A CLUSYER ANALYSIS OF AN OBJECTIVE MEASURE OF ACABEMIC MOIWAHON Thests {or Hm Degree of DH. D. MECHEGAN STATE UNIVERSITY Fred William Smith £965 u—- .h. - ._ 7- AA...- \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\l\\\\\\\\\§\\§\,\\\§l 3 1293 10359 Michtgan State University This is to certify that the thesis entitled A CLUSTER ANALYSIS OF AN OBJECTIVE MEASURE OF ACADEMIC MOTIVATION presented by Fred William Smith has been accepted towards fulfillment of the requirements for Ph.D. degree in Education Date 0c tober ' 0-169 FEB 2 0 2007 ABSTRACT A CLUSTER ANALYSIS OF AN OBJECTIVE MEASURE OF ACADEMIC MOTIVATION by Fred William Smith The major concern in this study was directed toward the isola- tion and description of the dimensions of an objective measure of aca- demic motivation called the Michigan M-Scales.1 It was hypothesized that meaningful and descriptive dimensions related to academic motiva- tion can be isolated and identified. The responses of 240 male eleventh grade high school subjects to the 139 male items in the M-Scales were subjected to McQuitty's Elementary Linkage Analysis (ELA) procedure. A cluster as identified by E.L.A. was defined as a structure in which every item is more like some other item in that cluster than any item in any other cluster. The clusters identified were labeled and de- scribed using the characteristics relative to academic motivation held in common by their constituent items. All of the clusters except Cluster XIV were internally consistent. Two elaborations of E.L.A., the Hierarchical Method and the Comprehensive Hierarchical Approach, each yielded two interpretable second order clusters. Only the clus- ters from the latter method were interpreted. labeled Fred William Smith The Elementary Linkage Analysis yielded fourteen clusters as follows: Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster I--Distractable II--Irresponsible III--Need Achiever IV--Competition Versus Ease of Meeting a Standard V--Intellectua1 Versus Material Accomplishment VI--Immediate Versus Long-Term Involvement VII--Occupationa1 ReSponsibility: High Versus Low VIII--OCCUpational Self-Development: High Versus Low ‘IX--Compulsivity X--Se1f-Perceived Competence XI--Self—Perceived ReSponsibleness XII-~Self-Perceived Intellectual Ambition XIII--Need Achievement: High Versus Low XIV--Self-Perceived Non-Conformity The second order clusters identified by the Comprehensive Approach were labeled as follows: Cluster Cluster A--Responsibleness: High Versus Low B--Self-Perceived Ability l The study was conducted as part of a larger research project Sponsored by the United States Office of Education under the direction of William W. Farquhar. A CLUSTER ANALYSIS OF AN OBJECTIVE MEASURE OF ACADEMIC MOTIVATION BY Fred William Smith A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Personnel Services, and Educational Psychology 1965 \ d) \3! 0;} "\l \J ACKNOWLEDGEMENTS The writer is gratefully appreciative of the invaluable assis- tance given by the following peOple toward the successful completion of this study: To Dr. William W. Farquhar, major advisor, for his interest, guidance and encouragement. To doctoral committee members Dr. James Costar, Dr. John X. Jamrich and Dr. John Useem for their guidance and cooperation. To Dr. Louis L. McQuitty, originator of the analysis procedure used, for his interest and technical assistance. And foremost, to his wife, Jo Anne, daughter, Sheryl, and son, Eric, for their sacrifice, patience and inSpiration. ii TABLE OF CONTENTS Page ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . ii LIST OF TABLES . . . . . . . . . . . . . . . . . . v Chapter I 0 THE PROBLEM O O O O O O O O O O O O O O O O O O 1 Purpose of the Study . . . . . . . . . . . . . 1 Need for the Study . . . . . . . . . . . . 2 Statement of the Problem . . . . . . . . . 4 Theory 0 O O O O O C O O O O O O O O O O O O O O 5 The Hypothesis . . . . . . . . . . . . . . . 7 Statement of the Hypothesis . . . . . . 7 Organization of the Study . . . . . . . . . . . 8 II. REVIEW OF THE LITERATURE . . . . . . . . . . . . . . . 9 Motivation as a Factor in Academic Achievement . . . 9 Personality Traits Related to Academic Motivation . 10 Dimensions of Measures of Academic Motivation . . . 14 sumary O O O O O O O O O O O O O O O O O O O 18 III. DESIGN AND METHODOLOGY . . . . . . . . . 19 Review of the Farquhar Study . . . . . . . . . . l9 sampling 0 O O I O O C O O O O O O O O O O O O O 19 Instrument Construction Techniques . . . . . 20 Analysis Procedures . . . . . 20 ~ Reliability of the M-Scales . . . . . . . 20 Validity of the M-Scales . . . . . . . . . 21 Multiple Correlation Estimation . . . . . 21 Hypothesis . . . . . . . . . . . . . 21 Statement of the Hypothesis . . . . . . . . 22 Elementary Linkage Analysis . . . . . . . . . . 22 Analysis Procedures . . . . . . . . . . . . . . . 25 Reference Items . . . . . . . . . . . . . . . 26 iii Chapter page Elaborations of Elementary Linkage Analysis . . . . . 28 Hierarchical Method . . . . . . . . . . . . . . . . 28 Comprehensive Hierarchical Approach . . . . . . . . 28 Summary . . . . . . . . . . . . . . . . . . . . . . . 30 IV. ANALYSIS OF RESULTS . . . . . . . . . . . . . . . . . . 32 Interpretation and Discussion of the Clusters . . . . 34 Discussion of Elaborations of Elementary Linkage Analysis .. ..... . . . ... . ..... . . . . . . . . . 60 Identification of Reference Items . . . . . . . . . 60 Second Order Clusters . . . . . . . . . . . . . . . 62 Summary . . . . . . . . . . . . . . . . . . . . . . . 68 V. SUMMARY, CONCLUSIONS AND IMPLICATIONS . . . . . . . . . 72 Summary . . . . . . . . . . . . . . . . . . . . . . . 72 Conclusions . . . . . . . . . . . . . . . . . . . . . 77 Discussion . . . . . . . . . . . . . . . . . . . . . . 79 Implications for Future Research . . . . . . . . . . . 82 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . 84 APPENDIX A, INTERNAL CONSISTENCY . . . . . . . . . . . . . . . . 86 APPENDIX B, CLUSTER MATRICES AND LINKAGE PATTERNS . . . . . . . 87 APPENDIX C, THE HIERARCHICAL METHOD . . . . . . . . . . . . . . 101 APPENDIX D, THE COMPREHENSIVE HIERARCHICAL APPROACH . . . . . . 106 iv Table 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 LIST OF TABLES Number of Items From Each Sub-Scale Fourteen Clusters ‘ Summary of Item Content and Indices for Cluster I Summary of Item Content and Indices for Cluster II . Summary of Item Content and Indices for Cluster III Summary of Item Content and Indices for Cluster IV . Summary of Item Content and Indices for Cluster V . . . . . . . Summary of Item Content and Indices for Cluster VI Summary of Item Content and Indices for Cluster VII . . . . . . Summary of Item Content and Indices for Cluster VIII . . . Summary of Item Content and Indices for Cluster IX . Summary of Item Content and Indices for Cluster X Summary of Item Content and Indices for Cluster XI . . . . . . . . Summary of Item Content and Indices for Cluster XII Summary of Item Content and Indices for Cluster XIII in of of of of of of of of of of of of of Each of the Association Association Association Association Association Association Association Association Association Association Association Association Association Page 33 35 37 41 42 44 45 47 50 51 53 54 55 57 Summary of Item Content and Indices of Association for Cluster XIV Comparative Summary of Cluster Associations in the Hierarchical Method and Comprehensive Hierarchical Approach . . . Summary of Cluster Labels and Indices of Association for Second Order Cluster A . Summary of Cluster Labels and Indices of Association for Second Order Cluster B . . . Cluster Rankings Regarding Their Dimensional Repre- sentation of the M-Scales Summary of Cluster Labels and Interpretive Emphases Summary of Labels and Interpretive Emphases of Second Order Clusters vi Page 60 63 66 66 67 74 78 CHAPTER I THE PROBLEM Purpose_of_the.Study Although the concept of academic motivation has attracted con- siderable interest and concern from educators it continues to be an inadequately defined and complex determinant of scholastic achievement. This is true, partially, because of the paucity of research that con- tributes to the understanding of the relationship of motivation to learning.1 Another reason has been the lack of comprehensiveness and objectivity in studies attempting to identify the dimensions which con- stitute academic motivation. It is the purpose of this study to isolate and examine the mean- ingful dimensions found in an objective, validated measure of academic motivation which has been constructed to incorporate a variety of moti- vation related variables. It is proposed that this investigation will contribute to further description of the dimensions underlying the motivation complex, to increased theory integration and ultimately to a better understanding of academic motivation. 1Raymond B. Fox, "Dearth of Research on Motivation," The Clear— ing House, XXXV, No. 4 (1960), 239-43. Need. for the- .Study The fact that academic motivation is deemed to be important in determining academic success, yet is so little understood, is evidence of the need for this type of study. Research findings have also indicated a need for further study on motivation. For example, Weiss, Wertheimer, and Groesbeck demonstrated that certain measures of motivation increase the prediction of scholastic performance when combined with an aptitude measure.1 This and similar evidence suggests that non-intellective variables influence academic achievement and warrant investigation to identify their dimensional characteristics. However, research contributions in this area have been limited by their restriction to comparisons of performance on various personality measures.2"3’4 Although these studies show that a relation- ship exists between certain non-intellective variables and academic achievement, they lack meaningfulness and conclusiveness since the 1Peter Weiss, Michael Wkrtheimer, and Byron Groesbeck, "Achieve- ment Motivation, Academic Aptitude, and College Grades," Educational and Psychological Measurement, XIX, No. 4 (1959), 663-66. 2H.H. Morgan, "A Psychometric Comparison of Achieving and Non- achieving College Students of High Ability," Journal of Consulting Psy- chology, XVI (1952), 292-98. 3James V. Mitchell, Jr., "Goal-Setting Behavior as a Function of Self-Acceptance, Over- and Under-achievement, and Related Person- ality Variables," The Journal of Educational Psychology, L, No. 3 (June, 1959), 93-104. 4P.J. Bond, "The Relationship Between Selected Nonintellective Factors and 'Concealed Failure' Among College Students of Superior Scholastic Ability," Unpublished doctoral dissertation, Purdue Univer- sity, 1960, cited in "Academic Underachievement in College," Student Personnel Methods Bulletin, Claremont, California: College Student Personnel Institute (August), 1964. nature of the relationship can only be Speculated. They are also characterized by fragmentation since each deals with only one aspect of motivation and by a lack of objectivity because the instruments used were not validated as measures of academic motivation. The result has been limited progress toward the identification and understanding of the constructs comprising academic motivation. An investigation based on an objective measure composed of validated items meets a need for this type of approach. Unfortunately the factors identified in past studies as having a relationship to academic over-and underachievement have not been integrated into an underlying theory which accounts for the major con- structs comprising academic motivation. The fact that theory compre- hensiveness is yet unattained was illustrated in the Farquhar study where it was concluded that McClelland's three factor theory was "too "1 simple to explain the academic motivation complex. The limited pre- 2 and dictability of personality measures as cited by Stone and Foster, Worell3 is further evidence that a more adequate theory is needed for improvement of future measure development. An analysis of the underly- ing psychological traits in an objective measure of academic motivation 1William W. Farquhar, "An Integrated Research Attack‘on Academic Idotivation," Journal of Counseling Psychology, IX (1962), 84-68. 2LeRoy A. Stone and James M. Foster, "Academic Achievement as a ,Function of Psychological Needs," The Personnel and Guidance Journal, XLIII (1964), 52-56. 3Leonard Worell, "Level of ASpiration and Academic Success," TTKB Journal of Educational Psychology, L, No. 2 (April, 1959), 47-54. which has predictive validity for over-and underachievement may sub- stantially contribute to more adequate theory building. Moreover, increased understanding of academic motivation has implications for the total educational process, eSpecially teaching, counseling, placement and curriculum development. A theoretical under- standing of academic motivation hepefully would enable teachers and counselors to help underachievers perform comensurate with their ability. Furthermore, prediction of underachievement may help prevent waste of academic talent. Such prevention might be at least partially achieved by placing predicted underachievers in an educational milieu that would be conducive to performance. This in turn has implications for curriculum development since the milieu would undoubtedly involve curricular considerations. Statement of the Problem The problem in this study is to isolate and describe the under- lying dimensions of an objective measure of academic motivation of eleventh grade male students. This investigation was conducted in con- junction with a larger research project Sponsored by the United States ()ffice of Education, under the direction of William W. FarquharJ' The Farquhar project was concerned with the develOpment of an objective battery of tests which measure various aSpects of motivation for academic agliievement. The resulting battery, The Michigan M-Scales, hereafter 1William W. Farquhar, Motivation Factors Related to Academic Achtievement. Final Report of U.S. Government Cooperative Research Pro- ject: No. 846, Michigan State University: Office of Research and Publi- catinans, College of Education, January, 1963. ' referred to as the M-Scales, are used in this investigation. The pro- blem of this study is to conduct an Elementary Linkage Analysis of the 139 male items in the total battery of the M-Scales to (1) determine what items group together into clusters on the basis of the agreement reSponses to the items; (2) examine the item content and determine the underlying characteristics held in common by the items which account for their being grouped together; (3) label the clusters in meaningful terms; (4) suggest implications for further theory building.1 The problem in essence then is to determine the number, nature, magnitude and structure of the dimensions functioning in the £2521 male battery of M-Scales. The theory on which the M-Scales was based follows. Theory Farquhar considered behavior theory as having three overlapping levels, two of which actually served as a basis for the development of the M-Scales. At the first level, focusing, an attempt was made to identify significant variables for the testing of null hypotheses con- cerning their relevance. Two scales were developed at this level: The 1The total battery of the M-Scales is composed of four sub- SCéiles: The Generalized Situational Choice Inventory (G.S.C.I.), Word Rat:ing List (W.R.L.), Human Trait Inventory (H.T.I.), and the Preferred Jot) Characteristics Scale (P.J.C.S.). Sample copies may be obtained erJUIDr. William W. Farquhar, Michigan State University. Factor analyses have been made on the individual sub-scales. It is the purpose of this study to analyze the battery in its entirety. Word Rating List which was designed to measure academic self-concept,1 and the Human Trait Inventory which was concerned with the measurement 2 of personality traits. At the second level, predictive, directional hypotheses were established on the basis of past theorizing. McClelland's N-achievement factors were altered and polarized to allow prediction of extremes in academic motivation, i.e., need for (a) long-term vs. short-term involve- ment; (b) unique vs. common accomplishement; (c) competing with a maximal vs. a minimal standard of excellence.3 Two scales were developed at level two: The Generalized Situational Choice Inventory which gathered 4 a person's preference for certain motivation related situations, and The Preferred Job Characteristics Scale which measured aspirations and values concerning certain future jobs. Theory was not developed at level three, integrative, where an attempt would be made to relate the constructs of the sub-scales. Although it is hOped that within the present study a greater degree of 1David A. Payne, A Dimension Analysis of the Academic Self-Con- cepts of Eleventh Grade Under- and Over-AchievinggStudents. Ph.D. Dissertation, Michigan State University, 1961, and Farquhar, Motivation Factors . . . . 2Ronald Taylor, "Personality Traits and Discrepant Achievement," Journal of Counseling Psychology, XI (Spring, 1964), 76-82, and Farquhar, Motivation Factors . . . . 3David C. McClelland, John W. Atkinson, Russell A. Clark and Edgar L. Lowell, The Achievement Motive. New York: Appleton-Century- Croft, Inc., 1953, pp. 110213. 