ABSTRACT LEVEL OF OCCUPATIONAL ASPIRATION, PERFORMANCE IN COLLEGE, AND FACILITATION: A PRELIMINARY TEST OF CERTAIN POSTULATES CONCERNING THE RELATIONSHIP BETWEEN ATTITUDES AND BEHAVIOR by Anthony J. Diekema Attitudes have been a central topic for many behav- ioral scientists during the past half-century. Perhaps no other concept used in the behavioral sciences has been applied to a greater variety of investigations of human behavior. Nevertheless, a review of the relevant literature makes it quite apparent that the general relations between attitudes and behavior have not yet been adequately devel— oped. Concern with the relationship between attitudes and behavior thus maintains a rather prominent place in the behavioral science literature today. This thesis develops three postulates concerning the relationship between attitudes and behavior. The adequacy of these postulates is ‘then partially determined by the testing of three derived general hypotheses and seventeen sub—hypotheses. The postulates and the derived hypotheses center on the fundamental problem of gaining insights into the nature of the relationship between attitudes and Anthony J. Diekema behavior when certain facilitation variables are considered and controlled. A theoretical framework from which the postulates are developed is presented. This framework is based upon a review of the literature, both theoretical and empirical, dealing with the concept ”attitude.” A conceptual framework is also formulated from the theoretical position which is then translated into the operational setting of the study. Throughout the thesis, the relationship between the theoret- ical framework and the Operational setting is emphasized. Operationally, the thesis is concerned with the nature of the relationship between level of occupational aSpiration (attitude variable) and performance in college (behavior variable) within the context of several controlled facilitation variables, both extra— and intra—personal. The facilitation variables which are controlled are: (l) per- ceived reference group support, (2) socioeconomic status, (3) mental ability, (4) perceived necessity of college, and (5) past academic performance. The subjects are 585 fresh— men students who entered Michigan State University in the Fall of 1960. The predictions are tested by a correlational anal- ysis of the data obtained through the administration of several instruments, including A. O. Haller‘s Occupational ASpiration Scale for measurement of the attitude variable. Anthony J. Diekema The predictions are, on the whole, supported. The three derived general hypotheses are tested, supported and con— sequently accepted as tenable. Twelve of the sub—hypotheses are accepted while five are rejected. Thus, the postulates are tentatively accepted as tenable on the basis of the data obtained in this operational setting. The findings of the study suggest several theoret— ical and methodological refinements for future research and the need for further testing of the postulates in other operational settings. Finally, some practical implications of the study are discussed. LEVEL OF OCCUPATIONAL ASPIRATION, PERFORMANCE IN COLLEGE, AND FACILITATION: A PRELIMINARY TEST OF CERTAIN POSTULATES CONCERNING THE RELATIONSHIP BETWEEN ATTITUDES AND BEHAVIOR By Anthony J. Diekema A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Sociology 1965 dedicated to JEANE AND THE THREE D‘s ii ACKNOWLEDGEMENTS The completion of a dissertation requires the time and labors of many persons, only some of whom can be appro- priately and publicly acknowledged. This opportunity is taken to express appreciation and gratitude to several whose assistance was most directly felt. My greatest appreciation and indebtedness is to Dr. Archibald O. Haller who from the beginning gave assistance far above and beyond the normal call of duty. As chairman of my guidance committee, major professor, constant motivator, rigorous critic, and true friend, he was indeed an indispensable resource to whom I shall always be grateful. Also, to the other members of my guidance committee, Drs. Donald W. Olmsted, Charles P. Loomis, and John X. Jamrich, my most sincere thanks are ex— tended. These men were always ready and willing to assist in all phases of the planning and completion of the grad- uate program and this dissertation. For assistance in the collection and analysis of data, I also wish to thank the Office of Evaluation Services and the Computer Laboratory at Michigan State University. Completion of the doctoral dissertation represents a sort of milestone in one's intellectual development. For this reason appropriate acknowledgement should extend far beyond the limits of the dissertation. Although it is in— deed difficult to trace clearly one’s intellectual develop- ment and to evaluate the contributions made to it by one's many teachers, colleagues and friends, the contributions of several tend to be felt most directly. To Dr. Donald H. Bouma, now at Western Michigan University, I shall always be grateful for a most stimulating