be invalid as measures of style. Hence, the question of the consistency with which cognitive styles are used cannot be answered with this data. (4) Will the different measures of cognitive styles be shown to have construct validity? This question was answered negatively in Sec- tions III and VI. Neither the problem solving questionaire measures of a priori styles nor the cognitive ability measures of style were found to have convergent or discriminant validity as style measures. Some evidence was obtained which indicated that the Matching Familiar Figures Test measured reflectiveness and that the Object Uses Task measured flexibility, but the adequacy of each as single measures of those char- acteristics is questionable. The Hidden Figures, Sign Changes, and Ver- bal Problems Tests behaved nere like tests of abilites than measures of cognitive styles. In general, the findings indicate that if cognitive styles do exist as individual traits, they are very difficult to identi- fy. If they are to be reliably and validly measured, batteries of tests will probably need to be constructed for each style. Using a single test to assess the use of a cognitive style is likely to be inadequate and may well lead to assumptions about its presence that are entirely Lurjustified. Although general cognitive processing characteristics were not found that fit the usual definitions of cognitive styles, a good deal of evidence was obtained which indicates that certain aspects of processing which are nere specific in nature than is suggested by the term "style" were identified. These were called cognitive strategies and were found 81 to have some generality across time and problem types. Until the valid~ ity of cognitive style measures are more firmly established, it may be more profitable in research on individual differences in cognitive pro- cessing to focus on specific aspects of functioning and to use caution when making assumptions which imply the existence of general processing traits. LIST OF REFERENCES LIST OF REFERENCES Alpert, R., and Haber, R. N. Anxiety in academic achievement situations. Journal 9f Abnormal and Social Psychology, 1960, 61, 207-215. Barratt, E. S. Anxiety and impulsiveness related to psychomotor effi- ciency. Perceptual and Motor Skills, 1959, 9, 191-198. Campbell, D. T., and Fiske, D. W. Convergent and discriminant valida- tion by the multitrait-multimethod.matrix. Psychological Bulletin, 1959, 56, 81-105. Cattell, R. B., and Scheier, I. H. The nature of anxiety:.A review of thirteen multivariate analyses comprising 814 variables. Ps2; chological Reports, 1958, 4, 351-388. Cattell, R. B., and Scheier, I. H. The meaning and measurement 9f.neuro~ ticism and anxiety. New York: Ronald, 1961. Cattell, R. B., and Scheier, I. H. Handbook for the IPAT Anxiety Scale (2nd ed.). Champaign, Illinois: Institute for Personality and Ability Testing, 1963. Cronbach, L. J. Coefficient alpha and the internal structure of tests. Psychometrika, 1951, 16, 297-334. Cronbach, L. J. Essentials 9f_Psychological Testing (3rd ed.). New York: Harper and Row, I970. Iiriksen, C. W., and Wechsler, H. Some effects of experimentally induced anxiety upon discrimination behavior. Journal gf_Abnormal and Social Psychology, 1955, 51, 458-463. Iiysenck, J. J. Check your own I. Q,_ Middlesex, England: Penguin, 1966. Frederiksen, C. H. Abilities, transfer, and information retrieval in verbal learning. Multivariate Behavioral Research Monographs, 1969, 2. Frederiksen, J. R. Cognitive factors in the recognition of ambiguous auditory and visual stimuli. Journal of Personality and Social PsychologngOnographs, 1967, 7, (Whole—No. 639{) French, J. W., Ekstrom, R. B., and Price, L. A. Manual for a kit of ref~ erence tests for cognitive factors. (Rev. ed.) Princetofij'New Jersey: EducaEIOnal Testing Servite, 1963. 82 83 Goldner, R. H. Individual differences in whole-part approach and flexi- bility-rigidity in problem solving. Psychological Monographs, 1957, (Whole No. 450). Guilford, J. P. Fmdanental statistics in psychology and education. New York; McGraw-Hill, 1965. Hunter, J. B., and Cohen, S. H. Package: A system of computer routines for the analysis of correlational data. Educational and Psy- chological Measurement, 1969, 29, 697-700. Jackson, D. N. Personality research form manual. Goshen, New York: Research Psychologists Press, Inc., 1967. Jackson, D. N., Messick, S., and Myers, C. T. Evaluation of group and individual forms of embedded- figures measures of field-inde- pendence. Educational and Psychological Measurement, 1964, 24, 177-192. Kagan, J. Impulsive and reflective children. In J. D. Krumboltz (Ed.) Learning and the educational process. Chicago: Rand McNally, 1965a. Kagan, J. Reflection-impulsivity and reading ability in primary grade children. Child Development, 1965b, 36, 609-628. Kagan, J. Preferred nodes of conception: consistency and significance of an analytic attitude. 1965c (Unpublished). Referred to in Allen, S. D. , The research on cognitive style, 1965 (Unpublished manuscript, Dept. of Psychology, University of California, Berkeley). Levitt, E. E. The Psychology 91; anxiety. New York: Bobbs-Merrill, 1967. Liebert, R. M., and Morris, L. W. Cognitive and enetional components of test anxiety: A distinction and some initial data. Psycholog- ical Reports, 1967, 20, 975-978. Maier, N. R. F., and Burke, R. J. Response availability as a factor in the problem solving performance of males and females. Journal 9f Personality and Social Psychology, 1967, 5, 304-310. Mandler, G., and Sarason, S. B. A study of anxiety and learning. Journ- §_l_9_f Abnormal and Social Psycholcgy, 1952, 47, 166-173. Marshall, R. L. Self-generated complexity, cognitive abilities and stra- te ies as determiners of clue utilization in problem solving. igctoral- dissertatfinj—University of Califarnia, Berkeley). Ann Arbor, Michigan: University Microfilms, 1972. No. 72-18, 541. Meresko, R., Rubin, M., Shontz, F. C., and Morrow, W. R. Rigidity of attitudes regarding personal habits and its idealogical corre- lates. Journal 9f Abnormal and Social Psychology, 1954, 49, 89-93. 