4Marion Dennis Thorpe, The Factored Dimensidns of an Objective Inventory of Academic Motivation Based on Eleventh Grade Male Over- and Ilgderachievers. Ph.D. Dissertation, Michigan State University, 1961, and Farquhar, Motivation Factors . . . . comprehensiveness will be achieved because many motivation related variables are taken into account, theory building will remain at level two, predictive, as in the Farquhar study. The Hypothesis Several assumptions underlie this study. Those initially made by the Farquhar team in the construction of the M-Scales that apply here are: that over-and underachievement represent extremes in motivation (See Chapter III for a statistical definition of under-and overachieve- ment); that the test items embody the related theory deSpite the possible bias of the test constructors; that the M-Scales actually measure academic motivation.1 Validity estimates from the Farquhar study give some credibility to the last assumption.2 An additional assumption to be made in this investigation is that the Elementary Linkage Analysis procedures to be used will accurately identify the underlying dimensions of the M-Scales so that they can be meaningfully described. It is hypothesized that the items can be arranged in clusters which can be meaningfully described in relation to academic motivation. Statement of the Hypothesis Psychologically meaningful and descriptive dimensions which re- late to academic motivation can be isolated and identified. Farquhar, Motivation Factors . . . , p. 181. 21bid., p. 176. .Organization of the Study The general plan of this study is as follows. In Chapter II a review of the related research is presented. A discussion of the design and methodology of the study including the sample selection and analytical procedures used in identifying the dimensions of the total male battery of M-Scales is presented in Chapter III. The analysis of the data including the discussion and interpretation of the cluster dimensions constitutes Chapter IV. The summary, conclusions and impli- cations for further research appear in Chapter V. CHAPTER II REVIEW OF LITERATURE A paucity of research designed to identify the dimensions of academic motivation via analyses of objective measures exists in the literature. The few studies relevant to the present investigation will be reviewed and evaluated. A summary of the contribution of these studies to the area of academic motivation will provide the frame of reference in which a further description of this investigation can be understood. Motivation as a Factor in Academic Achievement Weiss has demonstrated the existence of motivation as a factor influential in determining academic performance.1 He conducted a fac- tor analysis on the reSponses of a sample of 180 college students to ten questions concerning their expected performance on a verbal reasoning test. Two factors were identified, a judgmental or expectational factor and a motivational or aSpirational factor. Weiss concluded that the judgmental factor represents a realistic appraisal of one's ability 1Robert Frank Weiss, "ASpirations and Expectations: A Dimen- Sixanal Analysis," The Journal of Social Psychology, LIII (April, 1961), 249-54. 10 (non-emotional) whereas the motivational factor represents emotional aSpiration which he suggests is motivation and produces drive in learn- ing situations. Although Weiss' findings add evidence to the belief that non- intellective factors influence academic achievement they do not provide definite information regarding the motivation complex. Weiss' defini- tion of motivation is limited to the idea of asPiration which is a com- plex, undefined concept itself and does not incorporate all the vari- ables inherent in academic motivation.1 A more satisfactory and com- prehensive approach would be to identify the dimensions of a validated objective measure of academic motivation which has been constructed to incorporate a variety of non-intellective variables related to academic achievement. Personality Traits Related to Academic Motivation Middleton and Guthrie conducted a study which is relevant in a general way to the present investigation.2 They attempted to identify personality syndromes that relate to academic achievement. A 300 item personality questionnaire designed to measure 18 of the needs in Murray's system was administered to a sample of 14 high achieving and 14 low achieving students. All subjects were upperclass male business 1Leonard Worell, "Level of Aspiration and Academic Success," The Journal of Educational Psychology,L,No.2 (April, 1959), 47-54. 2George Middleton, Jr. and George M. Guthrie, "Personality Syndromes and Academic Achievement," The Journal of Educational Psy- chology,L',No.2 (April, 1959) , 66-69. 11 management students at Pennsylvania State University. -Low achievers were defined as those under a 2.0 G.P.A: and achievers as those over 2.5. Adjusted phi coefficients were computed between each pair of subjects and the 14 X L4nmtricesf6r high and low achievers were subjected to a tranSposed factor analysis. The rotated factor loadings were correlated with the scores on each of Murray's l8 scales. Five factors were extracted for the high achievers: 1. Power and Approval; II. Resentment and Independence; III. Dependence; IV. Social Prestige and Influence; V. Hostile Agressive Denial of Tender Socialized Feelings.. Four factors were extracted from the matrix of low achievers: I.-Preoccupied with.Pleasure; II. Extroversion; III. Disavowing Social Shortcomings; IV. Power and Acceptance. The findings of this study are limited in several ways. The sample used was too small and homogeneous to allow generalization of the results. The difference between 2.00 and 2.50 can hardly distinguish between low and high achieving students. To be sure, the subjects used could not be classified as over-and underachievers. As in many other studies of this type a causal relationship between the syndromes identi- fied and academic achievement can only be inferred since the validity' of the measure used is undetermined. The syndromes describe high and low achievers but do not define or describe academic motivation. The fact that high and low achievers were found to share some of the same needs further indicates that the results are limited in their conclu- siveness. In fact, the authors suggest that the ". . . factors must be regarded as more illustrative than confirmed." 12 AlthOUgh the study is limited to Murray's need system, the find- ings do indicate that some independent personality factors have some relationship to academic achievement. McQuary1 conducted a study in which he attempted to identify the factor pattern underlying twenty-three variables assumed to be related to scholastic achievement. The variables chosen were of both the in- tellectual (such as aptitude and achievement test scores, and grades) and non-intellectual (such as home background, and extracurricular: participation) variety. .Data pertaining to all of the variables was obtained from the University of Wisconsin Student Counseling Center on male freshmen in the 1948-49 and 1949-50 school years who had used the services of the center. A subject's inclusion in the sample was deter- mined by whether complete data on all 23 variables was available-on him. A sample of 174 resulted. .McQuary hypothesized that: "(1) Achievement.in college is significantly related to certain non-intellectual variables; and (2) The non-intellectual variables can be grouped into several factors.” 'Scores on the variables were normalized by converting them to stanine scores and were then intercorrelated. The multiple group method of factoring. was used and the resulting factor matrix was turned into an orthogonal solution (F). Single-plane and radial rotations were made on columns of the F matrix until an oblique solution (V) was obtained. 1John P. McQuary, "Some Relationshipszetween Non-Intellectual Characteristics and Achievement," The Journal of Educationalrsychology, XLIV, No. 4 (April, 1953), 215-28. 13 The analysis yielded six factors which were labeled: I. Academic Performance Predictor: II. Social-class Intelligence; III. Participating Urban Scholar; IV. Social Extroversion; V. Academic and Financial Secur- ity; and VI. Introvertive Reader. Two of the factors accounted for most of the correlation of grades with the various variables. The two in point were: first, the Academic Performance Predictor factor which grouped together ACE-Q and ACE-L scores, the Speed of Reading score, the Level of Comprehension and Vocabulary tests and the high school per- centile rank, credits carried and grade points earned; and second, the Social Class Intelligence factor which grouped the ACE-Q and L scores plus the educational level of the father and mother and the foreign- born parent variables. McQuary's study is a comprehensive analysis of performance vari- ables assumed to be related to academic achievement.1 Since McQuary dealt with externally observed performance variables in contrast to variables theoretically built into test items and measured by subject reSponse to these items as in this investigation, the two studies have limited comparability. Although McQuary's study resulted in some interesting groupings of the variables used it is limited in several ways. His assumption that the variables used are related to scholastic achievement is unsub- stantiated in certain cases. Perhaps other-variables would have been equally or more apprOpriate. In addition, the sample used was restricted 1Ibid. 14 by the data gathering procedures. To be included in the sample a stu- dent had to have used the Counseling Center and have complete data in his file concerning all 23 variables. This eliminated many students and produced a biased sample- Another limitation is the heterogeneity of the variables used. Labeling the.common elements of a group of variables became a difficult task resulting in factor labels with vague meaning. For example, "Participating Urban Scholar" is a broad label that covers the variables in the factor but has little descriptive mean- ing. Since the study was intended to measure non-intellective vari- ables related to academic achievement its major limitation is that the only factor with a substantial correlation with G.P.A. contains all intellective variables such as aptitude test scores, credits earned, and high school percentile rank. The study does not contribute to a more definitive understanding of academic motivation. Dimensions of Measures of Academic Motivation Mitchell conducted a factor analytic study on a combination of various purported measures of achievement motivation.1 His purpose was to investigate the question of whether achievement motivation was a single, invariable construct or whether it included a multiplicity of independent psychological dimensions. In the first phase of the study data was gathered on seven measures of achievement motivation and the A.C.E. all administered to a sample of 131 female college students 1James V. Mitchell, Jr., "An Analysis of the Factorial Dimen- sions of the Achievement Motivation Construct," The Journal of Educa- tional Psychology, LII, No. 4 (August, 1961), 179-87. 15 enrolled in his elementary educational psychology class at the Univer- sity of Texas. These measures included the McClelland Test of Achieve- ment Motivation; a sentence completion test, an adjective checklist containing achievement related items, an inventory-type test of achieve— ment motivation with items taken from the MMPI, and a multiple choice questionnaire constructed by the investigator; the Taylor Manifest Anxiety Scale; and two level of aspiration measures. Scores on each measure were correlated with one another and with a criterion of achieve- ment motivation, namely the discrepancy between G.P.A. and ACE scores. In the second phase, the individual items (29 in number) of the achievement motivation measures were subjected to a centroid factor analysis and rotated to orthogonal simple structure by means of Kaiser's varimax technique. The six factors identified were labeled as follows: Academic Motivation and Efficiency (Mitchell says this is the only fac- tor which is highly predictive of actual academic performance); Wish- Fulfillment Motivation; Non-academic Achievement Motivation; Self-Satis- faction; External Pressure To Achieve; and Imputed Generalized Motiva- tion Without Attendant Effort. Mitchell concludes that achievement motivation is not only multi- dimensional and complex, but that tests and test items vary widely in their ability to measure achievement motivation. Although Mitchell's findings convincingly demonstrate that moti- vation is complex and multidimentional, they add little definitive infor- mation useful for further theory development. The study itself is not based on any theory pertaining to the substance of academic motivation thus the results have limited meaning. From an examination of the factor loadings, only factor I seems to have much significance,but it is too general to 16 be meaningful. The results are also limited by the low predictive val- idity of the measures of achievement motivation used. The factors identified have limited significance because of the lack of validity of the items upon which they are based. Again, little can be generalized from the results since the sample consisted of only 131 subjects of the same sex, in the same class and school. Mitchell's study does point out the need for further research in this area and underscores the need for such investigations to be based on objective measures that have a theoretical base and are validated on an adequate sample of under-and over-achieving subjects. In another factor analytic study, Michael, Jones and Trembly analyzed a previously validated 34 item measure of academic motivation called the U.S.C. Inventory of Study Methods and Attitudes combined with a measure of verbal aptitude and a highly reliable achievement measure.1 The two samples used consisted of 236 men and 131 women enrolled at the University of Southern California. From the principle axis solutions, orthogonal rotations were effected analytically by Kaiser's Varimax Method. The factors identified for the females were: (1) general achieve- ment drive; (2) freedom from neurotic orientation to study tasks; absence of neurotic behavior patterns; self-confidence; (3) basic study skills; (4) self-reliance and self-sufficiency; (5) collaboration (giving and receiving help); dependence - sociability orientation; succorance- 1William B. Michael, R.A. Jones and W.A. Trembly, ”The Factored Dimensions of a Measure of Motivation for College Students," Educational and Psychological Measurement, XIX (1959), 667-71. 17 nurturance; (6) verbal aptitude; (7) cOnformdty to teacher expectation; (8) compulsive persistence; (9) systematic, methodical application; (10) efficiency in planning and studying; (11) positive affect intrinsic in the educational experience; (12) organizational perserverence. Male factors identified were female factors one through seven plus the fol- lowing: (l) shyness, seclusiveness (with negativistic and withdrawal tendencies); (2) aversion to work (laziness); (3) procrastination accom- panied by rationalization. Several limitations of this study prevent it from significantly clarifying the motivation complex. The authors do not indicate the magnitude of the validity of the items in the mmtivation measure nor do they state the criterion used in validation. It appears that validation was not based upon under- and over-“achieV'inthudent-s. In- addition, the aptitude and achievement measures are not identified and no coefficient of reliability is given for the latter which is said to be "highly reliable." The samples further-limit the conclusiveness of the study since they were confined to one class at one University. An examination of the factors identified adds further doubt to the validity of the U.S.C. Inventory of Study Methods and Attitudes as a motivation measure. .Although all the factors appear to be related to academic achievement, few of them suggest a motivational relatipnship. They more appropriately could be called a combination of personality traits, study methods and academic ability. The Michael Study again illustrates the need for dimensional studies to utilize valid measdres of academic motivation as well as. reliable criterion and an adequate sample. 18 Summary The research literature contains a limited number of studies deal- ing with the dimensions of objective measures of academic motivation. Generally the findings indicate that motivation is a complex multidimen- sional phenomenon. Several situational and personality factors such as independence, conformity, goal orientation, etc. have been shown to have a relationship to academic motivation. The nature of these relationships is left undefined. The studies cited have limited meaningfulness since they lack a theoretical base from which the results can be interpreted. In addition, most of the studies lack well defined criteria, sufficiently large samples and validated measures of motivation. A more satisfactory approach would be to objectively define over-and underachievement by statistical means, develOp an instrument to measure academic motivation validated on over- and under-achieving subjects as defined and then analyze the items in the instrument to identify the meaningful dimensions contained therein. Previous research lacks comprehensiveness since each study deals with a fragment of the motiVation complex. The paucity and fragmentation of past research underscores the need for additional research which integrates the various elements related to motivation into a meaningful theoretical framework. The present study is designed to overcome: a) weakness in instrumentation by using a device which evidences validity and reliability, b) generalization from small samples by employ- ing reSponses of 240 students, and c) inadequacies in number of assess- ment variables by using all 139 items on the M-Scales. CHAPTER III DESIGN AND METHODOLOGY Within this chapter may be found: a) a brief review of the Farquhar sampling and instrument construction technique; b) a state- ment of the hypothesis of this study; and c) a description of the analysis procedures employed. Review of the Farquhar Study Quotations from publications describing the sampling and instru- ment construction techniques in the Farquhar study are cited as back- ground to this investigation. Sampling Selection of under- and over-achieving students was ac- complished by the use of the Two Stage Regression Model de- scribed by Farquhar. The p0pu1ation of the larger project consisted of 4,200 eleventh grade students from nine high schools in eight Michigan cities. On an‘g priori basis, schools were selected to represent the total range of socio- economic environments. Individuals who varied more than 1 S.E.est. from the first to the second administration of two different aptitude measures were eliminated from the study to control for stability unreliability. Over-achievers were defined as falling at or above 1 S.E.est. relative to the linear regression of aptitude (Differential Aptitude Test, Verbal Reasoning sub-scale) on achievement (cumulative grade point average for academic subjects for 9th and 10th grades). Conversely, under-achievers were defined as falling l S.E.est. below the regression line. Regression equations were developed separately for each sex for each of the nine schools because of the lack of comparability in grading milieu. l9 20 Reliability of the achievement criterion (high school grades) was estimated to be {0.75 for males and {0.80 for females. The total number of under- and over-achievers for each sex was randomly dichotomized to provide validation and cross-validation groups. Validation samples contained 95 male over- and 62 male under-achievers; and 95 female over- and 90 female under-achievers. Cross-validation samples contained 80 male over- and 69 male under-achievers; and 91 female over- and 87 female under-achievers. Equivalent number of individuals are not found in each category because of sample loss through poor test motivation, inability to follow directions, absenteeism, and attrition.l The analysis in this study was limited to a sample of 240 males selected as part of the Farquhar study. Instrument Construction Techniques Analysis Procedures.