introduction to the disci- plines of sociology and anthropology during my undergraduate years at Calvin College; an introduction which certainly was most influential in my development. Also to Dr. Leo F. Schnore, now at the University of Wisconsin, and to Dr. William H. Form, I owe much for many insights gained during iii the early phases of the graduate program. Dr. Gordon A. Sabine, vice president and former dean of the College of Communication Arts at Michigan State University, must also be given a note of appreciation for supplying continued support and motivation during critical phases of the doc- toral program. For this, and much more, I shall always be grateful to him. There are many others at Michigan State University, both past and present, who in ways too numerous to mention have contributed to my graduate training. To them also I extend my sincere gratitude. Finally, I shall forever owe an overwhelming debt of gratitude to my family. My wife, Jeane, was a constant aid. Throughout the doctoral program her tolerance, understanding and love served to hold me "on course” and provided a con- tinual source of encouragement. To the three D's, Doug, Dave and Dan, who so often unknowingly played the role of the "hidden persuaders" and whose actions constantly chal- lenged my knowledge of human behavior, I hope I shall some day be able to adequately express my appreciation. To my mother and father, who from the start have provided encour— agement and support, I am also deeply grateful. iv TABLE OF CONTENTS Chapter Page I. INTRODUCTION . . . . . . . . . . . . . . . . . l The Purpose 1 The Problem 1 Importance of the Problem 2 The Contributions 4 The Limitations 6 The Scope 7 Organization of the Thesis 7 II. THEORETICAL FOUNDATION AND CONCEPTUAL BACK— GROUND OF THE POSTULATES . . . . . . . . . . 9 Introduction . . . . . . . . . . . . . . 9 Glossary of Terms . . . . . . . . . . 10 Theoretical Foundation and Conceptualizations . . . . . . . . . . . 15 The Problem . . . . . . . . . . . . . . . 32 The Variables . . . . . . . . . . . . . . 35 The Postulates . . . . . . . . . . . . . . 35 III. RESEARCH DESIGN . . . . . . . . . . . . . . . 38 The Operational Framework . . . . . . . . 38 The Attitude Variable . . . . . . . . 38 The Facilitation Variables . . . . . . 44 Perceived Reference Group Support . . . . - . . . . 47 Socioeconomic Status . . . . . . . 50 Mental Ability . . . 52 Perception of Importance of Means Behavior (Necessity of College). 53 Past Academic Performance (Self Concept of Ability) . . . . . . 54 The Behavioral Variable . . . . . . . 56 Chapter Page Measurement . . . . . . . . . . . . . . . 57 The Attitude Variable . . . . . . . . 58 The Facilitation Variables . . . . . . 6O Perceived Reference Group Support . . . . . . . . . 6O Socioeconomic Status . . . . . . . 65 Mental Ability . . . 67 Perception of Importance of Means Behavior (Necessity of College). 69 Past Academic Performance (Self Concept of Ability) . . . . . . 70 General Facilitation . . . . . . . 71 The Behavioral Variable . . . . . . . 73 The Hypotheses . . . . . . . . . . . . . . 75 The Sample . . . . . . . . . . . . . . . . 81 Limitations of the Sample . . . . . . 85 Summary of the Chapter . . . . . . . . . . 90 IV. THE RESULTS . . . . . . . . . . . . . . . . . 91 Introduction . . . . . . . . . . . 91 Tests of the Hypotheses . . . . . . . . . 92 Implications of Acceptance or Rejection . . . . . . . . . . . . . 93 Hypothesis 1 . . . . . . . . . . . . . 96 Hypothesis 2 . . . . . . U . , . . . . 114 Hypothesis 3 . . . . . . . . . . . . . 129 Summary of Results . . . . . . . 134 Summary of Behavioral Consequences . . . . 142 V. SUMMARY AND CONCLUSIONS . . . . . . . . . . . 149 Summary . . 149 Limitations of the Study and Suggestions for Future Research . . . . . . . . . 160 Theoretical Contributions . . . . . . . . 166 Practical Implications . . . . . . . . . . 170 REFERENCES . . . . . . . . . . . . . . . . . . . . . . 174 vi Appendix A. B C. D LIST OF APPENDICES Occupational Aspiration Scale and Code Data Card Format and Code About My College Plans Questionnaire Correlation Matrix of Attitude, Facilitation and Behavior Variables General Facilitation Index General Perceived Reference Group Support Index Distribution of Sample on Attitude and Facilitation Variables . . . . Mean Relationships of Attitude, Facilitation and Behavior Variables Behavioral Consequences of Combinations of Degrees of Occupational ASpiration and General Facilitation Selected Chi-Square Tests Cutting Points for Attitude and Facilitation Variables . . . . vii Page 180 187 201 204 206 209 212 216 219 222 225 Table LIST OF TABLES Comparison of Sample and Total Population on Certain Demographic Characteristics Summary of Tests of Hypotheses Mean Level of Performance for Major Combined Categories of Occupational ASpiration and General Facilitation Percent of Dropout and Non-Dropout for Major Combined Categories of Occupational ASpiration and General Facilitation viii Page 83 135 143 145 Figure LIST OF FIGURES Page The diagram portrays the general theoretical and conceptual framework within which the research is formulated. The concern is with the prediction of differential means behavior on the basis of differential attitudes within a situational framework. It is assumed that the various social structural variables have all had an impact upon the attitude measured, as well as upon other attitudes and the other personality variables, and that this impact is reflected in the attitude. It is further assumed that the environmental variables (attitude context variables) selected and controlled in this research are some of those most significant in their affect upon behavior . . . . . . . . . . . . . . . . . . . 