84 Messer, S. B. The effect of anxiety over intellectual performance on reflection-impulsivity in children. Child Development, 1970, 41, 723-735. Rehfisch, J. M. A scale for personality rigidity. Journal 9f Consult- ingPsychology, 1958, 22, 10-15. ""‘“ “"'““ Rosenberg, M. The association between self-esteem and anxiety. Journal 2f Psychiatric Research, 1962, 1, 135-151. Ruebush, B. K. Interfering and facilitating effects of test anxiety. Journal gt: Abnormal and Social Psychology, 1960, Saltz, E. Manifest anxiety: Have we misread the data? Psychological Review, 1970, 6, 568-573. Sarason, I. G. Interrelationships among individual difference variables, behavior in psycholtherapy and verbal conditioning. Journal 3f Abnormal and Social Psychology, 1958, 56, 339-344. Sarason, I. G. Empirical findings and theoretical problems in the use of anxiety scales. Psychological Bulletin, 1960, 57, 403-415. Sarason, S. B., Davidson, K. S., Lighthall, F. F., Waite, W. R., and Ruebush, B. K. Anxiety _i_p_ elementary school children. New York: John Wiley and Sons, 1960. Sarbin, T. R. Ontology recapitulates philology: The mythic nature of anxiety. Anerican Psychologist, 1968, 23, 411-418. Schroder, H. M. and Hunt, D. E. Failure-avoidance in situational inter- pretation and problem solving. Psychological Monographs, 1957, 71, (Whole No. 432). Shulman, L. S., Loupe, M. J ., and Piper, R. M. Studies 9f the inquiry rocess (Final report). East Lansing, MiEElgan: EducatTonaI PGBImation Services, College of Education, Michigan State University, 1968. Spielberger, C. D. and Gaudry, E. Anxiety and educational achievement. Sidney: Wiley, 1971. Spielberger, C. D., Gorsuch, R. L., and Lushene, R. E. Manual for the State-Trait Anxiety Inventog. Palo Alto, California: Consult- ing Psychologists Press, 1970. Suinn, R. M., and Hill, H. Influence of anxiety on the relationship between self-acceptance and acceptance of others. Journal gr: Consulting Psychology, 1964, 28, 116-119. Taylor, J. A. A personality scale of manifest anxiety. Journal 9: Abnormal and Social Psychology, 1953, 48, 285-290. 85 Summers, G. J. New puzzles ip logical deduction. New York: Dover, 1968. Wand, B. Flexibility in intellectual performance. Technical report NR 151-113, 1958. Educational Testing Service, Office of Naval Research. Wesley, E. Perseverative behavior, manifest anxiety, and rigidity. Journal pf Abnormal and Social Psychology, 1953, 48, 129-134. Wine, J. Test anxiety and direction of attention. Psychological Bulletin, 1971, 76, 92-104. Witkin, H. A. Individual differences in ease of perception of embedded figures. Journal 9_f_ Personality, 1950, 19, 1-15. Witkin, H. A. Origins of cognitive style, In C. Scheerer (Ed.) Cogr_1_ition: Theogz, research, promise. New York: Harper and Row, 1964. Wrightsman, L. S. The effects of anxiety, achievement, motivation, and task importance upon performance on an intelligence test. Journal _g_f_ Educational Psychology, 1962, 53, 150-156. Zuckerman, M. The development of an affect adjective check list for the neasurement of anxiety. Journal pf Consulting Psychology, 1960 , 24, 457-462. APPENDICES APPENDIX A: PREMEASURES 86 Directions: U o :2 .". O - n 'I Blacken in l for true and 2 for false. IL I am very slow in making up my mind..... ...... . ................. ... 2.1 rather like the idea of having my meals at odd hours and of...... going to bed when the mood strikes me. ii I am often the last one to give up trying to do a thing............ 1A I prefer work that requires a great deal of attention to detail.... &.My friends consider me to be happy-go-lucky........................ 0.I find it easy to stick to a certain schedule, once I have......... started on it. 'L I don't like to work with slow people..... ....... . ..... . ...... ..... 8. I usually have a ready answer 9. I like to do things on the spur of the moment.. ......... ........... um I am a careful person in whatever I do.............. ....... . ..... .. 11. I usually check more than once to be sure that I have locked a..... 2A. door, have everything I planned to-take when I leave the house,etc. . I am inclined to go from one activity to another without...........: continuing with any one for too long a time. . I sometimes do dangerous things Just for the thrill of it.......... . When reading a newspaper or magazine I skip around alot rather..... than going through it from beginning to end in a methodical manner. . I usually dislike to set asideratask that I have undertaken........ until it is finished. I like work requiring patience and carefulness..................... If I had.to choose which plays the greatest role in my decision..... making, I'd say its probably emotions rather than logic. .IusuallYLhinkbEfOI‘EIlea-1)..........o..................ooooo.... . I don't like to wait for traffic lights to change ...... ............ . I am always on time for social events..............................~ . In watching games I usually don't yell along with the others....... At times I feel I can make up my mind with unusually great ease.... Use Only a Number 2 Pencil — Do Not Fold MSU OS 10‘ Listed below are a number of statements concerning attitudes and traits. Read each item and decide whether the statement is true or false as it pertains to you personally. C . My interests tend to change quickly......... ......... .............. .5 U . I have less trouble concentrating than other people seem to have...—-~ [N1 am,“ Sun u‘wm") ""'"" V‘ r‘ Directions : 8 7 of the ways people react to taking examinations. never- l. 2. 3. h. 5. 6. 7. 8. 9. 10. ll. 12. 13. 1h. 16. W. L8. -9. occasionally about half the time often always ‘flf‘f‘b’ii’ I work most effectively under pressure, as when the test is. . . .. . . very important. Nervousness while taking an exam or test hinders me from..... . . . . . ' doing well. In a course where I have been doing poorly, my fear of a bad...... grade cuts down nor efficiency. While I may (or may not) be nervous before taking an exam, once... I start, I seem to forget to be nervous. When I am poorly prepared for an exam or test, I get upset, and. .. do less well than even my restricted knowledge should allow. The more important the examination, the less well I seem to do.... Nervousness while taking a test helps me do better.. ..... .. When I start a test, nothing is able to distract me............... In courses in which the total grade is based mainly on- 9_n_e_ exam,.. I seem to do better than other people. During exams or tests, I block on questions to which I know..... .. the answers , even though I might remember them as soon as the exam is over. I100kforwardtoexmooooooooooooocoo-o...