--The 2 x 2 chi square contingency table model was used for all item discrimination and questionnaire reaponse analyses (over- and under-achiever's responses to either of the two alternatives). Those scales employing a four point rating scale (never, sometimes, usually, and al- ways) were arbitrarily dichotomized between sometimes and usually. The level of significance for the chi square analysis was set at .20 for validation and .10 for cross- validation. Multiple regression analyses were conducted to determine the effect of adding the motivation measures to aptitude in predicting grade point average. Finally, the valid sub- scales were factor analyzed. The purpose of factoring was to l) empirically determine the homogeneity of the scale and 2) provide a psychologically meaningful interpretation of the item content groupings for retheorizing about the nature of academic motivation. Reliability of the M-Scales.--Using a sample of 240, a Hoyt's analysis of variance reliability estimate of 194 for 139 male cross-validated items was obtained. A female sample of 240 yielded a comparable .93 reliability estimate on 136 cross-validated items. For the most part, the reliability estimates for the sub-scales and various defined groups of 1William W. Farquhar and David A. Payne, ”Factors in the Aca- demic-Occupational Motivations of Eleventh Grade Under- and Over- Achievers," Personnel and Guidance Journal, XLII, No. 3 (Nov., 1963), 246-47. 21 achievers attain a satisfactory level (r = .68 to .92 for males and .60 to .93 for females). Validity of the M-Scales.--Based upon a sample of 254 males and 261 females the validity estimates of the total M-Scales against grades was .56 and .40 reSpectively. The cross- validation estimates were .49 and .48 for males and females. The correlation of the sub-scales with the grade point criterion (GPA) follows the same pattern with the female correlations lower in magnitude than the males. The range was .27 to .42 for females and .32 to .51 for males. The sub-scales correlate positively with each other indicating some measurement overlap. The range was from .32 to .51 for males and from .32 to .53 for females. The sub-scales findings for both sexes are comparable. Multiple Correlation Estimation.--The multiple correlation based on the total M-Scales and DAT-VR estimates of grade point average was .69 and .73 for the male and .64 and .63 for the female validation and cross-validation samples.1 The total M-Scales battery contains 139 male items. Factor analyses have been conducted on the separate sub-scales.2 In the present study an attempt is made to analyze the total male battery in order to identify factors or traits that relate to academic motiva- 5.2922221; Some of the assumptions underlying this study which were extended from the Farquhar investigation were cited in Chapter I. A further 1Farquhar, Motivation Factors . . . , pp. 173-77. 2Farquhar has conducted a factor analysis of each sub-scale. Results of Farquhar's analysis is presented in Motivation Factors Relgged to Academic Achievement, p. 178. 22 assumption forced upon the data by the analysis technique ‘was that all reSponses to the items have the same measurement indicant for all sub- jects in this study. It was assumed that when two or more subjects responded to a given item in the same way that they all meant the same thing.1 In addition, it was assumed that the analysis technique grouped the items according to their relationship with one another. From this, it was hypothesized that psychologically meaningful and descriptive clusters which relate to academic motivation could be isolated and identified. Statement of the Hypothesis Psychologically meaningful and descriptive clusters which relate to academic motivation can be isolated and identi- fied. Elementary Linkage Analysis The reSponses of the 240 male students to the 139 male items in the M-Scales were subjected to McQuitty's Elementary Linkage Analysis (ELA) procedure as programmed for the Control Data 3600 computer at Michigan State University.2 The data was in the form of "l” and "O" reSponses with ”1” being in the direction of high motivation. ELA is a cluster method and was selected for the analysis on the basis of its simplicity and objectivity. It accomplishes the same basic task as factor analysis. McQuitty states that although ELA is applied to 1Louis L. McQuitty, "Elementary Factor Analysis," Psychological Reports, IX (1961), 76. 2Louis L. McQuitty, "Elementary Linkage Analysis for Isolating Orthogonal and Oblique Types and Typal Relevancies," Educational and Psychological Measurement, XVII, No. 2 (Summer, 1957), 207-29. 23 a theory of psychological structure which is basically different from Thurstonian simple structure it ". . .sometimes yields results very similar to, or identical with, those of rotated factor-analytic solu- tions."1 In fact, when correlation coefficients are used as the indices of association, results from ELA can be rotated to factor analysis simple structure.2 As a cluster method ELA overcomes the weakness of similar methods of relying upon the experimenter to arbitrarily fix the lower limit of correlation coefficient acceptable for inclusion into a cluster. Since ELA requires that to be included in a cluster a variable must have a higher index of association with some other variable in the cluster than with any other variable in the matrix, the criteria for inclusion is set by the data exclusively. McQuitty uses the terms "type" and "typal structure" in describ- ing his technique. He defines a type as: . . .a category of persons (or items) of such a nature that everyone in the category is in some way more like some other person in the category than he is like anyone not in the category. In terms of coefficients of correlations be- tween persons, every person in a type would have a higher correlation with some other person in the type than he would with anyone not in the type. The psychological theory underlying typal structure as conceived by McQuitty applies to clusters of people. Although the procedure is, equally as applicable for clustering items (as was done in this study) interpretation of items cannot be based on typal theory. As mentioned 1Ibid., p. 207. 2 . Ibid., p. 212. 3McQuitty, "Elementary Linkage Analysis . . . ," 213. 24 above, grouping items or tests forces the assumption that all responses have the same measurement indicant for all subjects and therefore only an approximate solution can be achieved. By clustering peOple this difficulty is avoided and a pure solution is possible.1 .McQuitty sug- gests, therefore, that groups of items be called clusters rather than types.2 Because this possible limitation is not unique to ELA but applies to all factor analytical studies in which items or tests are factored and because the interest in this study and the data available pertained to items rather than people the apparent risk in preciseness seemed to be warranted. To conduct a typal study of the people involved in this study, more information on the subjects would have been neces- sary. Such an endeavor could legitimately become a separate research study. A minor potential limitation of ELA stems from the nature of the data used. The analysis used agreement scores, i.e., all 240 subjects were compared with one another with respect to their reSponses to each of the 139 items. The data, therefore, was unordered in the sense that 'the subjects were not ranked on a continuum of scores. The potential limitation was that of unreliability of the reSponses since reSponses to individual items are known to be less reliable when considered in isola- tion than when compared on the basis of their standing in a rank or profile.3 However, McQuitty states that: 1MCQUittY: "Elementary Factor Analysis . . . ," 76. 2Personal conversation with Dr. Louis L. McQuitty, June 25, 1965, Michigan State University, East Lansing, Michigan. 3MCQUittY: "Elementary Linkage Analysis . . . ," 14. 25 . . .even though reSponses to individual items may be in general unreliable, they may nevertheless have "differen- tial reliabilities' across people, i.e., a particular reSponse though generally unreliable may be highly reli- able for the members of a particular type of persons; all members of this type may give this reSponse invariably, even though other individuals fluctuate with reSpect to it. This limitation was,therefore, not considered to be serious. Another limitation was that the reliability of the clusters identified by the Elementary Linkage Analysis procedure has not been established. However, on the basis of related studies conducted by McQuitty, he suggests that ELA, when applied to matrices of more than 100 subjects will yield reSpectable reliabilities and validities.2 Analysis Procedures ELA procedures are relatively straight forward. Using the agree- ment scores for each item (which is simply the number of subjects who reSponded in the same fashion to that item) as the index of association, the first step is the identification of the highest reciprocal pair, i.e., the two items in the entire matrix which have the highest agreement of subject reSponse. These two items form the core of the first cluster. The next step is to identify additional items which have their highest index of association with either of the two reciprocals and include them in the cluster. Then the closest associates are identified for these items, and so on until all items are included in the cluster that have a lIbid. 21bid., p. 223. 26 higher association with some other item in the cluster than any item not in the cluster. After cluster one has been completed the next highest reciprocal pair is identified as well as the highest associates in order to form cluster two. This process is repeated until all items in the matrix have been assigned to a cluster. Reference Items When the constituent items of each cluster have been identified it is then possible to identify the items which are most representative of their respective clusters. The representative items, referred to as "reference items," can be identified by twoprocedures.1 The first method is accomplished by forming a sub-matrix of the indices of associa- ' tion (agreement scores) between the items for each cluster. Within each cluster the indices of association are summed in the column of each item so that the reference item can be identified by finding the item : with the highest sum of associations. .The second, third, fourth, etc. most representative items can also be identified in the same way. By definition the reference item is the item‘with the highest association of other items to it within its own cluster. McQuitty has demonstrated that when correlation coefficients are used as indices of association instead of agreement scores the loading of the reference item obtained by the procedure described above compares favorably with the first factor solution.2 The secondprocedureAinv61ves computing the average associa- tion each item in a clusterjhas with both members of the reciprocal 1McQuitty refers to reference items as "reference factors." "Elementary Factor Analysis," p. 74. 21bid. 27 pair.1 Identification begins with the third highest reference item be- cause the two members of the reciprocal pair are the first and second reference items. The latter procedure is preferable since it gives a more accurate indication of which items are most closely associated with the core of the cluster. In this investigation computations were made according to both procedures and the results compared. The comparison is discussed in Chapter IV. In deciding which items would be designated as reference items it was arbitrarily decided that not more than five reference items would be selected in each cluster. Only the reciprocals were selected in the smaller clusters. In the smallerclusters, those items whose indices of association were of a similar magnitude were selected. An important consideration applicable to both methods is that as purely statistical procedures they may not always concur with sub- jective or clinical judgment in identifying reference items. The prox- imity of the statistical and subjective judgments was noted in this study and is discussed in Chapter IV. When the clusters and their respective reference items have been identified it is then possible to describe a prototype for the items in each cluster by examining the content of all the items, eSpecially the reference items. Theoretically, the elements or characteristics they hold in common could then be meaningfully described which was the inten- tion of the study. It has been hypothesized that an examination of the clusters will yield psychologically meaningful descriptions of traits having some relationship to academic motivation. 1This procedure was suggested by Dr. Louis L. McQuitty in a per- sonal conversation July 8, 1965 at Michigan State University, East Lansing, Michigan. 28 Elaborations of Elementary Linkage Analysis Once the clusters are identified, two elaborations of ELA make it possible to investigate the relationship between the clusters. The first is called the Hierarchical Method.1 The reference items from the clusters (one from each cluster having the highest column sum) are entered into a sub-matrix using the indices of association between them which were ob- tained from the original matrix. These items then are grouped into second-order clusters by the same procedure that the first-order clusters were formed. If more than one second-order cluster results the reference items of these clusters can be entered into a sub-matrix to identify third- order clusters. This procedure can be continued until one cluster remains 'which, theoretically, best represents the entire matrix. A second method, the_Comprehensive_§ierarchical Approach provides additional comprehensiveness and refinement.2 Entire clusters can be compared with one another by pairing every cluster with every other cluster. The items of one cluster are entered into a sub-matrix along the tOp and the items from the other cluster along the side. The index of association between two clusters becomes the mean of all the indices in their sub-matrix. -As an alternative-to this extensive procedure one can, instead of using all the items, select some limited number of items that appear to be most representative of their clusters. This limited approach.was used in this study because it was assumed that the most 1Louis L. McQuitty, "Capabilities and Improvements of Linkage Analysis as a Clustering Method," Educgtional and Psychological Measure- ment, XXIV, No. 3 (Fall, 1964), 446. 21bid., p. 447. 