33 The diagram portrays the basic theoretical and conceptual framework within which the research is formulated. It constitutes a part of the larger framework presented in Figure l, and is that part upon which the present research focuses . . . . . . . . . . . 46 ix CHAPTER I INTRODUCTION The Purpose The purpose of this thesis is to develop postulates about the relationship between attitudes and behavior, and to test Specific hypotheses derived from the postulates. The Problem The problem centers around the basic question, ”What is the nature of the relationship between attitudes and behaviors when certain situational or environmental variables are considered and controlled?” A theoretical framework is developed from which several postulates are obtained and stated. The operational question upon which the specific hypotheses guiding this research are based is, ”What is the .nature of the relationship between level of occupational aSpiration and the level of performance in college when cer- tain facilitation variables and varying conditions of facil— itation are considered and controlled?” Within the opera- tional framework, level of occupational aSpiration consti— tutes the attitude variable and level of performance in college is the behavior variable. Selected intellectual and non—intellectual variables are measured and controlled, and these constitute the situational or environmental variables, both intra— and extra—personal, considered to be important when the individual attempts to enact an attitude. Specific hypotheses derived from the operational question are stated within the broader theoretical framework and tested to determine their adequacy. Importance of the Problem The problem is important for several Specific rea- sons. First, it is a problem which, if and when resolved, will contribute considerably to a more Specific and con— sensual conceptualization of the term ”attitude” and more consensus regarding its overt behavioral consequences under varying conditions of the actor within a social, psycholog- ical and physical environment. If so, it will also contrib— ute considerably to our present knowledge about human behav- ior. Second, the theoretical framework used may, if Sup— ported, provide new insights to the even broader concern with the development of a unified theory of social psychol- ogy and with the problem of causality. Third, the method may isolate some types of variables, and the relations between them, which could be applicable to future theory building and research in a variety of areas. Finally, the intensive investigation of the problem should lead to new knowledge regarding the predictability of behavior as well as concerning the relationship of certain socio—psycholog— ical factors to behavior and to each other. It Should also in time contribute appreciably to the theoretical purpose of increasing the generality of behavior theories. Current attempts to resolve some of the problems regarding the relationship of attitudes to behavior often ' are neither consistent nor systematic. Perhaps because of this many thoughtful social psychologists, and social sci— entists generally, declare the problems to be unresolvable, or at least unfruitful both empirically and theoretically. Unfortunately, in the writer's opinion, this general orienta— tion to the problems has impeded progress toward a more valid explanation of human behavior. From the ”applied” standpoint, the problem is also important. Interest in the predictive value of ”non—intel- lectual” factors in the academic achievement of college stu- dents has recently increased. The data used here, and the potential predictive value of the theoretical model of multi— variate analysis, will make a practical contribution to this problem possible. It is felt that occupational aspiration is Significantly related to performance in college. It is quite conceivable that this is so. Thus, if it is so, this research should enhance our present ability to predict differential success or failure in college. The research conducted here is concerned principally with the differential behavior of the actor as it results from his attitudes under various conditions offered by the environment of the attitude. The selection and use of var— iables is guided by the assumption that the prediction of behavior is possible, and that it is multivariate in nature. This research