-00.000000000000000.no I find that my mind goes blank at the beginning of an exam and... . it takes me a few minutes before I can function. Although "arming" under pre-examination tension is not..... ...... effective for most people, I find that if the need arises, I can learn material immediately before an exam, even under considerable pressure, and successfully retain it to use on the exam. I am so tired from worrying about an exam, that I find I.......... almost don't care how well I do by the time I start the test. I enjoy taking a difficult exam more than an easy one............. Time pressure on an exam causes me to do worse than the rest...... of the group under similar conditions. The more important the exam or test, the better I seem to do...... I find myself reading exam questions without understanding...... .. them, and I must go back over them so that they will make sense. When I don't do well on a difficult item at the beginning......... of an exam, it tends to upset me so that I block on even easy questions later on. Use Only a Number 2 Pencil — Do Not Fold M51) 0510‘ Blacken in = r f 5' 9 9‘ =97 4 °° a, a number from 1 to 5 to the right of each statement to indicate “ _ g R .. ‘ _ _ how closely it describes your behavior when taking examinations "f‘fffff‘fffn or tests in school. Listed below are a number of statements which describe some w) [NH (M) V...) (5‘. U‘ LAJ 1:: 0‘ d w w in) (a) (b: J‘ J‘ U.1 .nJ (a) J" 1 .. 5 I \J ..— 6 '8‘. (CD) (CD) M'i‘ it!" "~me ’""‘""' ’ 88 Directions: Listed below are a number of self-descriptive statements. Read each of them and decide whether the statement is true or Tii‘ ? 9 f D ” ? false as it pertains to you personally. Blacken in 1 for true ‘0 and 2 for false. 10. ll. l2. 19. 20. 21. 22. Once I.make up my mind about something, it is very hard to......... convince me to change. . Sudden unexpected changes of plans are usually upsetting to me..... . I prefer doing one thing at a time to keeping several projects..... going at once. . I like a great deal of variety in my work.......................... . I do not like to change my plans in the midst of an undertaking... . I often find.myself thinking the same tunes or phrases all......... day long. . I like to think up puns and plays on words... ..... 1 I have thought of several good inventions.......................... I dislike having to learn new ways of doing things I already....... know how to do. I always put on and take off my clothes in the same order.......... I like work which allows me to work at the same task rather than. .. having to switch around from one task to another. When I get stuck working a math problem, I generally keep trying... to solve it for a long time rather than taking a break and coming back to the problem later. When I am stuck on a.math problem I usually try to find a new...... approach rather than continuing with my original strategy. I often think of novel ways to use common objects.................. . When another person disagrees with me, it is often hard for me..... to see things from their point of view. When I can't solve a.math problem I usually try to forget what..... I have done and start all over from scratch. My interests probably range over a larger number of areas.......... than those of most of my friends. . When I can‘t get the right answer to czmath problem I usually...... try to find my mistake by doing the problem the same way over and over again until I find my error. Iliketo thirlkup new 'jOkeSOOOO......OOOOOOOOOOOO......OOOOOOOOOOO I often come up with new ways of doing things that others.......... haven't thought of. I get very tense and anxious when I think other people are......... disapproving of me. I tend to lack self-confidence in my academic ability.............. Use Only a Number 2 Pencil — Do Not Fold MSU 0510‘ P3 LN)! fiJ: u {a 30‘ [Ca V x I! r1 re r ¢ (£- (0‘ (0 u (£41 A on 3 NJ C...) I&’ ... V new.» Sut- war-run ’r‘ ' ‘ Directions: 89 1 2 (3 4 :5: 6 I R J “ “ I "' I l= almost never 2= sometimes 3= often h: almost always lIIfeel plewantOCCOOCOOO0.000.000.0000. 0000000000 ......OOOOCOCOOOOO 2OItire quiCklyOCOOOOOI.0.......OOOO-OOOOOO‘OOI0.0.0.0.0....0.0.0.0... 30Ifeel like eminECOCOOO ...... ......OOOOCCC......OOOOOCOOOOOOOOOOOO 1L I wish I could be as happy as others seem to be............ ...... .. &.I am losing out on things because I can't make up my mind.......... 10. 12. 13. it. soon enough. Ifeel rested...OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO0.00.000.00.09. I am "calm, cool, and collected"................ ...... ............. . I feel that difficulties are piling up so that I cannot............ overcome them. I worry too much over something that really doesn't matter......... I am.happy......................... ...... .......................... a I am inClined to take things hard.......................o.......... IlaCk self-confidenceOOOO0.0.0.000.........OOOOOOOOOOOOO00.0.00... I feel secure....................... ..... . ...... ................... I try to avoid facing a crisis or difficulity...................... 'OIfeel blueOOOOOOOOOOIOOO0............OOOOOOOOOOOOOOO0.00.00.00.00. I am contentoooocooo oooooo 000000000000000.000.000.000...000000.000. . Some unimportant thought runs through my mind and bothers me....... . I take disappointments so keenly that I can't put them out of...... my mind. 0 I am a Steady perSOnoooo.......................o............oooooo. . I become tense and upset when I think about my present concerns.... Copyright 1968 by Charles D. Spielberger Use Only a Number 2 Pencil — Do Not Fold MSU 05104 A number of statements which people have used to describe _ themselves are given below. Read each statement and then F.,.;. 9 s . m a blacken in the number to the right of the statement to indicate how you generally feel. There are no right or wrong answers. ; Do not spend too much time on any one statement but give the answer which seems to describe how you generally feel. ['07 L.“ ., \l T u. M!» snv- www-fir " APPENDIX B: INSTRUCTIONS 90 SECTION II: Problems The purpose of this study is to learn more about differences in the way people think as they work on problems. MOSt of the materials you are about to work on are designed to measure what is called "cognitive style." Your work will not be rated as "good" or "bad," rather measures of your performance will be used to separate the participants in this study into groups according to the different ways they like to work on problems. Then these preferences will be correlated with personality traits. You have alreamrfilled out some personality questionaires. Some people work fast, others prefer to take their time. Some people are global in their approach to problems while others are more analytical. What's best is what is most comfortable for you. We are interested in learning more about people's preferences along these lines. Some of the problems you are going to work on may seem difficult and others may seem.easy. This is to be expected. You will probably be doing Inatter'than you feel like you are doing. It is hoped that this explanation Inns relieved any sense of competition, anxiety, or worries you may have had. We want you to be as relaxed as possible as you go through these materials. You will be given five minutes to work on some practice problems before we begin. WAIT HERE UNTIL TOLD TO GO ON ... :44 . 