29 representative items would provide more meaningful results. The analysis yielded some implications regarding the tenability of this assumption which are discussed in Chapter IV. Because it was of interest in this investigation to know the relationship between the reference items as well as the relationship between the clusters as a whole both the Hierarchical Method and the Com- prehensive Hierarchical Approach were employed in the data analysis. As a further extension of the analysis, the internal consistency of the clusters was determined by comparing the mean association of items within a cluster with their mean association with items in other clusters. The magnitude of a given cluster's internal consistency is expressed by the extent to which the mean of its within cluster item associations is greater than its association with any other cluster. The within cluster mean of item associations is computed by dividing the sum of the agree- ment scores by the number of associations in each cluster sub-matrix. The associations between entire clusters is determined by the procedures outlined in the Comprehensive Hierarchical Approach. It is possible for the items in one cluster to have a higher association with another clus- ter than they have among themselves. Such clusters have relatively low internal consistency. Because the computation of the inter cluster asso- ciations in this study (in the Comprehensive Hierarchical Approach) was based on a few selected "most representative" (those that have the highest agreement with the reciprocal pair and with one another) items 1This procedure was suggested-by Dr. Louis L. McQuitty in a per- sonal conversation July 8, 1965 at‘Michigan State University, East Lansing, Michigan. ‘ i 30 from each cluster, the criteria for internal consistency was high. If the ”most representative" items in a cluster were more consistent (had a higher mean index of association) with another cluster than among themselves it was unlikely that all the items together would be more consistent (i.e., exceed the inter cluster association). Knowledge of cluster internal consistency was of particular im- W§portance in interpreting clusters identified by the linkage analysis. .By definition, an item must have its closest association with only one other item to be included in a cluster. The association with the other litems in the cluster is through their "linkage." .It is possible 2 although not probable for two items having a distant "linkage" in a cluster to be unalike in content. Interpretation, therefore, had to take into account not only which items were related but how closely they *were related. Summary Data gathered in the Farquhar project relative to the responses of 240 male eleventh grade high school subjects to the 139 male items in the M-Scales was used in this investigation. Factor analyses have 'been conducted on the separate sub-scales, therefore, it was the goal of this project to analyze the M-Scales in their entirety. It was hypothe- sized that psychologically meaningful and descriptive clusters could be isolated and identified. The Elementary Linkage Analysis procedure was selected for data analyzation because of its simplicity and objectivity. A cluster as identified by ELA is defined as a structure in which every ‘item is more like some other item in that cluster than any item in any other cluster. The clusters were described in terms of the characteristics 31 relative to academic motivation held in common by their constituent items. Reference items which best represented each cluster were iden- tified adding to the interpretability of the clusters. Two elaborations of ELA, the Hierarchical Method and the 92m; prehensive Hierarchical Approach, each yielded two interpretable second- order clusters. In addition, information regarding the relationship between the reference items of each cluster as well as the relationship between the clusters was obtained. Using information from these two elaborations an indication of the internal consistency of the clusters was determined which facilitated more meaningful cluster interpretation. CHAPTER IV ANALYSIS OF RESULTS In this chapter the clusters resulting from the analysis will be interpreted and discussed. The descriptive labels given to the clusters represent a subjective interpretation of the item content. Judgments of this nature are susceptable to various interpretations depending upon the meaning the items convey to the particular researcher concerned. Therefore, the results of this study must be interpreted with the sub- jective element in mind. It is assumed that academic motivation is a continuum and that in reality most peOple fall somewhere in the middle zone between high and low motivation. However, the clusters are characterized as describing either high or low motivation with no attempt to describe the dimen- sions between these two extremes. The discussion of each cluster will include the following: 1. The number of items in the cluster. 2. Item distribution from the four sub-scales. 3. Reference items. 1 4. Internal consistency. 5. Cluster label. 1Data regarding the internal consistency of the clusters is pre- sented in Appendix A. 32. 33 6. Cluster description. 7. Comparison with results from factor analyses in the Farquhar project, The description of the individual clusters is followed by a separate discussion of the second order solutions resulting from the Comprehensive Hierarchical Approach. The hypothesis formulated as a basis for analysis was that psychologically meaningful and descriptive clusters which relate to academic motivation can be isolated and identified. The distribution of items from the sub-scales in each cluster is shown in Table 4.1. The analysis yielded a total of fourteen clusters. TABLE 4.1 NUMBER OF ITEMS FROM EACH SUB-SCALE IN EACH OF THE FOURTEEN CLUSTERS Cluster H.T.I. G.S.C.I. P.J.C.S. W.R.L. Total I 4 O 0 0 4 II 12 8 O 2 22 III 1 2 0 0 3 IV ,0 2 0 0 2 V O 5 O 0 5 VI 0 6 0 O 6 VII 6 5 12 4 27 VIII 1 1 7 O 9 IX 0 O O 3 3 X l O 0 12 13 XI 1 0 O 13 14 XII O O O 4 XIII O 15 1 19 XIV 1 0 , 0 7 8 Total 139 34 Clusters II, VII, and XIII had the greatest number of items, 22, 27, and 19 respectively, totaling almost half of the 139 male items in the en- tire battery. Items were drawn from more than one sub-scale in eight of the clusters. Clusters II, VII and.XIIIpu11ed items more frequently from all four sub-scales than any of the other clusters, thus achieving the highest degree of sub-scale integration. None of the other clusters contained more than two items from a second sub-scale. The Elementary Linkage Analysis (ELA) technique arranged the items into clusters indicating the linkage or relationship of the items to one another and to the reciprocal pair in each cluster. An inter- pretation and discussion of the clusters follows. A graphic presenta- tion showing the item associations and linkage patterns within each cluster may be found in Appendix B. Interpretation and Discussion of the Clusters Because of the nature in which the items were structured they pre- sented, when grouped into clusters, varying descriptions of motivation. Some of the clusters characterized low motivation, others high motivation and others polar traits, i.e., traits whose possession are associated with extremes in achievement. Because some traits have a positive in- fluence on achievement, their absence results in underachievement. Con- versely, other traits inhibit achievement while their absence facilitates it. Cluster I; Item content (four items) and indices of association for Cluster I are shown in Table 4.2. Items 12 and 4 formed the reciprocal pair and were designated as the reference items. Cluster I was found to be internally consistent when the association of the reciprocal items 35 TABLE 4.2 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER I. Index of Item Number and Content8 Associationu 4. It is difficult for me to keep interested in most of my school subjects. 181 9. I find it difficult to find the time to study my assignment for the next day. 164 12. Even when I do sit down to study I find that my mind tends to wander. 181 13. I have to be in the mood before I can study. 171 LEGEND (TABLES 4.2 through 4.15): aEach item has its highest agreement with the item under which it is indented. For example, in Cluster I, item 12 has its highest agree- ment with item 4, since they are the reciprocal pair. Item 9 asso- ciates highest with item 4, and item 13 is highest with item 12. Appendix B contains a graphic illustration of this arrangement in each cluster. bThe maximum agreement possible between two items is 240, i.e., the total number of subjects in the sample. was compared with the cluster's highest mean intercluster association, but not when the mean association of all the items was used. Thus, items 9 and 13 did not contribute substantially to the central meaning of the cluster because they had relatively low indices of association with the reciprocal pair. Cluster I was labeled Distractible because it characterizes a lowly motivated student who finds it difficult to concentrate on his studies and is easily diverted from academic pursuits. Item 4 indicates a lack of interest in school work and because item 9, which is associated with item 4, expresses a lack of time for studying, "interest" in non- academic matters appears to contribute to distraction from academics. 36 Item 12 had the highest column sum of associations and best ex- presses the idea of "distractability." Cluster I is similar to Factor I, "Agitation," of the Human Trait Inventory (HTI) reported in the Farquhar Report. Factor I which contained many of the same items as Cluster I was described as "low motivation related to excitation, distractibility, and escapism.“1 Cluster II: Item content (twenty-two items) and indices of association for Cluster II are presented in Table 4.3. The items clustered heaviest on reciprocal item number 17. The cluster integrated items from three sub-scales (Human Trait Inventory, Generalized Situational Choice Inventory, and Word Rating List). The selected representative items in Cluster II indicate internal consistency but not the total cluster of items. As in Cluster I, the implication is that some of the items con- tribute little to the meaning of the cluster. Cluster II describes a lowly motivated student and is labeled Irre3ponsible. The type of person described lacks an internalization of accepted behavior norms. He dislikes conforming to expectations set by others. At the same time, he desires significant achievement through non-conventional and non-academic means. Item 11 and its recipro- cal, item 17, both involve socially unacceptable and irresponsible be- havior, i.e., stealing and skipping school. Items 55, 62, 46, and 2,all associated with item 11, convey the idea of independence, so that taken together these items present a 1Farquhar, Motivation Factors . . . , p. 147. W _-—_-__———_____——__.= 37 TABLE 4.3 ITEM CONTENT.AND INDICES OF ASSOCIATION FOR CLUSTER II. 17. 51. 19. Index of Item Number and Content Association 11. When I was a youngster I stole things. 219 55. I would prefer to: (a) do what I think is right, or (b) do what others think is right. 205 62. I would prefer to: (a) carry out the plans of others, or (b) create something of my own. 197 46. I would prefer to: (a) work rapidly just "skimming along," or‘ (b) work slowly with great thoroughness. 190 2. I have been quite independent and free from family rule. 178 I have played hooky from school. 219 23. I feel that I haven't any goals or purpose in life. 211 20. I work under a great deal of tension. 199 10. I have done something that is considered ,dangerous just for the thrill of it. 176 27. I would prefer to: (a) avoid failing in school, or (b) do well in school. 209 14. I like to make the best grades possible. 196 22. I get disgusted with myself if I don't do as well as I should. 169 35. I would prefer to: (a) be successful in finishing a job, or (b) finish a job. 195 I would prefer to: (a) receive a grade on the basis of how much my teacher thinks I have learned, or (b) take a course from an instructor who only gives ”C‘s". 205 I want very much to be a success. 205 3. When I have an opinion, I stand up for it. 177 1. I worry about my grades. 171 38 Table 4.3 - Continued Index of IteunNumber and Content Association 38. I would prefer to: (a) make quick decisions and sometimes be right and sometimes wrong, or (b) deliberate over decisions and usually be right. 199 135. Teachers'feel that I am passive. 196 132. Teachers feel that I am inconsistent. 194 71. I would prefer to: (a) have one of my children win a beauty contest, or (b) have one of my children win a college scholarship. 195 5. I have difficulty working under strict rules and regulations. 192 39 picture of rejection of socially acceptable behavior and authoritarian control. Item 13 expresses a lack of goals and life purpose. Items associated with number 23 indicate concern and anxiety over achievement (20, 3, l, 22) along with the concept of competing with a maximum versus a minimum standard of excellence in school (27, 35, 51). At the same time, items 10, 19 and 71 indicate a desire for personal significance. The characterization given is one of ambivalence and in- consistency which is supported by the self—perception indicated by this type of student in items 135 and 132. In Cluster II is found an interesting integration of the con- cepts of irreSponsibility, non-conformity, rejection of authority, goal ambivalence, competition with maximum versus minimum standards, and desire for unique or significant personal experiences. These concepts are expressed in the five highest reference items, 17, ll, 23, 27 and 51 (in that order). It could be Speculated that the irreSponsible student who rebels against authoritarian expectations has few long-term goals, thus he has no maximum standards for which he competes. Instead he gains significance through immediate, non-academic behavior. This non-con- formity and ambivalence is accompanied by anxiety (item 2). Based on the column sum of associations of its representative items used in the "Comprehensive" computation, Cluster II is the third most representative cluster of the entire male battery of M-Scales. Although Cluster II does not corre3pond identically to any one factor cited in the Farquhar study, it does incorporate the concepts ex- 1 pressed in six of the Farquhar factors. The two items from the Word 11bid., pp. 138-57. 40 Rating List loaded on Farquhar's Factor II, "Educationally Resistive," which was interpreted by Farquhar to indicate that "the student sees himself through the teacher's eyes as resisting involvement with edu- cational tasks, either actively or passively." Three factors of the Human Trait Inventory included some of the same items and eXpressed concepts similar to Cluster II: Factor IV, "Purposelessness--1ow motiva- tion is related to lack of feeling of direction with concomitant feel- ings of anxiety"; Factor V, "Success Drive--high motivation for success is based on attempt to prove self as a defense against family interfer- ence"; and Factor VI, "Normlessness--low motivation is related to hostility towards social standards. Contains overtones of psychopathic characteristics." Cluster II also has similarities with two factors from the General Situational Choice Inventory: Factor I, "Unique Versus Common Accomplishment--the high motivated male choses responses descrip- tive of unusual tasks for the typical male, the low motivated male choses to do what most of his classmates will do"; and Factor III, "Competition With Versus Ease of Meeting a Standard--the high motivated male choses reSponses describing a standard as personal challenge, the low motivated male choses the reSponses which require the least effort." The concepts expressed in Cluster II logically relate to the factor descriptions from the Farquhar study. Cluster III: Item content (three items) and indices of association for Cluster III are presented in Table 4.4. There is little sub-scale over- lap with item 15 being from the Human Trait Inventory and items 65 and 49 from the General Situational Choice Inventory. Cluster III was intern- ally consistent by both methods of computation used. The high internal consistency was a result of a low index of association with other clusters. 41 TABLE 4.4 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER III. 1 Index of Item Number and Content Association 15. I like to study ; 163 65. I would prefer to: (a) be known as a person who knows his own mind, or (b) be known as a person who gets help in making decisions. . 163 49. I would prefer to: (a) have average ability and be liked by many people, or ' (b) have superior ability but not be liked by as many people. 154 Cluster III was labeled Need Achiever. Item content indicates that the type of person characterized possesses need achievement to the ex- clusion of need affiliation. He is inclined toward academic interests rather than interpersonal interests. Both items 65 and 49 express the feeling of independence from other people and a desire for self-accomp- lishment. The type of person described is highly motivated academically. Cluster III has no correlate factor in the Farquhar study indi- cating that this cluster describes an independent concept missed by the factor analysis. In fact, the legitimacy of Cluster III as a valid, meaningful cluster is questionable because of its small number of items, its rank as the least representative cluster in the total matrix, and the fact that two of the items (65 and 49) did not meet the criteria for in- clusion into a factor in the Farquhar study.1 As a result, its contribu- tion to the basic conceptual patterns composing the M-Scales is limited. 11bid., pp. 149-51. 42 Cluster IV: The item content and indices of association for Cluster IV are presented in Table 4.5. Cluster IV contains two items, both from the General Situational Choice Inventory, which are practically identical in content. It is internally consistent by virtue of being composed of just the reciprocal pair. Cluster IV was labeled Competition Versus Ease of Meeting a Standard. Both items indicate a contrast between the type of student who seeks maximum accomplishment and the student who wants to fulfill only minimum requirements. The idea of risk taking is also involved in both items. A polarized concept is characterized in which the highly motivated student competes for maximum achievement and the lowly motivated student avoids competition and settles for achieving a mini- mum standard. TABLE 4.5 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER IV. Index of Item Number and Content Association 32. I would prefer to: (a) have the teacher give everyone the same grade at the beginning of the term and know I had passed, or (b) take chances on getting a higher or lower grade at the end of the course. 