is not expected to resolve the problem, but it is expected to test initially the usefulness and adequacy of the theoretical framework guiding it, and to determine its worthiness for future research. If the theoretical frame— work is supported, more intensive studies with more rigid controls, particularly of the Situation within which the attitude is carried into behavior (facilitation variables), should be completed in various operational settings. The Contributions Several contributions of this study may be isolated. First, it provides initial support, although tentative, for the utility of the theoretical framework which guided it and which is purported to be general. Theoretically, it suggests that differential behaviors toward one type of social object (a "means object") may be predicted on the basis of differen— tial attitudes toward the more distant object (for which the former is a means) and differential facilitation. In the operational setting of this study it was shown that differen— tial college performance (means behavior) may be predicted on the basis of differential level of occupational aSpira- tion (attitude toward the more distant object—-the occupa— tional heirarchy) and differential faciliatation for perform- ance in college (means structure).1 The evidence suggests that the theoretical framework may be extended profitably to other situations and operational settings toward which it has logical relevance. Second, this study emphasizes and apparently sub- stantiates the critical role of the attitude context (the extra— and intra-personal situation in which the attitude is enacted in behavior) in the attitude—behavior relationship. It supports the notion that the reasons overt behaviors are not always consistent with attitudes are to be found by the intensive investigation and control of these facilitation variables. Methodologically, the study also provides some possible insights into the development of new methods for combining attitude and facilitation variables for optimum prediction of corresponding overt behavior. Third, the research provides empirical support for the idea that there exists a means—ends relationship between the educational and occupational structures in American 1These concepts are fully defined and discussed in Chapter II. society. Thus, it provides indirect support for the notion that persons' positions in one structure differentially determine or allocate their positions in the other. As sug- gested earlier, then, it may be concluded tentatively that where such a means—ends relationship does exist between two or more objects, overt action behavior toward any one may be differentially predicted on the basis of attitudes (used together with facilitation variables) regarding any one. The Limitations The limitations of this study lie in two general Spheres. First, the sample imposed several problems by its relatively limited range of variability on several of the key independent variables. The possible implications of this limitation for the tests of significance employed and the observed correlation coefficients cannot be ignored and, thus, they are discussed at various appropriate points in the thesis. Second, the research design posed several un- anticipated problems in respect to the control and measure— ment of variables. These problems are also cited at appro- priate points in the thesis and are isolated and discussed in the final chapter. The Scope Technically, the findings of this study are limited to the 585 freshmen entering Michigan State University in the Fall of 1960 who made up the sample used in this re— search. In all probability, however, generalizations may be made to the total freshman class which entered Michigan State University in the Fall of 1960. Finally, assuming that the data used here in the test of a theoretical frame- work which is purported to be general has all of the proper parameters, the conclusions may be tentatively extended to any situation in which the theoretical framework may be logically employed. Organization of the Thesis The thesis contains five chapters. Following this brief introductory chapter, the larger theoretical and con— ceptual framework is presented in Chapter II. It also pro- vides a general statement of the problem and presents the postulates which constitute the focus of this study. Chap- ter III discusses the operational framework and the research design. In it the theoretical and conceptual framework is applied to the specific problem and the research design is presented. The sample and the variables are described. Finally, the relationship of the hypotheses to the postulates is shown and the former are presented in testable form. In Chapter IV the tests of the hypotheses are presented. Chap— ter V summarizes and evaluates the empirical findings in view of both the theoretical and the operational frameworks of the study. Several limitations of the present study are discussed and suggestions for future research are presented. CHAPTER II THEORETICAL FOUNDATION AND CONCEPTUAL BACKGROUND OF THE