91 SECTION III: Measuring Intellectual Abilities Due to the nature of some of the variables under study, it was not pos— sible to tell you the full purpose of this-study until now. Our primary con- cern is with intelligence. The problems you worked on in Section II were taken from.an intelligence test we are developing that will be used with high school students. Students with college-level aptitudes generally find little difficulty with materials designed for use with people of the high school level. That is why the problems were so easy for you. An important part of intellgence test development has to do with isolating fectors which may affect performance on the test. Since there is a large pool of subjects available here at M.S.U., we are using college students to study the relationships between certain personality traits and IQ. We have developed some intelligence measures for students with college- level aptitudes that are very similar to the problems you have Just finished. Thus, the problems you will work on in this section are part of a more advanced test of the same intellectual abilities. It is crucial that you do your very best on this section, otherwise the results will be invalid. WOrk quickly, but do not make foolish mistakes. You will have five minutes to do some practice problems before beginning the actual IQ test. WAIT HERE UNTIL TOLD TO GO ON 92 SECTION II: Measuring Intellectual Abilities We are currently in the process of developing an IQ test that will be used with high school students. An important part of intel- ligence test development has to do with isolating factors which may affect performance on the test. Since there is a large pool of subjects available here at M.S.U., we are using college students to study the relationships between certain personality traits and per- formance on this measure of intelligence. You are about to work on a series of problems that have been established as valid measures of IQ. You have already filled out the personality questionaires. It is crucial that you do your very best on these problems, otherwise the results of this study will be invalid. Work quickly but do not make foolish mistakes. You probably have college—level aptitudes, so these problems, which were designed fer use with high school students, may not seem very dif- ficult. In fact, most college students find them quite easy. You will be given five minutes to do some practice problems before beginning on the actual IQ test. WAIT HERE UNTIL TOLD TO GO ON 93 SECTIONzIII: Problems Due to the nature of the variables under study, it was not possible to explain the exact purpose of this experiment until now. These materials were designed to help us learn more about individual differences in the way people work on problems. We are trying to find out if a stressful testing situation will affect people's style or method of approaching problem solving tasks. In an effort to make you anxious, we told you that you were taking an IQ test and we gave you some unaflyable practice problems. The problems you worked on were not taken from an IQ test. They are designed to measure preferences in thinking styles, not intelligence. Your work will not be rated as "good" or "bad," rather performance measures will be used to separate participants into groups according to the different ways they like to work on problems. Some people work fast, others prefer to take their time. Some peOple are more global or analytic than others. What's best is what is most comfortable for you. We are interested in learning more about people's preferences along these lines. Don't worry if you found some of the problems difficult. Most people do much better on them than they feel like they are doing. It is hoped that this ex- planation will relieve any sense of competition, anxiety, or worries you.may have had, We want you to be as relaxed as possible as you go through similar materials in this section. You will be given five minutes to work on some practice problems. WAIT HERE UNTIL TOLD TO GO ON APPENDIX C: PRACTICE PROBLEMS 94 Directions : Circle your answer or fill in the blank. 1. Holy is to Slob as Cold is to? A. Born B. Old C. Glow D. Bow 2. Find the one that doesn't belong: /? //l/ l! !> \J l A B ./ . i\ D E 3. Complete the number series: 25, 2o, 16, 13, ________ ’4. Rearrange the letters to make a word: BACHE 5. Complete the letter series: A, C, F, J, 6. Find the one that doesn't belong: // \ / / A B D WAIT HERE UNTIL TOLD TO GO ON \ ,3” \ s1: OI 95 Directions: Circle your answer or fill in the blank. 1. Find the one that doesn't belong: ...- “"——K". ._._n-. i l ".7 ! ,<§ i Ll 3......3/713 L_;___ ‘ [34 l A B C l TI; r-" Eifi ' l "' ‘ 1 9 l ‘. ; ILL“ ’ quggfl .__il;i D E F 2. Rearrange the letters to make a word: GOHAW I Fill in the missing number: 18, 20, , 25, 32, hi Unusual is to Unprepossessing as Undulating_is to? L., A. Dubious B. Preponderance C. Congruent D. Continuous E. Roundabout -—— h -... _li._. !! U/fic l .u #33371 ' I / I <1 11 h ..-..- ”ii ( iiiq. :7 [ i, £_+:UJ 4;“ _ A B C I": A " 4 :23:- L :4 ______-f ., 441; I -fl__.1 l ‘. v g..- Lilli D E F i Complete the series of letters: X, V, S, L, WAIT HERE UNTIL TOLD TO GO ON APPENDIX D: CDCNITIVE ABILITY MEASURES OF COGNITIVE STYLES 96 Verbal Problems Directions: In this task you are to find the words in the paragraph in the box.below which could have the same meanings as each of the numbered.words or phrases. When you.have found the required word for each numbered item.write it in the space provided. Choose only one word for each blank space. Here is a sample item: unclouded CL£I%7,L/ On line 3 in the paragraph below you will seethe word ”clear” which can mean “unclouded" and therefore it has been written in the space provided. Now go right ahead, working as quickly as possible. You will have 10 minutes. I managed to get into action earlier than usual this morning, dressing by the first beam.of sunlight which crept in through my window. When I started off for town, looking very trim in my new outfit, the air was clear and it was silent except for the lonely bark of a dog. Two boys were making their way to the playing field, but otherwise I was quite alone on the street. I ploughed through a mass of leaves at the gate thinking of the errands I must go on. ”First, I should go to the bank and draw out some money in case the music for there is that meeting to attend..." -- l. steep SIOpe 13. disentangle 2. piece of timber 14. care for! 3. publicize 15. deed 4. law suit 16. box 5. submit 17. border of a lake 6. manner 18. lay away 7. incline 19. smile 8. performing 20. 