197 50. I would prefer to: (a) have everybody in the class get a "C" at the beginning of the course, or (b) be graded at the end of the course with the possibility of getting a higher or lower mark. 197 43 Both items in Cluster IV loaded on Factor I of the General Situational Choice Inventory, "Unique Versus Common Accomplishment," cited above in the Farquhar study.1 Cluster V: The item content and indices of association are summarized in Table 4.6. Five items were associated in Cluster V, all from the General Situational Choice Inventory. Based on the reciprocal items Cluster V was internally consistent but not when all items were used. Cluster V was labeled Intellectual Versus Material Accomplish- .EEEE' Cluster V is a polar dimension which attributes high motivation as having a different goal direction than low motivation. The highly motivated student is interested in intellectual achievement in school and the lowly motivated student is interested in material or, at least, non-academic accomplishment. Both of the reciprocal items (70 and 54) express the basic dimensional concepts. In addition, item 33, as well as items 54 and 70, suggest that a desire to be innovative or creative also contributes to high motivation. The significance of Cluster V to the interpretation of the dimensions of the M-Scales is limited by its lack of dimensional repre- sentation as indicated by a rank of thirteen among the cluster mean indices of association. Although this cluster is not comparable to any factor in the Farquhar study, it shares some of the same items as Factor II, "Immediate Versus Long-Term Gratification--The highly motivated male chooses responses describing delayed reward, the lowly motivated male chooses the immediate rewards."2 11bid., p. 157. 21bid. 44 TABLE 4.6 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER V. Index of Item Number and Content Association 54. I would prefer to: (a) think of an idea that nobody has ever thought of, or I (b) set a world's Speed record. 158 70. I would prefer to: (a) discover a gold mine, or (b) discover a new medicine. 158 48. I would prefer to: (a) have a great deal of money, or (b) be an eXpert in my favorite school subject. 156 39. I would prefer to: (a) be allowed to take extra courses before or after school, or (b) just take courses offered during the school day. 145 33. I would prefer to: (a) develop a new_product which may or may not be good, or (b) make a product as good as the best one available. 141 Cluster 2;: Item content and indices of association for Cluster VI are summarized in Table 4.7. Six items, all from the General Situational Choice Inventory, are associated in Cluster VI. Items 58 and 60 formed the reciprocal pair. Internal consistency was evidenced when computa- tion was based on the reciprocal items but not when all of the items were used. Cluster VI was labeled Immediate Versus Long-Term Involvement. Cluster VI polarizes motivation by distinguishing between students who are able to defer immediate satisfactions to obtain future, more 45 important accomplishments against those who prefer immediate gratifica- tion. The contrast in temporal orientation takes the form of academic versus non-academic interests. The two reciprocal items illustrate this: number 58 relates to a non-academic situation and number 60 to an academic situation. The student who is not able to set long-term goals for him- self is less likely to see the value of an education. Cluster VI correSponds to Factor II of the Farquhar study cited above "Immediate Versus Long-Term Gratification."1 TABLE 4.7 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER VI. Index of Item Number and Content Association 58. I would prefer to: (a) wait ten years and receive fame throughout the nation, or (b) receive fame in my community overnight. 197 56. I would prefer to: (a) work overtime to make more money, or (b) get more schooling to make more money. 190 31. I would prefer to: (a) be well prepared for a job after gradua- tion from high school, or (b) be well prepared to continue learning. 178 60. I would prefer to: (a) study to go to college, or (b) study to get out of high school. 197 30. I would prefer to: (a) buy a car, or (b) continue my education. 193 67. I would prefer to: (a) put together a new object, or (b) develop new ideas. 147 11bid. 46 Cluster VII: Item content and indices of association for Cluster VII are summarized in Table 4.8. Cluster VII contains twenty-seven items drawn from all of the sub-scales with the Preferred Job Characteristics Scale most heavily represented with thirteen items. All of the items are associated with item 86 of the reciprocal pair. On the basis of its column sum of associations item 86 is the most representative item in the entire male battery of M-Scales. Selection of the second, third and fourth most representative items (90, 87 and 79 respectively) in Cluster VII illustrated the advantage of basing reference item selec- tion on the mean association items have with the reciprocal pair over using the sum of the column associations. An example is item 7 (ranked 6th) which, when computed by the sum of associations procedure was ranked second, even above item 90, one of the reciprocals. Item 7 is not highly representative of the basic conceptual emphasis in Cluster VII. The substantial integration of items from the sub-scales in Cluster VII suggests along with Cluster II that the sub-scales overlap in the dimensions they measure. Cluster VII ranked as the second most representative cluster in the total male battery of M-Scales. Cluster VII was labeled Occupational Reaponsibility: High Versus Log, The four items selected as the most representative of the Cluster (86, 90, 79 and 87) all express the concept of high versus low occupa- tional reaponsibility. The types of students contrasted are those who see a job as requiring personal involvement, competence and leadership as against lowly motivated students who desire work that requires the least amount of responsibility. Presumably both types view their school work in the same manner. Along with the idea of reSponsibility is the element of ambition for (versus a lack of) unique, creative or signifi- cant accomplishment as evidenced in items 61, 66, 69, and 47. 47 TABLE 4.8 ITEM CONTENT.AND INDICES OF ASSOCIATION FOR CLUSTER VII. W" Index of Item Number and Content Association 90. I would prefer: (a) a job where my opinion is valued, or (b) a job where I make few if any decisions. 216 86. I prefer: (a) a job which requires little thinking, or (b) a job where my opinion is valued. 216 87. I prefer: (a) a job where I make few if any decisions, or (b) a job where I could become known for out- standing accomplishments. 213 7. Most of my school subjects are a complete waste of time. . 212 25. I like to go to the movies more than once a week. 205 8. Most of my school subjects are useful. 201 16. I like to plan very carefully what courses I will take in school. 176 119. Teachers feel that I am lazy. 201 117. Teachers feel that I am a person who postpones. 195 61. I would prefer to: (a) have a great deal of influence over people, or (b) have a great deal of ambition. 164 66. I would prefer to: (a) do something like everyone else, or (b) do something outstanding. 199 69. I would prefer to: (a) do something that I have done before, or (b) do something that I have never done before. . 194 47. I would prefer to: (a) have a better job than my father. has, or (b) have a job like my father has. 183 48 Table 4.8 - Continued :L— . Index of Item Number and Content Association 79. I prefer: (a) a job where I could decide how the work is to be done, or (b) a job where I make few if any decisions. 211 82. I prefer: (a) a job where I make few if any decisions, or (b) a job where I solve problems no one else can. 211 37. I would prefer to: (a) be graded at the end of a course with . the possibility of making an "A", or (b) get a "C" at the beginning of a course along with everyone else. 190 78. I prefer: (a) a job which requires little thinking, or (b) a job where I solve problems no one else can. 210 75. I prefer: (a) a job where my opinion is valued, or (b) a job where I could not be fired. 210 88. I prefer: (a) a job where I could not be fired, or (b) a job where I could decide how the work is to be done. 201 81. I prefer: (a) a job where I could not be fired, or (b) a job which absorbs my interests. 200 83. I prefer: (a) a job where I could become known for outstand- ing accomplishments, or (b) a job which requires little thinking. 208 94. Teachers feel that I am inefficient. 207 104. Teachers feel that I am uninterested. 203 64. I would prefer to: (a) be very happy, or (b) have lots of money. 178 76. I prefer: (a) a job which does not tie me down, or (b) a job which absorbs my interests. 198 80. I prefer: (a) a job which does not tie me down, or (b) a job where I could continue to learn the rest of my life;‘ 196 26. I would like to belong to a motorcycle club. 175 49 As expected from the fact that Cluster VII contained the greatest variety of items from the sub-scales, it also incorporated several of the concepts expressed in factors identified in the Farquhar study. They are as follows: G.S.C.I. Factor I: "Unique Versus Common Accomplishment"1 W.R.L. Factor II: "Educational Resistive"2 H.T.I. Factor III: "Academic Negativism"3 P.J.C.S. Factor I: "High Versus Low Job Involvement" Factor II: "Long-Term Versus Short-Term Job Involvement"4 Cluster VIII: Item content and indices of association for Cluster VIII are presented in Table 4.9. Nine items composed Cluster VIII with all but two of these items coming from the Preferred Job Characteristics Scale. Items 91 and 89 formed the reciprocal pair. The cluster was internally consistent based on selected items but not when all its items were computed. Cluster VIII was labeled Occupational Self-Development: High Versus Low. It characterizes the highly motivated male who sees a job as a means of personal and professional self-development versus the lowly motivated male who presumably views work as merely a necessity for earning a living. There is also an element of unique, creative accom- plishment or competition versus common achievement relative to occupa- tional involvement expressed primarily in items 85, 77, 74 and 72. 1Farquhar, Motivation Factors . . . , p. 157. 21bid., p. 138. 31bid., p. 147. 4Ibid., p. 168. 50 TABLE 4.9 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER VIII. Index of Item Number and Content Association 89. I prefer: (a) a job which does not require a college educa- tion,.or (b) a job where I could continue to learn the rest of my life. 215 85. I prefer: (a) a job with short working hours, or (b) a job where I solve problems no one else can. 192 91. I prefer: (a) a job which does not require a college education, or (b) a job where I could express my ideas, talents, and skills. 215 77. I prefer: (a) a job where I could become known for out- standing accomplishments, or (b) a job which does not require a college education. 211 74. I prefer: (a) a job where I solve problems no one else can, or (b) a job which does not require a college education. 195 73. I prefer: (a) a job which does not require a college education, or (b) a job where I could decide how the work is to be done. 194 72. I prefer: (a) a job where I solve problems no one else can, or (b) a job which permits me to take days off when I want. 169 41. I would prefer to: (a) do as well as most of my classmates, or (b) do better than most of my classmates. 185 21., I have trouble waiting for a class to be over. 177 51 Five of the eight items contain the alternative response, "A job which does not require a college education." The lowly motivated student described by these items may be expressing a disinterest in going to college more than a lack of interest in occupational self- development. In this sense, the cluster may be mostly an expression of pure intellectual versus non-intellectual interest. Statistically, Cluster VIII best represents the basic dimensions measured by the male items of the M-Scales as indicated by the extent to which other clusters associated to it. Elements of all three Preferred Job Characteristics Scale fac- tors in the Farquhar study were evident in Cluster VIII with Factor III, ”Unique Accomplishment Versus Avoidance of Education," being most com- parable.1 Cluster IX: Item content and indices of association for Cluster IX are summarized in Table 4.10. Three items, all from the Word Rating List, are associated in Cluster IX. Items 101 and 94 were reciprocal. Inter- nal consistency was evident in the reciprocal items but not in the com- putation using all of the items. TABLE 4.10 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER IX. Index of Item Number and Content Association 101. Teachers feel that I am thorough. 176 114. Teachers feel that I am systematic. 162 97. Teachers feel that I am logical. 176 11bid., p. 168. 52 Cluster IX was labeled Compulsivity. It characterizes the male who perceives himself through his teachers' eyes as being well organized, disciplined and systematic. He is typically a highly motivated student. His achievement possibly is more a result of his organization and effort than his intelligence. It suggests that some academic motivation is neurotically produced and is, therefore, undesirable. Although this cluster is not directly analogous to any of Farquhar's factors three of the same items loaded on Factor I of the Word Rating List, "Academician." The student described by the Farquhar factor "...sees himself through the teachers' eyes as effective, bright, orderly, goal oriented and amenable to learning."1 Cluster X: Item content and indices of association for Cluster X are presented in Table 4.11. Cluster X contains thirteen items with all but one coming from the Word Rating List. Items 98 and 124 were reciprocal and item 99 was also selected as a reference item. The discrepency between the results obtained by the two procedures used for selecting reference items is illustrated again in Cluster X in most of the items. Cluster X showed high internal consistency because it was consistent by both computations used. Cluster X was labeled Self-Perceived Competence. The type of student described perceives himself through his teachers' eyes as having a high degree of academic ability. He possesses a general positive academic self-image and high motivation. ' Cluster X ranks twelth regarding its representativeness of the central dimension of the male M-Scales. 1Ib1d., p. 138. 53 TABLE 4.11 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER X. Index of Item Number and Content Association 98. Teachers feel that I am smart. 175 93. Teachers feel that I am talented. 174 99. Teachers feel that I am successful. 173 131. Teachers feel that I am competent. 166 130. Teachers feel that I am a planner. 164 122. Teachers feel that I am intellectual. 163 18. I plan my activities in advance. 4 151 108. Teachers feel that I am original. 165 124. Teachers feel that I am above average. 175 125. Teachers feel that I am productive. 171 105. Teachers feel that I am studious. 170 111. Teachers feel that I am in-the—know. ’ 164. 118. Teachers feel that I am exacting. 160. Eleven of the thirteen items appear in Word Rating List Factor I, "Academician," in the Farquhar study.1 Cluster XI: Item content and indices of association for Cluster XI are presented in Table 4.12. Cluster XI contains fourteen items with thirteen of them contributed by the Word Rating List. Items 107 and 138 formed the reciprocal pair with 138 being the most representative of the cluster. Items 116 and 123 were also selected as reference items. Internal consistency was evident when based upon the selected items but not when all items were used. Ibid. 54 TABLE 4.12 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER XI. Index of Item Number and Content Association 107. Teachers feel that I am responsible. 189 116. Teachers feel that I am dependable. 186 102. Teachers feel that I am orderly. 176 109. Teachers feel that I am consistent. 170 100. Teachers feel that I am careful. 177 110. Teachers feel that I am intelligent. 171 128. Teachers feel that I am contented. 158 138. Teachers feel that I am reliable. 189 133. Teachers feel that I am teachable. 182 92. Teachers feel that I am patient. 168 139. Teachers feel that I am serious. 178 123. Teachers feel that I am alert. 176 95. Teachers feel that I am practical. 174 24. I like to be consistent in the things I do. 160 Cluster XI was labeled Self-Perceived Re8ponsib1eness. The 'student characterized here perceives himself as being sincere, dependable and hard working. He possesses high motivation. Cluster XI ranked fifth in representativeness of the basic dimension of the M-Scales. All of the Word Rating List items in Cluster XI loaded on Factor I, l "Academician," in the Farquhar study. Apparently this dimension contributes substantially to the meaning of Factor I. 11bid., p. 134. 55 Cluster XII: Item content and indices of association for Cluster XII are presented in Table 4.13. Four items, all from the Word Rating List, contribute to Cluster XII. Items 127 and 136 were reciprocal. Internal consistency was indicated in both computations. Cluster XII was labeled Self-Perceived Intellectual Ambition. The highly motivated male perceives himself to be actively agressive and ambitious regarding academic studies. No doubt he possesses many of the characteristics found in Clustersliand XI since words like "thinker" and "efficient" are practically synonomous with words like "intelligent" ' (Cluster XI) and "productive" (Cluster X). The emphasis, however, is on the ambition. Cluster XII ranks eleventh in statistical representativeness of the central male dimension of the M-Scales. All of Cluster XII's items loaded heavily on Word Rating List Factor I, "Academician," in Farquhar's study.1 In fact, items 129 and 126 had the first and third highest loadings reSpectively. TABLE 4.13 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER XII. Index of Item Number and Content Association 127. Teachers feel that I am ambitious. 175 129. Teachers feel that I am an achiever. 169 126. Teachers feel that I am a thinker. 172 136. Teachers feel that I am efficient. 175 Ibid. 56 Cluster XIII: Item content and indices of association for Cluster XIII are summarized in Table 4.14. Nineteen items were associated in Cluster XIII, fifteen of which were contributed by the General Situational Choice Inventory. Cluster XIII was the only cluster to have items with tie linkages, namely, items 28 and 36. Another irregularity was the occur- -rence of two sets of reciprocal pairs (items 36 and 43 and items 68 and 64).1 Internal consistency was evidenced in the computation using selected items but not when all items were used. The four reciprocals were selected as the reference items. Reciprocal items 68 and 84 were statistically more representative of the cluster dimensions than items 36 and 43 on the basis of their column sum of associations. Item 113 had a high linkage to reciprocal items 68 and 84 and, in fact, on this basis it was more representative than reciprocal items 36 and 43 since they had a lower linkage with items.68 and 84 than item 113. Item 113 again illustrates the discrepency between the item rankings in the two procedures used to identify reference items; the column sum ranked item 113 ninth and the mean association with the reciprocals ranked it third. Cluster XIII was labeled Need Achievement: High Versus Low. The achiever is characterized as placing high personal value on accomplish- ment and having a desire to excell. He is competitive and considers himself to be competent. He is self-disciplined, purposeful, and a hard worker. The male with low need achievement meets minimum standards. He 1Cluster XIII could have been Split into two separate clusters with a reciprocal pair in each one. The logical dividing points would have been between items 28 and 96 and between items 36 and 68 (See Appendix B). However, in light of the high linkage between items 36 and 68 it was decided to maintain the items as one cluster. This decision was influenced by a suggestion from Dr. Louis McQuitty in a personal con- versation June 25, 1965. 57 TABLE 4.14 ITEM OONTENT.AND INDICES OF ASSOCIATION FOR CLUSTER XIII. l Index of Item Number and Content Association 68. I would prefer to: (a) be demanding on myself to do good work, or (b) be demanding on my friends so that they will do good work. 192 36. I would prefer to: (a) get excellent grades because I have a great deal of ability, or (b) get average grades because I have average ability. 187 29. I would prefer to: (a) have the best teachers in the state in my school, or (b) have a large recreation center in my school. 173 42. I would prefer to: (a) be considered as being strong but not very smart, or (b) be considered as being weak but smart. 163 34. I would prefer to: (a) receive money for my good grades, or (b) be allowed to take any course I wanted because of good grades. 156 57. I would prefer to: . (a) inherit a great deal of money,or (b) earn a great deal of money. 153 40. I would prefer to: (a) complete a job which I recognize as difficult, or (b) complete a job which others recognize as difficult. 142 44. I would prefer to: (a) work at many less important jobs which I know I could finish, or (b) work at one very important job which may never be entirely finished in my lifetime. 141 58 Table 4.14 - Continued -. ._-I m: 11 Index of Item Number and Content Association 28. I would prefer to: (a) receive a grade on the basis of how well I did on the teacher‘s test, or (b) get a grade on the basis of how hard I tried. 131 43. I would prefer to: (a) be known as a person with much ability, or (b) be anwn as a person with adequate ability. 187 113. Teachers feel that I am nervous. 183 96. Teachers feel that I am confident. 157 45. I would prefer to: (a) be paid for how well I did a job, or (b) be paid the same amount no matter how I did the job. 167 52. I would prefer to: (a) be paid for the amount of work I did, or (b) be paid by the hour. 147 84. I would prefer: (a)'a job which has high work standards. (b) a job where I make few if any decisions. 192 53. I would prefer to: (a) study my assignments during study hall, or (b) wait to study until the mood strikes me. 175 59. I would prefer to: (a) plan my life in advance, or (b) live my life from day to day. 166 63. I would prefer to: (a) be known as being a "good guy" or a "good gal," or (b) be known as a person who "does things well." 152 103. Teachers feel that I am purposeful. 149 59 is less competitive and lacks the drive to obtain extraordinary accom- plishment. Cluster XIII is distinguished from a similar cluster and its closest associate, Cluster VII. The latter emphasized an attitude toward work whereas Cluster XIII emphasizes obtaining a result, namely, accomplishment. Items 68 and 84, the reciprocals, give the impression that this type of person is demanding of himself but not necessarily of others. Cluster XIII was ranked as the sixth most representative cluster. Cluster XIII does not have a counterpart factor in the Farquhar study. Cluster XIV: Item content and indices of association for Cluster XIV are presented in Table 4.15. Eight items clustered into Cluster XIV with seven contributed by the Word Rating List. In addition to the reciprocals (items 115 and 120), items 112 and 134 were selected as being most representative of the cluster. Item 112 tied with 115 for the highest column sum. Because Cluster XIV was not internally consis- tent by either procedure used and is distinctly similar in content to Cluster II, it could be interpreted as part of Cluster II. Cluster XIV was labeled Self Perceived Non-Conformity. Cluster XIV characterizes a lowly motivated student. He perceives himself through his teachers' eyes as being reckless, rebellious, different, distracted, carefree, stubborn and impatient. A question which cannot be answered here is whether the student actually perCeives himself in the same way as he feels his teachers do.1 1The'Word RatingJList was constructed on the premise that "sig- nificant others" influence the develOpment of one's self concept. It was further assumed that teachers were significant others and therefore a picture of a student's self concept could be observed in his impres- sion of how he was perceived by his teachers. Payne, A Dimension Analysis . . . , p. 6. 60 TABLE 4.15 ITEM CONTENT AND INDICES OF ASSOCIATION FOR CLUSTER XIV. Index of Item Number and Content Association 115. Teachers feel that I am reckless. 199 112. Teachers feel that I am rebellious. 196 106. Teachers feel that I am different. 177 137. Teachers feel that I am easily distracted. 171 121. Teachers feel that I am carefree. 167 6. I f1irt.i 165 120. Teachers feel that I am stubborn.. 199 134. Teachers feel that I am impatient. 188 Cluster XIV contains most of the same items as Farquhar's Factor II, "Educationally Resistive."1 Cluster XIV was ranked as the eighth most representative cluster. Discussion of Elaborations of E.L.A. In addition to the results of the basic Elementary Linkage Analysis, concern in this Study was extended to include a comparison of two procedures for selecting reference items and computation of second order solutions by means of the Hierarchflufl Method and the Comprehensive Hierarchical Approach. Identification of Reference Items A reference item has been defined as the item which is most repre- sentative of the dimensional core of its cluster. In the present .w- 1Farquhar, Motivation Factor . . . , p. 133. 61 investigation several reference items were identified in each cluster. Theoretically, the two most representative items in each cluster are the reciprocal pair. The representativeness of the other items was determined by their average agreement or index of association with the reciprocals. This procedure was preferable to selection based on the magnitude of their column sums of associations, because theoretically an item's linkage with the reciprocal pair is the best indication of its contribution to the dimensional core of its cluster. It was found that these two procedures yield comparable results, although in certain clusters there are distinct discrepencies.1 For example, in Cluster VII item 7 was ranked second on the basis of its column sum of associations but ranked sixth in its association with the reciprocals. Similarly, in Cluster VIII, item 77, which is not a member of the reciprocal pair, was ranked first on the basis of its column sum but ranked third in its relationship to the reciprocals.2 Other less significant discrepent rankings resulted in addition to these cited. It was mentioned in Chapter III that the same reference items, statistically identified, may not always be selected on a subjective basis. In every case subjective judgment in this investigation agreed with reference items selected on the basis of their association with the reciprocals. In fact, support for the preference of this procedure over. 1The comparative item rankings by these two procedures is pre- sented in the cluster matrices in Appendix B. 2Cluster VIII was the only cluster in which a non-reciprocal was ranked first on the basis of column sums of association. 62 using column sums was illustrated in Cluster VII where the column sums method selected an item which did not directly contribute to descrip- tion of the dimensional core. The item in point is item 7 which was ranked second by the column sums method but does not directly relate to the concept of "occupational reSponsibility."1 Second Order Solutions In both the Hierarchical and Comprehensive procedures the clus- ters were forced to associate into second order clusters. The procedures did not establish a cut—off point for determining when a cluster should not be linked to another cluster but instead described an independent dimension. Therefore, the magnitude a cluster's association and its internal consistency had to be considered in determining the meaning of the second order linkages. The internal consistency based on the selected most representative items of each cluster indicates that all of the clusters except XIV represent independent dimensions. The rela- tionship of Cluster XIV to Cluster XIII described above accounts for Cluster XIV's lack of independence. Six of the fourteen clusters associated differently in the two second order procedures. ‘A comparison of the associations in the two procedures is presented in Table 4.16. As stated in Chapter III it was assumed that using selected items that were most representative of their cluster dimension is a sounder procedure for associating clusters than using all of the items 1Item 7 content is: "Most of my school subjects are a waste of time.” 63 TABLE 4.16 COMPARATIVE SUMMARY OF THE CLUSTER LINKAGES IN THE HIERARCHICAL METHOD AND THE COMPREHENSIVE HIERARCHICAL APPROACH Cluster‘Linkage‘in Cluster Linkage in the the Comprehensive Cluster Number Hierarchical Method Hierarchical Method I VII II II VII VII III X X IV IV VI V VI VI VI VIII VIII VII II II VIII VII VII IX x1 x1 X IX XII XI XIII IX XII XI XI XIII II VII XIV VII II in each cluster. The reason for this assumption was that in Elementary Linkage Analysis an item must have a close association with only one other item in the cluster to be included. Therefore, it is possible for those items most distantly linked to the reciprocals to be quite different from the reciprocals.in content. The fact that only four of the clusters were internally consistent when all the items were used in the computation suggested that items distantly linked did not contribute substantially to their reSpective cluster dimensions. Therefore, the assumption made regarding the desirability of using selected items seemed to be justified. 64 Hierarchical Method-~The item with the highest column sum of associations was selected from each cluster for the second order solution by this method.152 Since Cluster IV contained only a reciprocal pair both items were included as were items 112 and 115 in Cluster XIII since they had identical column sums. As a result, Cluster IV (reciprocal items 32 and 50) did not link to any other cluster nor did any cluster link to it. The implication from this is that this cluster is highly internally con- sistent and independent, having little relationship to the main dimensions measured in the male M-Scales. Since the Comprehensive Hierarchical Approach was considered to be the more valid procedure no attempt was made to interpret the clusters formed by the Hierarchical Method. Comprehensive Hierarchical Approachi-Nine clusters associated in second order Cluster A. Clusters II, VII and VIII had the closest associations and formed a three way reciprocal linkage. Based on its column sum, Cluster VIII was the most representative of the entire male M-Scales battery. Clusters II and VII also indicated high representativeness with column sums nearly as high as Cluster VIII. High indices occurred in the association of Cluster XIV with Cluster II (190) and Cluster XIII 1The item content and indices of association along with the matrix and graphic description of the item associations resulting from the Hierarchical Method are presented in-Appendix C. 2Since only one item was used, selection was based on column sums rather.than associations to reciprocals. This facilitated deciding which reciprocal would be used. 3The item matrix, cluster matrix and a graphic illustration of the second order cluster linkages resulting from the Comprehensive Hierarchical Approach are presented in Appendix D. 65 with Cluster VII (189). Other associations were relatively low. Cluster I had a tie association with Clusters II and VII. Second order Cluster A was labeled ReSponsibleness: High Versus £93. The concept was expressed in the two reciprocal clusters (II and VII). The relationship of responsibility to achievement is also suggested in the high association of Cluster XIII, "Need Achievement" with Cluster VII, "Occupational Reaponsibility." The achiever is characterized as being reSponsible, having long-range goals, preferring intellectual over material accomplishment, and viewing education and work as instruments of self—develOpment and self-esteem. He is also a conformist. The linkage of Clusters II, ”IrreSponsible," and XIV, "Self-Perceived Non- Conformity," indicates a relationship between conformity and reSponsi- bility. The achiever is also competitive and the association of Cluster IV, "Competition...," with Cluster VI, "Immediate Versus Long-Term Involvement," further suggests a relationship between competitiveness and an ability to establish long-term goals. Also, the linkage of Cluster V, "Intellectual Versus Material Accomplishment," with Cluster VI suggests that the student who is more interested in material achievement has difficulty establishing long-term goals and is more inclined toward immediate gratification. The association of Cluster V1 with Cluster VIII, "Occupational Self-Development," further suggests that perceiv- ing work as a means of self-develOpment involves delaying immediate wishes to achieve long-range goals. The highly motivated student is able to delay immediate material enjoyment and invest his time and money into an education in order to achieve occupational goals. Second order Cluster A achieved better integration of the sub- scales than second order Cluster B. It was also a better expression of 66 TABLE 4.17 SUMMARY OF CLUSTER LABELS AND INDICES OF ASSOCIATION FOR SECOND ORDER CLUSTER A Index of Cluster Number and Label Association VIII. Occupational Self-Development: High Versus Low 203 VI. Immediate Versus Long-Term Involvement 179 IV. Competition Versus Ease of Meeting a Standard 174 V. Intellectual Versus Material Accom- plishment 143 VII. Occupational Reaponsibility: High Versus Low 203 XII. Need Achievement: High Versus Low . 189 I. Distractable *Tie 165 II. Irresponsible 203 XIV. Self-Perceived Non-Conformity 190 I. Distractable *Tie . 165 TABLE 4.18 SUMMARY OF CLUSTER LABELS AND INDICES OF ASSOCIATION FOR SECOND ORDER CLUSTER B Index of Cluster Number and Label Association IX. Compulsivity 168 XI. Self-Perceived Reaponsibleness 168 XII. Self-Perceived Intellectual Ambition 165 X. Self-Perceived Competence 151 »III. Need Achiever 135 non-intellectual variables than Cluster B. The fact that second order Cluster A contained most of the clusters ranking high in dimensional representativeness further suggests that this cluster better expresses the basic dimensions measured by the male M-Scales. are presented in Table 4.19. Second order Cluster B linked five clusters with Cluster IX and XI which were associated reciprocally. Compared to those in second order Cluster A the indices of association in second order Cluster B were relatively low indicating a lack of dimensional independence. A lesser degree of dimensional representation was also suggested by the low rankings of its constituent clusters. All of the clusters (IX, X, XI, and X11)1 loaded heavily with Word Rating List items and all of the clusters characterized high motivation. TABLE 4.19 CLUSTER RANKINGS REGARDING THEIR DIMENSIONAL REPRESENTATION OF THE M-SCALES Rank Cluster Label 1 VIII Occupational Self-Development 2 VII Occupational Reaponsibility 3 II IrreSponsible 4 VI Immediate Versus Long-Term Involvement 5 XI Self-Perceived ReSponsibleness 6 XIII Need Achievement: High Versus Low 7 IV Competition Versus Ease of Meeting a Standard 8 XIV Self-Perceived Non-Conformity 9 I Distractable 10 IX Compulsivity ,ll XII Self-Perceived Intellectual Ambition 12 V Intellectual Versus Material Accomplishment 13 X Self-Perceived Competence 14 III Need Achiever 1The clustering of W.R.L. items indicates that this sub-scale measures a relatively independent dimension. The cluster rankings 68 Second order Cluster B was labeled Self-Perceived Ability. Although the type of person described may achieve because of his intel- lectual ability rather than his motivation, the important emphasis is on his perception of his ability. The achiever is characterized as hav- ing a positive academic self-concept. He perceives himself as being reSponsible, ambitious, and competent. The significant position of Cluster IX suggests that compulsivity is related to the traits just mentioned and contributestx>achievement. In fact compulsivity may at times be mistaken for reSponsibleness, ambition, and competence. The weak linkage of Cluster III is further evidence of its limited contri- bution to the dimensional interpretation of the male M-Scales. To the extent that Cluster III is relevant it suggests that need achievement contributes to high motivation. The cluster labels and indices of association for second order Cluster B are summarized in Table 4.18. Viewed together the second order clusters consolidate the dimen- sional concepts of the male M-Scales into two c0mprehensive traits. The male achiever is reSponsible and has a positive impression of his ability. Summary The Elementary Linkage Analysis yielded fourteen clusters. Three clusters, II, VII and XII, accounted for almost half of the items (68 out of 139) in the entire battery. The same clusters contained the most heterogenious representation of items from all of the sub-scales thus achieving the greatest degree of sub-scale integration. The fact that nearly half of the items clustered in an integrated fashion suggests that the sub-scales overlap in the dimensions they measure relative to academic motivation. 