'part.of'the body 9. pages 21. prune 10. sauce to add to 22. small sailing certain dishes vessel 11. equip 23. application to a wound 12. to give material or:iora1 support 24. in a flag, he to background ..-- Mpfi_ .—-——-—__-—.—._—-l _ -".--— WAIT HERE UNTIL TOLD TO GO ON 1 store wants cash for the violin bow I ordered, and then I must be back by eleven “ l GYW AAAAA AAAAA AAAAA AAAAA AAAAAA AAAAA w A m . ram AAAAA AAAAA AAAAAA 98 Part 2 l 8-li= 2 6-5= 3 5+1i-7= h h—2x2= 5 5+3—2= 6 6+2+3= 7 l+3-S+h= 10. 3+2+9-T= ll. h+h+h-l= 12. 6+3-2x2= 13- 7+3-5X3= lh. 8+ U.) I ..a N \J'l ll 15. 6+h—Ax7= WAIT HERE UNTIL TOLD TO GO ON 99 Object Uses Directions: On the following page is the name of a familiar object. Write down all the different ways you can think of in which the object might be used. Do not hesitate to write down whatever waijyou can think of in which the object might be used as long as they are possible uses for the object that is named. You will have 5 minutes. WAIT HERE UNTIL TOLD TO GO ON 100 A CARDBOARD BOX 101 APPENDIX E: PROBLEM SOLVING TASKS 102 Malice and.Alice Instructions: USe the infermation presented in the fellowing paragraph to answer the questions which follow it. When you have solved this mystery turn the page and record the time. Here are the facts: (1) Alice, Alice's brother, her son, and.her daughter were involved.in a.murder. (2) One of the feur killed one of the other three. (3) TWO of them who are of the same sex were in a bar at the time of the murder. (4) The victim.and the killer were together on a deserted beach at the time of the murder. (5) The vicitm's twin and the killer are of the opposite sex. (6) The victim.and the killer are of the same age. Fill in the blanks: (Alice, brother, son, daughter) A. The killer was B. The victim was__‘ .,.- ,- ...‘.—.—.—._.m."———- C. The two in the bar'were_fl_ _._..—.- .v-..—._.,_-_._._————_-..-_l_..-—. ¢Hm GO ON TO THE NEXT PAGE 103 Murder in the Family Instructions: Use the information presented in the following paragraph to answer to questions which follow it. When you have solved this mystery problem, turn the page and record the time. Here are the facts: (1)Murder occurred one evening in the home of a married couple and their son and daughter. (2)0ne member of the family murdered another member, the third member witnessed the crime, and the fourth member was an accessory (an accessory is one who was absent but who contributed to the crime). (3)T‘he accessory and the witness were of opposite sex. (4)The oldest member and the witness were of opposite sex. (5)The youngest member and the victim were of opposite sex. (6)The accessory was older than the victim. (7) The father was the oldest member. (8) The killer was not the youngest member. Fill in the blanks: (father, mother, son, daughter) A. The killer was The victim was B C. The witness was D The accessory was GO ON TO THE NEXT PAGE 104 The Hotel Room Problem Three traveling salesmen stopped at a hotel one night and stayed in three separate rooms. They were charged $10 fer each room, so the total bill came to $30. The next day, however, the desk clerk discovered that a mistake had been made and that the bill should.have been only $25. Accordingly, he gave the bellboy the $5 to distribute among the three men. As it happened, the bellboy was not entirely honest. He gave each man only $1 in return and kept $2 for himself. Choose from the calculations on your paper the one which most accurately depicts how the men's money was spent. ____1 30-1=27+2 ___2 30-2=27+1 _____3 1=3-2 __4 25+2=27 __5 s-3=2 105 The Horse Trading Problem A farmer went to an auction and bought a horse fer $60. But on his way home he met a neighbor who wanted to buy the horse, so he sold it to him for $ 70. Later that night the farmer decided.he still wanted the horse, so he went over to his neighbor's and.bought it back paying $80. The following day he was offered $90 for the horse so he sold it. How did the farmer come out financially in the horse trading business? 1 Lost $10 2. Broke even 3. Made $10 4 ‘Made $20 5 Made $30 APPENDIX F: PROBLEM SOLVING QUESTIONAI RES 106 . ‘_...— -Ndnu Listed below are a number of statements that describe the way people may think and feel as they work on mystery problems ,z* - ‘ ‘ ‘ like the one you have just finished. Read each item and decide ‘ ' ' “ ' ‘ whether it is true or false as it pertains to you personally. ( 1 completed. .more than once. othis particular problem rather than just using a familiar attack. 9. 1o. 11. 12. 13. 1n. 15. 16. 17. 1a. 19. 20. ’21. 22. f ' Blacken in,l for true and 2 for false depending on how you ‘ “ ‘ “ thought or felt as you worked on the problem you have just T F I have worked this problem before and remembered how to solve it.. I changed my strategy or approach for solving the problem at....... least once. I changed my strategy or approach for solving the problem.......... My first approach was based on my previous experience with......... similar problems. My first approach was based on what seemed to be the demands of.... O .Q'. The first strategy I used.was basically trial and arror, i.e.,..... 0 I chose a person at random to be the killer, etc., then checked to ' ' C ‘5 "'L' 'L see if that role for the person fit the facts. I m , My first approach was to list all the facts given about each....... 0 individual or role. ' My first strategy was to list all possible roles for each of the.... (3 four persons, then systematically eliminate them by checking all the facts .- a u .- . '- , a When I discovered that a hypothesis I was trying out ran counter... 6 M ' to the facts, I quickly abandoned it, never to try it again. It seems to me that this problem could have more than one.......... d :': correct solution. I found myself trying to use the same hypothesis again and again... 5 2' I always checked out the facts and clues in the same order......... a i .as they were presented in the problem. 