69 Reference items which statistically best represent the basic di- mension of their reSpective clusters were selected by two different pro- cedures for purposes of comparing the two techniques. In most clusters the same items were identified by both procedures although there were, in a few instances, distinct discrepancies. It was also observed that in six instances the second order cluster associations based on one reference item from each cluster differed from those produced by using several reference items. The method of considering the reciprocal pair to be the most representative items in their reSpective clusters and selecting other items on the basis of the magnitude of their average association with the reciprocals was the more satisfactory method because it gave a more representative picture of the cluster's dimen- sional core. The clusters were interpreted and described in terms of the traits each characterized relative to academic motivation. Interpreta- tion~was based on subjective judgment of the meaning conveyed by the items with special weight given to the reference items. Each cluster was given a label representing the main characteristics it portrayed. The clusters were labeled as follows: Cluster I: Distractable-low motivation Cluster II: IrreSponsible-low motivation Cluster III: Need Achiever-polar Cluster IV: Competition Versus Ease of Meeting a Standard-polar Cluster V: Intellectual Versus Material Accomplish- ment-polar Cluster VI: Immediate Versus Long-Term Involvement-polar Cluster VII: Occupational Responsibility: High Versus Low-polar Cluster VIII: Occupational Self-Development: High Versus Low-polar Cluster IX: Cumpulsivity-high motivation 70 Cluster X: Self-Perceived Competence-high motivation Cluster XI: Self-Perceived Responsibleness-high motivation Cluster XII: Self-Perceived Intellectual Ambition-high motivation Cluster XIII: Need Achievement: High Versus Low-polar Cluster XIV: Self-Perceived Intellectual Ambition-high motivation An index of the internal consistency of each cluster was derived on two different bases: (1) by using all of the items in each cluster in the computation; and (2) by using only the several selected reference items. A cluster was internally consistent if its mean within-cluster item association was greater than its mean intercluster linkage. Based on the latter procedure, which proved to be the more meaningful of the two methods, all of the clusters except Cluster XIV were internally con- sistent. Cluster XIV was short of the criterion by a narrow margin. The implication was that Clusters I through XIII represent independent dimensions and that Cluster XIV could be interpreted as part of Cluster'II. As an elaboration of E.L.A. second order solutions were com- puted by two procedures, the Hierarchical Method and the Comprehensive Hierarchical Approach. -A1though each method yielded two second order clusters, they differed in the manner in which the original clusters associated. The latter procedure was considered to be more meaningful because of its comprehensiveness. Therefore, interpretation was limited to the results obtained by the Comprehensive Hierarchical Approach which included two second order clusters. Second order Cluster A contained nine clusters including most of those which were statisti- cally most representative of the dimensional core of the total male battery of M-Scales as indicated by their column sums of associations .in the second order matrix. The high degree of item integration in 71 Cluster A suggested that the sub—scales overlap in the dimensions they measure 0 Second order Cluster A was labeled Re3ponsibleness: High Versus L93. Second order Cluster B contained five clusters composed almost entirely of items from the Word Rating List indicating that academic self-concept is an independent dimension. Cluster B was labeled S91E- Perceived Ability. The academically motivated male student as characterized by the dimensions of the male M-Scales has a high sense of reSponsibility and perceives himself to be academically competent. The concept of respon- sibility is more representative of nonJIntellective variables than the concept of self-perceived ability. It is also more meaningful because the clusters in which it was characterized achieved a higher degree of item integration and were the most dimensionally representative clusters as indicated by their high indices of association. On the basis of the analysis, the hypothesis, underlying this study, that meaningful and descriptive dimensions which relate to academic motivation can be isolated and interpreted, was accepted as tenable. CHAPTER V .SUMMARY, CONCLUSIONS, AND IMPLICATIONS Summary The purpose of this study was to achieve a more comprehensive, meaningful description of the dimensions underlying academic motivation than has been achieved previously. The problem was to isolate and describe the dimensions of an objective measure of academic motivation called the Michigan M-Scales. The four sub-scales composing the M-Scales were each developed from a different theory of academic motivation. Therefore, the purpose of this investigation was to analyze the entire battery to achieve an integrated comprehensive description of the con- stituent dimensions. It was hypothesized that meaningful and descrip- tive traits related to academic motivation could be isolated and identified. This study was conducted as part of a larger research project Sponsored by the United States Office of Education under the direction of William W. Farquhar.1 Data gathered in previous stages of the Farquhar project, relative to the reSponses of 240 male eleventh grade high school subjects to the 139 male items in the M-Scales, was used in this investigation. 1Farquhar,,Motivation Factors . . . 72 73 The Elementary Linkage Analysis procedure was selected for data analyzation because of its simplicity and objectivity. A cluster as identified by E.L.A. was defined as a structure in which every item is more like some other item (i.e., individuals who responded one way to one item responded similarly to another item) in that cluster than any item in any other cluster. Giving special weight to the reference items, which statistically are most representative of the dimensional core, the clusters were described according to the characteristics relative to academic motivation held in common by their constituent items. Two elaborations of E.L.A., the Hierarchical Method and the Comprehensive Hierarchical Approach,.each yielded two interpretable second order clusters. Only the clusters from the latter method were interpreted. In addition, information regarding the relationship be- tween the reference items from each cluster as well as the relationships between the clusters was obtained- Using information from the Compre- phensive Approach an indication of the internal consistency of the clusters was determined which facilitated more meaningful cluster interpretation. The Elementary Linkage Analysis yielded fourteen clusters. Three clusters, II, VII, and XIII, accounted for almost half of the items (68 out of.l39) in the entire battery- The same clusters con- tained the best representation of items from the sub-scales thus achiev- ing the highest degree of subescale integration. The two techniques used to select reference items identified the same items in most clus— ters. However, in several clusters discrepancies in the items identi- fied by the two procedures resulted. In addition, six of the second order clusters had different linkages when the computation was based on one reference item than they had on the basis of several selected reference items- The labels assigned and interpretive emphasis of each cluster 74 TABLE 5.1 SUMMARY OF CLUSTER LABELS AND INTERPRETIVE EMPHASES Cluster Label Interpretive Emphasis II III IV Distractable Irresponsible Need Achiever Competition Versus Ease of Meeting a Standard Intellectual Versus Material Accomplishment Low Low High Polar Polar ‘ I The lowly motivated male has difficulty concentrat- ing on his studies and is easily diverted from studying and acadeudc pursuits. The low1y_motivated male lacks an internalization of accepted behavior norus, dislikes conform— ing to expectations and authoritarian control, lacks competitiveness and goal direction, and finds personal significance in non—conventional and non- academic activities. The highly motivated male possesses achievement needs to the exclusion of affiliation needs. He is inclined toward academic rather than interpersonal interests. The highly motivated male competes for maximum achievement. The lowly motivated male avoids competition and seeks the easiest means of meeting requirements. The highly motivated male seeks intellectual or vacademic accomplishment. The lowly motivated male prefers immediate grati- fication. 75 Table 5.1 - Continued ’—Y Cluster Label Interpretive Emphasis VI VII VIII IX XI Immediate Versus Long- Term Involvement Occupational Responsibil- ity: High Versus Low Occupational Self—Develop— ment: High Versus Low Compulsivity Self-Perceived Competence Self-Perceived Respons- ibleness Polar Polar Polar High High High The highly motivated male is able to defer immedflne satisfactions to obtain future, more important accomplishments. The lowly motivated male pre- fers immediate gratifi-~ cation. The highly motivated male views an occupation as requiring personal in— volvement, compefience and leadership and as an opportunity for unique or creative accomplishment. The lowly motivated male desires work that requirw little responsibility. The highly motivated male views a job as a means of personal and professional self-development. The lowly motivated male views- work as a necessity for earning a living. He is not interested in college. The highly motivated male sees himself through his teachers' eyes as being well organized, discir plined, and systematic. The highly motivated male sees himselfrthrough his teachers' eyes as.possess— ing high intellectual competence. The highly motivated male sees himself through his teachers' eyes as being sincere, dependable and hard—working. 76 Table 5.1 - Continued Cluster Label Interpretive Emphasis XII XIII XIV Self-Perceived Intellectual High Ambition Need Achievement: High Polar Versus Low Self—Perceived Non- Low Conformity The highly motivated male sees himself through his teachers' eyes as academ- ically agressive and am— bitious. The highly motivated male places high personal value on accomplishment and excellence. He is competitive, self—dis- ciplined, purposeful, and sees himself as being competent. The lowly motivated male meets min- imum standards. He lacks competitiveness and the drive to obtain extra- ordinary accomplishment. The lowly motivated male sees himself through his teachers' eyes as be- ing reckless, rebellious, different, distracted, carefree, stubborn and impatient. 77 Internal consistency was evident in four clusters when all of each cluster's items were used in the computation. However, all except Cluster XIV indicated internal consistency when only the selected refer- ence items were included in the computation. Two elaborations of ELA, the Hierarchicaljyethod and the £227 prehensive Hierarchical Approach, each produced two second order clusters but differed in the patterns in which the clusters associated. The labels and interpretive emphases for the second order clusters (identified by the Comprehensive Approach) are summarized in Table 5.2. Conclusions Several conclusions were drawn that pertain to the Elementary Linkage Analysis procedures: 1. Reference items selected on the basis of their average association with the reciprocals were more representative of their cluster dimension than reference items selected on the basis of their column sums of association. 2. Cluster internal consistency was more accurately deter- mined when computed from selected reference items than when all items in each cluster were used. 3. The Comprehensive Hierarchical Approach produced a more accurate and meaningful second order solution than the Hierarchical Methpd. 4. Items with distant linkages in clusters produced by Ele- mentary Linkage Analysis sometimes make little contri- bution to their cluster's dimension. A number of conclusions were made based on the logical interpre- tation of the results of the Elementary Linkage Analysis: 78 TABLE 5.2 SUMMARY OF LABELS AND INTERPRETIVE EMPHASIS 0F SECOND ORDER CLUSTERS Y Cluster Label Interpretive Emphasis A Reaponsibleness: High Versus Low Self-Perceived Ability The highly motivated male is reSponsible, conforming and competitive. He is guided by long-range goals and views an occupation as a means of self-develop- ment and personal accom- plishment. The lowly motivated male is irresponsible, non- conforming and non-com- petitive. He seeks immediate gratifications and has a weak apprecia- tion for occupational self-deve10pment. The highly motivated male has a positive academic self-concept. He per- ceives himself through his teachers' eyes as being responsible, ambitious, and competent. He is also compulsive. 79 1. The sub-scales overlapein the-dimensions they measure and a meaningful integration of the.sub-sca1es was achieved by the analysis procedures. 2. The dimensions relative to academic motivation as measured by the male items of the Michigan M-Scales were isolated and meaningfully interpreted and.described. 3. The clusters represent independent dimensions.relating to academic motivation-, The clusters.are also internally consistent. 4. The clusters identified.by the Elementary Linkage Analysis were comparable inwpart to factors produced by factor analysis. 5. The dimensions identified in the MrScales indicate that the non—intellective.traits.mostlsignificant in determining the extent of a.ma1e studentis.academic motivation are his sense of responsibility and selfjperception of ability. 6. Self-concept as measured by the Word Rating List functions as an independent construct with regard to academic motiva- tion. Discussion The results of this investigation represent a more comprehensive and objective attack on the problem of defining and understanding academic motivation than previous studies.. However, the dimensions produced here do not account for all of the non-intellective variance in academic per- formance. -Academic motivation continues to be a vaguely defined, elusive concept. Further integration of factors affecting motivation is needed. Researchers are obviously limited to existing notions and hunches about non-intellective factors that influence motivation. Item construction ,80 is based on these concepts, therefore, factors identified by dimensional analyses are described with old-terminology. Perhaps.new dimensions are being isolated by the analysis.techniques.but go unrecognized because they are described with familiar.terms. An encouraging indication of the validity and reliability of the dimensions underlying the MrScales was made by the comparable labeling of the factors in the Farquhar study to the.clusters in this investiga- tion. This is significant in light of.the fact that labelings in the two studies were conducted independently- The cluster dimensions were also comparable to.some of the fac- tors cited in research.studies.reviewed in Chapter II. Clusters II and XIV characterized lowly.motivated students as being non-conforming. Correspondingly, in the Middleton and Guthrie1 study "Dependence" (which presumably often leads to conformity) was iden-:. tified as a factor characteristic of high achievers. Mitchell's2 "Academic Motivation and.Efficiency" factor is similar to ”Need Achievement" as described.in Cluster XIII. Also the "Nonacademic Achievement Motivation" factor in the Mitchell study has some of the same emphasis as the "Material:Accomplishment" aSpect of Cluster V. In the Michael, Jones and Trembly3 study the factor labeled "General Achievement Drive" corresponds with Cluster XIII, "Need Achieve- ment" and a "Compulsivity" factor was identified corresponding to Cluster IX in the present study. In addition, the factor labeled "Syste- matic, Methodical Application" has the same emphasis as Cluster IX, "Compulsivity." 1Middleton and Guthrie, pp. 66-69. ZMitchell, "An Analysis . . . ,," pp. 179-87. 3Michael, Jones and Trembly, pp. 667-71. 81 Because the results cited in the.research literature were not derived in any of the studies, within a theoretical framework the simi- larity between the factors identified with the clusters in this study has limited significance. The results of this study lend themselves to two concluding observations: First, the concept of gesponsibility implies doing what is necessary or expected. .Although-most educators have encountered a few students who enjoy and are excited about learning, the relationship of reSponsibleness to academic motivation.suggests that most students do not really enjoy studying and must discipline themselves to study because they realize it is necessatyw. Although this finding may be only statistical confirmation of what most practitioners already know it does raise the serious question of why our society and educational system do not cultivate a greater.appreciation and enjoyment of learning. Secondly, the immergence of selfeperceived ability as a major dimension in conjunction with-"reSponsibleness" raises the suggestion that a sense of reSponsibility is at least partially developed in response to one's perception of his ability. It can be conjectured that the stu- dent who is above average academically learns to assess his ability from evaluations of his performance made by significant others. These signi- ficant others (such as parents and teachers) encourage able students to use their ability even at times to the extent of making the students feel obligated to achieve or in a sense to feel reSponsible for their ability. Although, perpetual conjecturing about the interplay of variables that influence academic motivation continues, exceptions to almost every hunch are revealed. For example, the relationship suggested above be- tween self-perceived ability and reSponsibleness does not account for the 82 overachievement of students with low ability. The need for further moti- vation research is obvious. Implications for Future Research Several implications for further investigation are noteworthy: l. 2. Conduct a similar study using the female sample. Develop a new theoretical base from the results of this study; construct a revised scale giving weight to the basic dimensions identified here, and determine if the revised scale has increased predictive validity over the present M-Scales. Conduct a study to determine the reliability of the dimen— sions of the M—Scales. An identical analysis to the present study could be conducted using a new sample. Conduct an Elementary Linkage Analysis in which people rather than items would be clustered. Additional information regarding the subjects in each cluster would have to be gathered in order to describe the "types" of pe0ple charac- teristic of each cluster. Conduct a study identical to the present study on a sample of over- and under-achievers to compare the resulting di- mensions with those produced in this study. Rearrange the order in which the items are presented in the M-Scales so that the items in each sub-scale are not grouped together. The present investigation could then be duplicated and the resulting clusters compared with those identified in this study. Design a study to identify the variables that contribute to or inhibit the development of a sense of reSponsibility and a positive academic self-concept. 83 8. Compare the dimensions identified in this study with those identified by a clinical study on a sample of comparable subjects. The clinical study would involve interviews, autobiographical data, and test scores. The intention would be to compare subjective description with statistical description. 9. A study identical to the present investigation could be conducted using college students to determine the extent to which the same dimensions exist at the college level. 10. The present study should be duplicated on students in various parts of the United States and foreign countries to deter- mine the universality of the dimensions found here. 11. Rearrange the MrScales battery so that subjects' performance could be scored on each of the fourteen dimensions separately. The attempt to integrate variables related to academic motivation in this study achieved some comprehensiveness resulting in the identifica- tion of "reSponsibleness" and "self-perceived ability" as basic components of the motivation complex. Future research should be directed toward reducing the fragmentary conceptions of the variables that compose academic motivation. New items should also be generated for the other clusters to determine their actual contribution to explaining academic motivation. Finally, new studies need to be conducted which assess the predictiveness of the clusters as well as their interrelatedness. BIBLIOGRAPHY Bond, P.J. "The Relationship Between Selected Non-intellective Factors and Concealed Failure Among College Students of Superior Scho- lastic Ability." Unpublished doctoral dissertation, Purdue University, 1960, cited in "Academic Underachievement in College," Student Personnel Methods Bulletin, Claremont, California, College Student Personnel Institute (August), 1964. Farquhar, William W. '"An Integrated Research Attack on Academic Moti- vation," Journal of Counseling Psychology, IX (1962), 84-86. . -Motivation Factors Related to Academic Achievement. Final Report of U.S. Government Cooperative Research Project No. 846, Michigan State University: Office of Research and Pub- lications, College of Education (January), 1963. Farquhar, William W. .and Payne, David A. "Factors in the Academic- Occupational Motivations of Eleventh Grade Under- and Over- Achievers," Personnel and Guidance Journal, XLIII, No. 3 (Nov., 1963), 246-47. Fox, Raymond B. "Dearth of Research on Motivation," The Clearing House, XXXV, NO. 4 (1960), 239-43. McClelland, David C., Atkinson, John W., Clark, Russell A., and Lowell, ,Edgar L. The Achievement Motive. New York: Appleton-Century- Crofts, 1953. McQuary, John P. "Some Relationships Between Non-Intellectual Charac- teristics and Achievement," The Journal of Educational Psychology, XLIV, No. 4 (April, 1953), 215-28. McQuitty, Louis L. "Capabilities and Improvements of Linkage Analysis as a Clustering Method," Educational and Psychological Measure- ment, XXIV, No. 3 (Fall, 1964), 441-56. .. "Elementary Factor Analysis," Psychological Reports, 1x (Nov., 1961), 71-77. "Elementary Linkage Analysis for Isolating Orthogonal and Oblique Types and Typal Relevancies," Educational and Psy- chological Measurement, XVII, No. 2 (Summer, 1957), 207-29. 84 85 . Personal conversations, June 25 and July 8, Michigan Michael, State University, East Lansing, Michigan. William B., Jones, RQAs, and Trembly, W.A. "The Factored Dimensions of a Measure of Motivation for College Students," Educational and Psychological Measurement, XIX (1959), 667-71. Middleton, George and Guthrie, George M. "Personality Syndromes and Academic Achievement," The Journal of Educational Psychology, L, No. 2 (April, 1959), 66-69. Mitchell, James V., Jr. "An Analysis of the Factorial Dimensions of the Achievement Motivation Construct," The Journal of Educa- tional ngchology, LII, No. 4 (August, 1961), 179-87. . "Goal-Setting Behavior as a Function of Self-Accep- Morgan, tance, Over- and Underachievement, and Related Personality Variables," The Journal of Educational Psychology, L, No. 3 (June, 1959), 95-104. H.H. "A Psychometric Comparison of Achieving and Nonachieving College Students of High Ability," Journal of Consulting Psy- chology, XVI (1952), 292-98. Payne, David A. -A Dimension Analysis of the Academic Self-Concepts of Eleventh Grade Under- and Over-Achieving Students. Ph.D. Dissertation, Michigan State University, 1961. Payne, David and Farquhar,.William W. "The Dimensions of an Objective ‘Measure of Academic Self-Concept," Journal of Educational ng- chology, LIII, No. 4 (1962), 187-92. Stone, LeRoy A. and Foster, James M. "Academic Achievement as a Func- Taylor, Thorpe, tion of Psychological Needs," The Personnel and Guidance Journal, XLIII (1964), 52-56. Ronald. "Personality Traits and Discrepant Achievement," Journal of Counseling Psychology, XI (Spring, 1964), 76-82. Marion Dennis. The Factored Dimensions of an Objective Inven- tory of Academic Motivation Based on Eleventh Grade Male Over- and Under-Achievers. Ph.D. Dissertation, Michigan State University, 1961. Weiss, Peter, wertheimer, Michael and Groesbeck, Byron. "Achievement Motivation, Academic Aptitude, and College Grades," Educational and Psychological Measurement, XIX, No. 4 (1959), 663-66. Weiss, Robert Frank. "ASpirations and Expectations: A Dimensional Worell, Analysis," The Journal of Social Psychology, LIII (April, 1961), 249-54. Leonard. "Level of Aspiration and Academic Success,” The Journal of Educational Psychology, L, No. 2 (April, 1959), 47-54. APPENDIX A INTERNAL CONSISTENCY 86 mouoom “cosmopwm poucommudou mHHoo Cannes : on on mm» o: mom on. o: 0: on o: mom wok o: o: :cH NmH mofi mmH mmH HoH :NH mna on 03H Rafi mmH MNH moH on mm» mm» mm» mm» mom was mph mom no» mom no» no» mo» mmH BAH mmH omH HRH pea mam mom an mmH an mod mom HmH. omH mmH moa moH HmH moa mom mom aha mda :nH mmH mom moH >Hx HHHx HHX Hx x xH HHH> HH> H> > >H HHH HH H monoummmcoo HmcuoucH msouH HH¢.mo cOAucwoOmm<1cmoz monoumwmcoo HmcuoucH maouH oocououom mo coaumHUOmm<.cmoz commemoOmmm noumsfio noucH cwoz amonwem nocasz mouwsau MMHMDAU m0 .520 we: mozmmmmmm Efiomzoo AC “wozmemHmzoo AczmmezH mmBmDAU OZHZHEmWHmQ mom mmmDQmoomm 039 mo_wmHHanocoudaou on» :H com: maouH on» moumoncH s .mopoon undemoumm mucomoudmn mHHoo :HnuHs : HsH.m sweeps» H.m sapaoaaH MMHmDHU mom ZMMHfimm moaezHa nz< .mezaa .moneaHoomm< zmHH ¢.m MHmmH .3 Hindu mp 12 2 m H a x53 3.3 HmH "OH MOH than Haoouaaoom £33 .coHpmHoonnHH one: m H N seem sopH Hon an on psoauamooup<_uo seam emH HOH as ONH NOH me as” mm: mH Od kme «OH FHHHHHHHHHHHHHflHHHHHHHHHHHHHHHHHHHHHHH" HHH anemone mom zmmeeam‘ moaezHH n24 .mezae .szHHaHOowma zmHH m.m MHm MMHODHU mom zmeH seemeHo mom zemeeam moaszH n24 .mezaa .monH MMHODHU mom HAMMER/PH MUEHH 92¢ OMZHDH .mZOHHOOHUOOmO. HAM—EH N.O OHOOH 93 HN OH ON OH O OH O N O OH. O HH saOOAEOOH OOH NOH OOH HON OON NOH OON OON OON OOH OON HON “HOH HOOOOOHOOO SUNS .GOHumH 00m m< cmmz HN OH ON OH. N NH O O O OH O NH same sOOH NOOO NONO OHOO NON OOOO NOON OHOO OOOO NNOO OONO NOOO OONO OsoHuOHuoOOO Oo sum "LII Ali ONH wmw ONH ONH ONH OOH OOH ONH NOH ONH ONH OO ONH OOH OOH NOH OOH OOH OOH OOH OOH OOH OOH OHH OOH OOH OOH HOH OOH ONH ONH OOH ONH OOH NOH NHH ONH OOH OOH OOH ONH OOH OOH OOH HOH OOH HOH OOH ONH NOH HOH OOH OOH OOH NOH OOH OOH OOH OOH OO ONH OOH OOH ONH OOH OOH OOH OOH OOH OOH OOH OO OOH OOH ONH OOH OOH OOH NON OHN OOH OON OON NO OOH OOH ONH OOH OOH OOH NON NON OOH NOH OOH mO ONH OOH OOH OOH OOH OOH OHN NON NOH OON OON NO NOH OOH ONH HOH OOH OOH OOH OOH NOH OOH OOH HO ONH OOH OOH OOH OOH OOH OON NOH OON OOH NOH ON mwm. OOH NOH HOH OOH OOH OON OOH OON OOH NOH ON OOH OOH OOH HOH OOH wa_ HOH OOH HOH OOH NOH NOH ON NOH NOH NNH OOH OOH HON OON OOH HOH OON HOH OOH ON OOH ONH HNH NOH ONH ONH, NOH ONH OOH NNH OOH ONH OO OOH OOH ONH NOH OOH ONH OOH HOH OOH ONH OOH ONH OO NOH ONH NOH ONH OOH NOH ONH OOH NNH NNH NNH OOH OO OOH OOH OOH NOH NOH NOH OOH NOH OOH OOH OOH OOH HO OOH OOH OOH OOH OOH NOH NOH OOH OOH OOH NOH OOH NO OOH OOH ONH ONH OOH NOH NOH NOH OOH ONH OOH NOH NO OOH OOH NOH OOH OOH OOH OOH OOH OOH OOH OOH OOH ON ONH OOH NNH OOH OOH ONH OOH NOH OOH NOH OOH HOH ON HOH OOH OOH OOH OOH NOH NOH OOH OOH HOH OOH OOH OH ONH OOH OOH ONH ONH OOH ONH NNH NNH OOH ONH HNH O OOH HON NOH OOH OON OOH OOH OOH OOH NOH OON NOH N OOH OOH OOH .mmw .mww OOH OON OON OON NOH NOW. OOH. OO OOH OOH OOH OON NON OOH OHN OON HHN OOH HHN OHN OO OO OHH NHH OOH OO OO NO MO NO HO ON ON .u.coo N.m mgm<9 94 69 (002) 18 761 1°) 7,8 00 O OOH. NO 99 5? SE/ 0 3x1 012 91 (b 661 N OO ONH OOH NOON“u .J ». I Am OO OHN OO ANNHNM N HON O ONHO OH ? 3 HO (9 v. 0 i; 9L 861‘ ‘z4égfip (3) 92 (902') a [II (:60, j 961 08 .u.aoo N.m mqde. 95 OHH dquHOH 3.3.30 NO O H N OOOO amuH OOH ONH ONH OHOO HmoouOHoOO :33 :oHumNoomm< :32 O H N OOOO amOH OOO OOO ON 36338236 2:5 NOH OHM OHH NOH In. ONH HOH OOH ONH NO OHH «HOH ONO <§K§T7‘zz OH OOOOOOO OOO ZOOOHOO OOOOZHO Oz< .OOZOO .OZOHOOHOOOOO zOeH O . m MHO OOOOOOO OOO ZOOHHOO OOOMZHO Oz< .OOZOO .OonHOHOOOOO zOOH O . m mHmOfi. 96 V T... T. OOH NONHO n w mm m I 9 C... < r. /\ OHH AdudeHH dfldfl< ONH AQVO Oj HOH T. 9 9 S co ,1. m u HOH OH O NH O N HH O N OH O O O O .OH Ogaqmgamaa OOH OOH OOH ONH OOH OHOH OOH OOH OOH OOH ONH NOH NOH OHOO HuoauOHomO . 5.53 coauflnoommOO cam: O HH O N N NH O O OH O H O. OH OOOO OOOH NOOH OOOH OOOH OOOH OOOH NONH OOOH NHOH OHOH OOOH OOOH anOquooOOO.Oo OOOO OOH HOH OOH OOH NOH HOH OOH OOH OOH OOH NOH OOH NOH mum. OOH OOH OOH OOH OOH OOH OOH .11: HNH NOH OOH mmw OOH ONH OOH Own. OOH OOH ONH OOH OOH HNH OOH OOH OOH OOH OOH OOH ONH OOH OMM.ONH OOH OOH NOH OOH OOH OOH NOH OOH NOH OOH HOH HOH NNH HOH NOH OOH OOH OOH III OOH OOH HOH OOH NOH OOH OOH OHH NOH OOH NOH OOH OOH OOH OOH OOH OOO OOH HOH OOH HHH OOH OOH OOH OOH NOH OOH OOH OOH OOH OOH OOH HOH OOH OOH NOH ONH OOH OOH HOH OOH OOH NOH OOH OOH OOH OOH mmw OOH OOH HMO NOH OOH OOH mmw NOH III ONH OOO ONH OO OOH OOH OOH ONH OOH NOH OOH OOH OOH ONH ONH OOH OO NOH OOH OOH OOH HOH OOH HOH OOH OOH OOH ONH OOH OO OOH OOH OOH NOH HOH OOH OOH HOH OOH ONH OOH OOH OH HOH OOH ONH OONH NNH OHH HHH OOH OOH. OOO OOO OO OH x OOHOOOO OOO zOOOOOO OOOOzHO.Oz< .OOZOO .OZOHOOHOOOOO zOHH OH.O OOOOH 97 OOH .336 814%qu OOH ) I I. 0 CK OHH A muN a NOH 981 NONHA NNH OOH .L m .6 U OWN 6 NO OOH a Ow H C. to N H O OH 3 O OH NH N O O O HH OH Kama amuH NNH OOH ONH NOH ONH OOH OOH NOH OOH HNH ONH ONH OOH NOH “Ham HwoouOHoom fiuHS coHumH00mm< cam: O H O OH O O OH HH N O N O NH 3H xamm EouH IOOOOINHNNENOON OOOH-OmHWOOOWHOOm-MOON OOHOOOHNHOOOMOON OOOHOOOO 2633682 O6 25 Ill. ONH MMM. OOH DOM. ANH OOH OOH .ONH OOH OOH OOH OOH pmw OOH ONH NOH OOH ONH HOH OOH OOH OOH HNH ONH ONH mmw. OOH OOH OOH NOH OOH OOH OOH OOH HOH NOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH NOH OOH OOH OOH OOH OOH WWW OOH NOH ONH ONH ONH OOH OOH . ONH OOH WNW ONH mmm. OOH :NH #OH NOH ONH HNH HOH OOH NOH ONH NOH ONH. OOH ONH ONH OOH OOH .OOH OHH OOH OOH OOH OOH OOH NOH NOH HNH OOH OOH OOH OOH OOH OHH OOH WNW HOH WNW OOH pmw WWW OOH OOH mum. OOH OOH HOH OOH NOH OOH NOH OOH ONH OOH HNH OOH ONH NNH OOH. OOH OOH NOH OOH HNH OOH OOH HOH ONH OOH OOH ONH OOH NOH OOH NOH NOH OOH ONH OOH OOH OOH ONH OOH OOH NNH OOH OOH NOH .NOH OOH OOH ONH OOH OOH ONH OOH OOH OOH OOH NOH OOH OOH NOH OO OOH NOH OOH OOH :OH OOH OOH OOH OOH OOH NOH OOH 33H NO OOH OOH OOHHWMOH NOH OOH OOH HOH OOH NOH NdH NOH 33H :N OOH OOOH OOH ONH OONH «OHH OHH OOH ONOH NOH OOH O Hx OOOOOOO OOO ZOOHHOO OOOOzHH.Oz< .OOZOO .OonOOHOOOOO zOHH . HH.O OOOOH 98 9ZI ZLI OOH A NOOHuv NNH 691” 6ZI O H O O gnaw aouH ONH ONH OOH OOH “Hum HaoouOHOOm H.323» gwuww 00mm< flwmz N H O m xcmm aopH OOO OHO OOO OOO OnoHuaHooOm<.Oo masm NNH OOH POM. OOH ONH OOH «OH- NNH OOH OOH HOH ONH OOH NOH HOH OOH «OOH ¥NNH ONH ONH HHx MMHODHU mom ZMMHHOrm O.~w non.n. 414 3 CO5 VOO OOOJ vm 1 3 1 O. “HOONVMO “HNONVOO EOOACONNVOOOPKHNON ON “HOON mOmmHOMW Ow m\/ 2 h». 141 K m5) OOO aHOONv Iv 43 18 <( 7)9 (167) 7 103 1.13.9.1) 57 45 001 147 52 101 121 (L91) OOH HOOHOv OOH nOOHuw OHH HHOO: OHH wOOHO OOH ‘ ) \II T. T. 9 I. HO nu I 9 Co I. O O O H O O O O OOOO aOOH OOH OOH OOH OOH OOH OOH OOH OOH OOOO HOOOOOOuOm OOOs OOOOOOOOOOO OOO: O O O H O O O O OOOO aOOH OOHH OOOH OOOH OOOH OOHH OOHH OOOH OOHH .mmOOOOOOOOOO Oo asO nun. OOH OOH. MOO. OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOH OOO OOH OOH OOH OOO OOH O OOH OOH OOH OOH OOH OOH OOH OHH OOH OOH OOH OOH OOH HOH OOH OOH OOH OOH OOH OOH HMO OOH MOO. HOH OOH OOH OOH OOH HOH OOH OOH OHH OOH OOH OOH OOH OOH OOH OOH OOH OOOH OONH O OOHH HNH OOHH >Hx mmHmDHU mom OZMMHH HH>. HO. > > HHH OH H .oz OOOOOHO 989mg Agmgmg ZOHHDHOm mun—MO 9200mm E mom mZmEV on O>HO A>HO AOOOHOVOO z umumsHo v3.5 35 NdH 00 NH _flu “u 8 I. Q, T. ij /O AHHH>V AHW>V nu NOON“ cm TL 6 6 O>Hxv NHH Noaauv mHH A>HNV AHHV AHHHNV mw AHHHV ANHHNmV NH ANOONV o mmNHv m0 m“ a“ U U. Sac v < NNH ”NOH“ mma OH HOH Sc 35 2 “mumsau . .uHGOU N.U m4m<fi APPENDIX D THE COMPREHENSIVE HI ERARCHIAL APPROACH 106 .csaAoo some GA ououu ucusuwuwu unoch; «so unencumou Ononsac vocAAOovcb .aoOOun ucuauuuwc auauuounwu OAAou cAzuAa c NA «A 0N NN NN ON RA NA R NN NN RN AN 0A 0A @N RN mN ON AN 0 N CA AA 0 A wN NA NN 0N 0A ¢N ON 0Q Q m N N NN ON scum EOuA 99s999999sgsg99§g§SS9 99 9S 7. 7» 6 0 0 O I z r. V C— 8 0 I T. h! 0 I. C. (L w r). (c C. 9 C. C. 9 9 9! c 9 9 C. C. ”COAUQNUOwg MO OOOOuOuOOOOOOanmuOOOOOmmummmmmmmmmmmmmm 55.33 00A NNA NNA NRA RRA NRA omA ooA AQA NNA NNA GOA woA NNA ANA NAA OAA ANA NQA owA RNA GOA NoA owA ooA ONA NRA NNA NNA NRA NNA ow o¢ NoA RNA NOA 00A ANA wNA NAA NNA owA mwA NRA NRA NRA NQA RwA NNA ANA QQA NoA RoA ooA NNA 00A NAA NNA §¢A NRA owA omA woA AGA woA NNA NRA NNA ONA NoA 00A RR QN 00A #oA woA NoA ORA ONA ONA 00A mwA NRA woA NRA NOA NNA omA NNA NNA RNA NNA QNA RNA NNA OOA ¢AA NQA NNA QRA RRA NRA mmA owA AoA oNA RoA NNA NNA wNA NoA mR aN RwA NNA AoA 00A 00A 00A NAA NNA mwA ORA oHx HHAX HHX Hx x xA HAA> AA> A> > >H AAA HA A uvnfldz heunnAU mzoAHHx 35H NmH omH mm“. mnH amfi amH moH omH amH cu omH ooH HHHx Rafi NmH .ul. mofi HmH mm“. mmH an mmH omfi HaH ONH :nH mnfi HHx mmH omfi moH 03H moH on HoH mmH mmH mmfl pmw mcH Hmfl Hx oHH mNH HmH mmw mafl amfi aHH omH mmH mmH mmH add mnfi x 33H mmH Hofi on maH mafi mmw mm“. mNH Rafi HHH mafi osfl xu aka mmw mmH :0H :NH mafi 1:: new aha mmH mcfl oh mmm.wmw HHH> mmfi awH mmH Hog aHH mag mom NNH mmw. HRH. aw mom moH HH> mmfi mofl mmH amH oMH HmH as” NNH mafi aha am ONH aaa H> Had end aNH mmH mm“ aNH mmfi nmfl mzfi onfi “HA smH mmH > .moH amfi Aafi mmfi mmH 53H moH HNH aBH amH um nNH mmH >H phi. on ONH ooH mnH HHH om mm: mm NHH mm :o mpw HHH omH omfi :mH mofi OHH maH oaH no“ oNH. an ”NA so mcfi HH Hmfi oofi mmfl HmH mam oafl Hofl mcH 03H mmfi mm”. NOH mofi H >Hx HAHN HHx Hx Kn NH HHH> HH> H> >2 >H. HHH HH H umnssz umumsAo u 1! mzmmhfi¢m mO