3I stuck with an unproductive hypothesis for a long time even....... 8 'though it didn't seem to be getting me anywhere. I did most of my thinking without the aid of a pencil.............. 5 " I used pencil and paper to make notes as I worked on the problem... é u I used a rational systematic approach at all times as I worked..... é f on the problem. ” There were times when I abandoned logic and used a non-............ 9 ‘systematic approach. ‘ After working for some time, I discovered that I had been.......... 9 overlooking some important element of the problem. I was careful and cautious throughout the entire time I worked..... 9 :' on.the problem. Sometimes I checked out hunches that seemed improbable even... s . though I knew they would take time and energy. I read the problem several times before deciding how to first...... attack it. I decided on my first approach or strategy immediately after....... reading the problem.once. I didn't write down any answers until after I was positive......... that all the ffig'fingaquanbggPencil — Do Not Fold CONTINUED WIOQ 1M (‘1 \. w: (b) J‘ *1 \J on . inn 1.7-14si.789’ ON NEXT PAGE Tan: ’37 9 ‘4 Mum-'flMW" ”My ' 107 I 23. More than once, I wrote down an answer that I later changed. . . . . . . '29- I guessed at some answers before time was up '25- I guessed at some answers when time was called '26-'One thing I did was reading the problem over and over in an. . . . effort to get the entire picture. 270 Each time I got an idea about what r'ole a particular person. .. might play, I checked the facts for that person as well as check- 5the facts fer all the other persons in their new roles. 280-I tried to visualize the entire problem as a unit 29. I concentrated mostly nn the details presented in the problem..... 3°. I concentrated mostly on the generalizations or assumptions. . . . . . . 1, one could make from the facts. 310 I really didn't expect to get the answers by logical deduction.... .I expected them to come as insights from thinking about the .. problem. 32.“ I had a little trouble deciding which facts were necessary........ ifor solving the problem and which were not. 33.31 enjoyed working on this problem '3“.~At times I felt discouraged.ooooooooooooooono.00000009000000.0000.i.I, 35. I took this problem as an interesting challenge. . . . . . . . . . . . . . . . . . . 35. Isuspected some trick solution.... ......... . ..... 370‘ I quit working on the problem at least once. . . . . . . . . . ......... . . . . L? 38. When I first read the problem I didn't think I'd be able to ..... . . solve it. 390 There were times when I couldn't seem to think “0011 had quite a bit of trouble concentrating ...... .................. MVSometimes my mind wandered........................................ “2.' At times I worried that I might not be able to get the right..... . ? answer. “3. éSometimes I wondered how well the other studeutb'vcrc (10'! ng... . . . . 1 on this problem. . lAf‘tni' work-ing on tho lu‘nflflt'm 8 Vhl' 1.8 I began to feel frustrated... i lu5.:This prOblem‘waS easy for me to SOlveooooocoo-00000000000000.0000. TF U .a Use Only a Number 2 Pencil — Do Not Fold WAIT HERE UNTIL TOLD 'm 00 ON 1 ‘ l usu 05104 U-r'fl'" 3r!" ”“‘" ' Directions: 103 Listed below are a number of statements that describe the way people may think and feel as they work on math problems like the one you just finished. Read each item and decide whether the statement is true or false as it pertains to you personally. Blacken in l for true and 2 for false depending on how you thought or felt as you worked on the problem you have just completed. T F I have worked this problem.before and remembered how to solve it... 1. I changed my strategy or approach for solving the problem at....... ' least once. 2. I changed my strategy or approach for solving the problem more..... than once. 3. My first approach was based on my previous experience with......... 'similar problems. 9. My first approach was chosen not because I had used it before...... but'because it seemed appropriate to use for this specific problem. jaliworked the problem only once..................................... 6. I worked the problem more than once, doing it more than one way.. .. 7.]Zobtained more than one answer before deciding what the........... , 1 correct answer was. &.fllworked the problem more than once, the same way each time........ i just to check my computational accuracy. 9. I wrote most of my calculations down on paper. ....... .............. .1 10,; I worked most of my calculations in my head. . . . . . . . . . . . . . . ..... . . . . w :r [1. I made notes as the problem was read............................... 12. I tried to use a rational systematic plan to solve the problem..... 13. I worked the problem with no specific plan in mind......... ..... ... .1 l“. I thought about the problem briefly before I went to work.......... with the actual calculations. . [5._I began.calculating immediately after the problem was read......... - ' l [6,.I guessed at the answer before time was up..... ..... ............... I7. I guessed at the answer when time was called....................... 18. I tried to visualize the overall picture painted by the............ 'information given. 19. I forgot about the persons and objects mentioned in the... ........ . problem and figured it with manipulations of numbers. 20. This problem was easy for me to solve.............................. 21. I enjoyed.working on this problem.................................. 22. At times I felt discouraged 23. I took this problem as an interesting challenge.................... 24. After working on “the prob] em for a while I began to feel frustrated Use Only a Number 2 Pencil - Do Not Fold CONTIM'ED NEXT PAGE MSU 05-104 109 I 23- 26. 37. 25. ' 29. 9. . 31. . 9. 33. I suspected some trick solution.............. ............... ....... Sometimes I wondered how well the other students were doing ........ -on this problem. I quit working on the problem at least once............. ...... ..... At times I worried that I might not be able to get the right....... answer. When I first read the problem I didn't think I'd be able to ....... . solve it. Sometimes my mind wandered........................................ fiThere were times when I couldn't seem to think..................... ;I had quite a bit of trouble concentrating......................... ‘It seems to me that this problem could have more than one ...... ... correct solution. I changed.my answer at least once ..... . ....... . ......... .......... WAIT HERE UNTIL TOLD TO GO ON Use Only a Number 2 Pencil — Do Not Fold MSU»OS-104 T F APPENDIX G: REVISED STAI-STATE SCALE Directions: 110 a r 3 a s f c i s s A number of statements which people have used to describe themselves are given below. Read each statement and then blacken in the number to the right of the statement to indicate 0 l 1 1 4 '3 h 7 b H . how you felt while working on the problems. There are no right 1 “ t T 3 f‘?“f “ or wrong answers. Do not spend too much time on any one state- ,_ ‘ H q ,, ,' ment but give the answer which seems to most accurately describe ° ’ 2 T f T‘?Z:Ef’ how you felt while working on the problems. l= not at all 2‘ somewhat 3= moderately so 0 , J 3 a 5,, h= very much so 1. I felt calm....................................................... j ; 2 , , 1.; 2. I felt secure..................................................... n , , ,.,.,g , , 3. I was tense....................................................... C 1.1. 4 r g {p ’4. I was regretful................................................... :1 1:11,, 91.81pm 5. I felt at ease.................................................... o 1 r'i: 5 b ;;-1 6. I felt upset...................................................... 01.11.15u1... 7. I was worrying over possible misfortunes.......................... 1 1; 3 1 5 1 ;1 8. I felt rested..................................................... u 1 I 31.12.11 y w 9. I felt anxious.................................................... 0 1 1 :1: 1 L. r. 1. 10. I felt comfortable................................................ 0'11 3 Q 5 5.1;.g 11. I felt self-confident............................................. 0 E g j i 517} § 12. IfeltnervouSOO......OOOODOOOOIOOOOO0..........OOOOOOOOOOOOOOOOOO f;:“:23:;;561;8” 13. Iww JittewOOOOOO000......0.0.000.........OOOOOOOOOO.........Ol. 0‘“ ‘2 3‘23: 5.2,.71'8‘.” *J L- J‘ 11‘. Iffilt "high stmgnooooooooooooooooo00.0....0.0000000000000000...Ci“, r. ....;§_/ 9 H A t." C: \J (I; t. 15. IV” relaxed.........OOOOOOOOOIOOOOOOO00.0.00.........OOOOOOOOOOO«:;.:“.r,2‘:3:_::.U l6. Ifelt contentOOOO......OOOOOOOOOOO0.0.0.........OOOOOOOOOOOOOOOOO f). ... “4.1.: 17. IVES worriedooooooooo000.00.00.00...ooooooooooooooooooooooooooooo u... 18. Ifelt over-eXCited wd rattled-............OOOOOOOOOOOOCOOOO...... 0 ‘1 2) 3 2‘: r‘ b 2 g ” 19. Ifelt Joymooooooooococoa-00.00.00.ooooooooooooocooooooooooooooo"‘- :“4~ 20. Ifeltp1888811t.............o...o..................o........o..o.. “ “ “ .4 «UV '1 l " 3 4 5 l. I g '1 H l 3 3 1 '5‘ 1 7 11 1' l 2 7 .4 r. t R i 11 l .7 1 I. a I P 1 Use Only a Number 2 Pencil — Do Not Fold 'm Una—w. F ...... ...... MC.“ OS :04 APPENDIX H: CLUSTER TABLES 111 Table H-l Cluster 509: Mystery Concentrated .. £10. I had quite a bit of trouble concentrating. - 39. There were times when I couldn't seem to think. . £11. Sometimes my mind wandered. - 37. I quit working on the problem at least once. Form A Form 8 £10 39 £11 37 £10 39 £11 37 £10 £13 £13 31 111 61 51 £11 33 39 ‘43 39 22 20 51 36 26 28 £11 31 22 17 11 £11 26 26 25 37 1’4 20 11 08 33 28 25 21 508 ~35 -50 -18 ~10 ~18 -7+5 1’31 32 Frustrat—ed 509 66 63 £11 27 79 60 50 £16 Concentrated 510 20 oo 33 17 2£1 07 39 15 Enjoyed 511 -26 411 -19 ~08 -° -31 ~15 -32 Strategy Change 512 3:1 2L1 19 08 39 £19 32 18 Careful . 513 06 02 -1O -11 -20 OO 05 -2£1 Deliberate 51£1 -15 --32 -11 23 -2 5 -1£1 -1£1 ~16 Global 515 -03 ~04 y 03 08 05 09 18 Notes 516 25 17 27 16 25 33 21 21 Systematic 51? ~04 ~07 -2 5 ~06 ~2£1 ~09 ~03 ~0£+ Trial and Error 518 ~11 22 -03 ~11 --20 01+ -07 --17 New Approach 519 -07 ~09 - 02 Oil -20 ~27 '01 -18 Fixated 599 -0£1 -0£1 -12 09 05 oo 11 00 rpm 112 Table 1H-2 Cluster 510: Mystery Enjoyed 33. I enjoyed working on this problem. 35. I took this problem as an interesting challenge. .lzazigi. 1_§aza_§_ 33 35 33 35 33 52 19 68 66 35 “9 52 66 68 308 -39 -22 -30 -22 Frustrated ' 509 38 12 29 30 Concentrated 510 71 71 82 82 Enjoyed 511 -17 -10 ~03 10 Strategy Change' 512 1? 111 20 13 Careful - 513 ~16 -O6 -08 07 Deliberate 514 -04 01 03 05 Global 515 -19 10 15 1? Notes 516 25 16 15 14 Systematic 517 -51 -12 -03 -17 Trial and Error 518 -05 ~17 -28 ~05 New Approach 519 -02 -16 -09 .06 Fixated 599 12 17 02 ~02 rpbi 113 Table H.3 Cluster 511: Mystery Strategy Change ... ”no- —_-—.._ 2. I changed my strategy or approach for solving the problem more than once. 1. I changed my strategy or approach for solving the problem at least once. 31. I really didn't expect to get the answers by logical deduction: I expected them to come as insights from thinking about the problem. 17. After working for some time I discovered that I had been overlooking _flfi some important element of the problem. ' I Form A Form B 2 1 31 17 2 1 31 17 2 48 149 29 12 36 65 05 17 1 £19 36 20 13 65 62 15 25 31 29 20 18 15 05 15 07 25 #1] 12 13 15 06 17 25 25 17 508 27 ’28 28 32 25 26 09 31 Frustrated 509 -19 ~27 -28 -19 -1L; -19 -33 -32 Concentrated 51o ~03 -20 -15 00 -16 08 -01 ~14 Enjoyed 511 70 6O #1 23 60 81 25 41 Strategy Change 512 -32 ~21 ~49 -33 -29 -27 -22 -3# Careful 513 11 08' 26 ~07 - ' 28 25 19 Deliberate 51a 32 11+ 15 20 14 20 19 22 Global 515 O9 22 -O6 O7 28 10 ~05 16 Notes 516 .27 ~30 -36 -13 -17 -26 ~27 -20 Systematic 517 17 18 04' ~07 05 ~12 -04 13 Trial and Error 518 00 ~05 22 -01 -011 08 20 03 New Approach 519 10 -11 -02 06 -21 ~24 03 -12 Fixated 599 12 02 ~10 09 07 ~06 02 -07 rpm 114 Table H~4 Cluster 512 : Mys tery Careful . ~ 24. I guessed at some answers before time was up. ~ 23. More than once I wrote down an answer that I later changed. ~ 12. I stuck with an unproductive hypothesis for a long time even though it didn't seem to be getting me anywhere. 22. I didn't write down any answers until after I was positive that all the facts checked out. 18. I was careful and cautious throughout the entire time I worked on the problem. ~ 19. Sometimes I checked out hunches that seemed improbable even though I knew they would. take time and energy. Form A Form B 24 23 12 22. 18 19 24 23 12 22 18 ' 19 24 49 41 29 27 31 12 19 15 17 20 21 . 25 23 41 26 24 31 08 06 15 15 11 25 27 11 12 29 24 23 14 22 14 17 11 22 25 25 26 22 27 31 14 16 04 15 20 25 25 20 19 13 18 31 08 22 01+ , 11 13 21 27 25 19 28 21+ 19 12 O6 14 15 13 06 25 11 26 13 21+ 19 508 ~38 -12 ~28 -19 - 18 -21 4+1 ~32 J40 ~25 ~38 ~23 Frustratef 509 17 05 33 23 24 14 33 07 25 09 43 43 Concentrated 510 12 -11 -02 24 16 14 20 ~18 12 17 1“ 09 Enjoyed 511 4+6 -31 -17 -21 -116 -23 ~37 ~12 ~23 05 ~L15 ~36 Strategy Change 512 71 51 I17 110 34 25 44 39 47 1&5 53 144 Careful . 513 ~03 ~10 .03 06 -09 -15 09 11+ 15 29 01 ~01 Deliberate 514 ~23 -01 ~09 _ 05 ~16 -42 ~18 08 -06 13 03 -02 Global 1115 -01 -02 -21 03 -10 -06 ~08 O8 , 09 02 ~08 08 Notes 516 24 1 11 26 111 08 L18 1’4 211 17 26 35 Systematic 517 -35 -08 07 --‘3,4 -07 -20 ~15 ~12 ~28 01 ~18 ~24 Trial and Error 518 -06 03 03 -05 -13 -03 03 27 11+ 20 10 04 New Approach 519 ~14 ~04 ..33 ~10 ~08 ~11 10 25 -211 02 ~08 -12 Fixated 599 ~09 14 -11 -05 00 -09 02 ~08 00 «01+ 01; ~13 rpbi 115 Table H~5 Cluster 5131 Mystery Deliberate 20. I read the problem several times before deciding how to first attack it. - 21. I decided on my first approach or strategy immediately after reading the problem once. Form A Form B 20 21 20 21 . 20 9 45 6o 57 r?“ 21 49 452 57 60 508 06 -03 ~06 O6 Frustrated 509 ~03 ~07 ~15 ~11 Concentrated 510 -03 -19 -03 02 Enjoyed 511 07 21 28 22 Strategy Change 512 ~09 ~09 24 24 Careful 513 71 71 77 77 Deliberate 514 O7 13 22 17 Global 515 09 02 ~08 ~06 Notes 516 ~16 ~17 01 ~09 Systematic 517 00 13 06 08 Trial and Error 518 03 08 39 36 New Approach 519 09 '06 05 ~07 Fixated 599 09 -04 ~02 -02 rpm Cluster 514: 116 TR bl 6 117-6 Mystery Global ~ 29. I concentrated mostly on the details presented in the problem. 26. One thing I did was reading the problem over and over in an effort to get the entire picture. 30. I concentrated mostly on the generalizations or assumptions one could make from the facts. 36. I suspected some trick solution. 28. I tried to visualize the entire problem as a unit. Ford A Form B 29 26 3O ' 36 28 29 26 3O 36 28 29 24 09‘ 39 07 09 32» 07 33 18 35 26 O9 21 19 27 ~28 O7 11 10 14 14 30 39 O6 11 11 ~07 33 10 15 13 03 36 O7 19 11 O9 07 18 14 13 12 12 28 09 27 ~07 07 06 35 14 13 12 35 508 ~06— 28 12 18 09 ~02 08 -03 08 32 Frustratedw 509 ~21 12 ~23 ~03 03 ~12 ~19 ~06 ~14 ~14 Concentrated 510 03 04 ~01 ~19 08 ~15 02 16 ~01 09 Enjoyed 511 24 03 18 18 11 01 28 18 11 22 Strategy Change 512 ~06 ~18 ~18 ~06 ~09 08 ~ ' 03 ~12 C4 Careful 513 O4 18 03 03 ~02 00 10 19 14 15 Deliberate 514' “9 “5 33 30 23 56 33 39 311 59 GlObal 515 01 12 02 O9 20 ~08 ~12 07 ~10 ~07 Notes 516 ~22 03 ~11 03 12 19 ~10 01 ~04 12 Systematic 517 00 ~09 ‘ 26 07 ~17 03 ~14 ~13 16 ~15 Trial and Error 518 20 ~02 ~03 00 06 ~02 18 20 00 13 New Approach 519 '06 26 '03 28 10 '13 07 00 ‘13 02 Fixated 599 ~04 00 00 c6 02 04 ~13 09 07 02 rpm 117. "Table H-7 'Cluster 515: Mystery Notes 14., I used pencil and paper to make notes as I worked on the problem. ~ 13. I did most of my thinking without the aid of a pencil. Form A Form B 14 13 14 13 14 49 “5 55 52 13 “5 49 52 55 508 23 12 “03 03 Frustrated 509 O8 11 11 14 Concentrated 510 ~04 ~05 17 11 Enjoyed 511 13 11 24 17 Strategy Change 512 ~06 ~13 ~01 06 Careful 513 05 05 ~15 01 Deliberate 514’ 31 02 ~12 ~07 Global 515 69 69 73 73 NOtGS 516 08 08 ~02 06 Systematic 517 ~23 03 ~24 —O9 Trial and Error 518 ~12 ~13 ~09 ~11 New Approach 519 07 20 -01 '07 Fixated 599 04 -02 -02 -02 rpbi 118 Table H-8 Cluster 516: Mystery Systematic 15. I used a rational systematic approach at all times as I worked on the problem. - 16. There were times when I abandoned logic and used a nonsystematic approach. Form A Form B 15 16 15 16 15 76 71+ 71+ 72 15 74 76 72 74 508 '18 ‘-22 -33 ‘-36—Frustrated 509 35 4O 44 29 Concentrated 510 24 26 20 11 Enjoyed 511 -44 ~50 ~37 -3 Strategy Change 512 43 37 55 ' 50 Careful 513 -24 -16 03 -11 Deliberate 514 -02 -12 06 08 Global 515 00 20 03 01 Notes 516 87 87 86 86 Systematic 517 ~30 -52 -24 -25 Trial and Error ‘ 518 -26 -21 -17 ~01 New Approach 519 14 32 '06 04 Fixated 599 ~04 ~05 02 05 rpm 119 Table H-9 Cluster 51?: Mystery Trial and Error -‘6. My first approach was to list all the facts given about each individual role. 5. The first strategy I used was basically trial and error, i. e., I chose a perSOn at random to be the killer, etc., then checked ‘ to see if that role for the person fit the facts. Form A Form B _ . _ m5 6 5 6 '5 6 32 28 “0 26 5 28 32 36 0 508 '01 23 01 “17 Frustrated 509 -14 -09 -12 ~10 Concentrated 510 -28 -21 ~16 oo Enjoyed 511 -09 26 ~02 04 Strategy Change 512 ~06 -33 -19 -25 Careful 513 ~05 15 ~06 17 Deliberate 514 -14 18 -15 02 Global 515 -17 -04 -26 -02 Notes 516 -29 ~22 ~16 ~20 Systematic 517 55 55 62 62 Trial and Error 518 05 ~27 06 -08 New Approach 519 '21 '07 '15 '08 Fixated 120 Table H-IO Cluster 518: Mystery New Approach A My first approach was based on my previous experience with similar My first approach was based on what seemedto‘be the demands of this particular problem rather than just using a familiar attack. ., 3, problems. 4. Form A _fgrm B 3 u 3 u 3 38 3’4 47 an :1; =3"; 38 41+ 41 508 "21 ~08 ~04 '19 Frustrated 509 -08 06 -21 01 Concentrated 510 ~02 ~16 ~18 ~09 Enjoyed 511 ~02 12 21 -03 Strategy Change 512 ~12 O3 05 33 Careful 513 13 -03 36 30 Deliberate 514 ~02 16 20 10 Global 515 -15 -O7 -13 -05 Notes 516 -21+ -08 ~2L; 1o Systematic 517 O7 -31 11 -13 Trial and Error 518 60 6O 67 67 New Approach 519 02 15 ~02 ~11 Fixated 599 00 13 -07 09 rpbl 121 Table 3911 Cluster 519: Mystery Fixated 10. I found myself trying to use the same hypothesis again and again.‘ - 8. When I discovered that a hypothesis I was trying out ran counter to the facts I quickly abandonediit never to try it again. Form A _Ij‘orm B 10 8 10 '8 10 35 33 36 3’3 8 . 32 36 32 36 508’ 26 02 15 -1O Frustrated 509 ~07' '01 ~23 -20 Concentrated 510 -21 05 01 ~03 Enjoyed 511 -01 03 -12 ~19 Strategy Change 512 -24 -13 -O4 O7 Careful 513 -01 O4 -04 O3 Deliberate 514 20 16 -06 -03 Global 515 08 16 ~08 02 Notes 516 18 05 - 12 10 Systematic 51? ~03 ~27 —03 -18 Trial and Error 518 ~02 19 —10 -01 New Approach 519 58 58 58 58 Fixated 122 Table H-12 Mystery Questionaire Residual Items --7. My first strategy was to list all possible roles for each of the fcur persons, then systematically eliminate them by checking all the facts. 9. It seems to me that this problem could have more than one correct solution. -11. I always checked out the facts and clues in the same order as they were presented in the problem. -25. I guessed at some answers when time was called. ~27. Each time I got an idea about what role a particular person might play, I checked the facts for that person as well as checking the facts for all the other persons in their new roles. 32. I had a little trouble deciding which facts were necessary for solving the problem and which were not. ' 438. When I first read the problem I didn't think I'd be able to solve it. Form A ‘ 'Form B ' 7 9 11‘ 25 27" 32 38 7 9 11 25 27 .32 38 7 100 -10 05 02 05 02 23 21 01+ 17 00 "17 -