POSTULATES This chapter sets forth the larger theoretical framework within which the general problem of research is formulated, the types of variables are stated, and several postulates are constructed for later use in the structuring of the Specific problem and testable hypotheses. Introduction It is felt that behavior may be predicted on the basis of attitudes. It is also recognized that in attempt- ing to predict behavior on the basis of Specific factors one inevitably must become involved in the problem of causation.1 However, causal relationships in the field of social psychol— ogy, or social science generally, are difficult to establish rigorously. At the present time we can often only say that there is a certain pattern of association between two or more phenomena. Usually, the best we can do is assert that 1For a thorough discussion of the problem of causa— tion, see Roy G. Francis, The Rhetoric of Science (Minneap— olis: The University of Minnesota Press, 1961), Chapter 3 and 7. 10 certain types of phenomena regularly precede others in a temporal sequence, for example, that certain types of attitudes regularly precede certain types of behavior. Although we cannot rigorously prove causality, we may build a theoretical model of multiple causation which will assist in isolating many variables, which when used together may be more highly predictive of human behavior. Assuming, then, that causal or explanatory relations in social psychology are primarily patterns of association between phenomena and that the designata of the social sci- ences are multivariate in nature, we must proceed with a theoretical model and research design which reflects these assumptions if a contribution to this field of knowledge is desired. Glossary of Terms The purpose of this section of the thesis is to present a preliminary definition of the key terms used here and thus facilitate communication with the reader. These terms are developed further, and related to each other where appropriate, in the following section. The definitions are not necessarily intended to improve upon other definitions found elsewhere in the literature. Behavior. This concept is narrowly defined as attitudinally— directed action of individuals. It is not broadly-defined 11 to include all changes in the state of the organism, such as reflexes, Spasms, etc. It is used to denote that Specific area of behavior often called ”action,” that is, the goal— directed or attitudinally-directed behavior of the individ- ual. Differential Behavior. This concept refers to the possible variation among actors in behavior with reSpect to any object. It implies that alternative behaviors are possible with respect to an object. Object Behavior. This term is defined as the individual‘s behavior with respect to the object of an attitude. Means Behavior. This term is defined as the individual's behavior with reSpect to Something which is only a means to achieving some more distant end, that is, a means to even- tually behaving with reSpect to the object of an attitude. Orientations to Behavior. This concept refers to the gen- eral prediSposition within an individual to act in a given manner toward a given object. It does not imply that actual overt behavior will occur. It does imply, however, that this predisposition will determine in part the nature of the overt behavior toward an object, if overt behavior does in fact occur. 12 Attitude. This term is defined as an affective orientation to an object. The definition refers to attitudes as Special instances of orientations to behavior. Attitude Context. This term refers to the environment or Situation, both intrapersonal and extrapersonal, in which an attitude is differentially facilitated in its expression in behavior. It thus includes the social (extrapersonal) context, which may in turn include other individuals, norms, and material objects. It also includes the intrapersonal context, which may in turn include any or all aspects of the personality other than the attitude itself, such as ability, self-control, perception, self—concept, and the like. Attitude Constellation. This term is defined as a group of separate attitudes which are perceived by the actor as belong- ing together because they generally are centered around a given activity or object. Attitude Strength. This concept is defined as the degree of positive affect invested in objects by individuals. Hence, attitude strengths may vary between individuals or groups of individuals. Independent Attitude. The term is defined as an attitude which tends in any actor to be independent of other l3 attitudes or attitude constellations, and which may or may not be congruent with them, because it arises out of Spe— cific and possibly unique personal experiences of the actor. Differential Attitudes. This concept refers to attitudes characterized by varying degrees or directions of affect invested in objects by two or more individuals. Facilitation. This term is defined as the assistance or hinderance which the context of the attitude, or situation, provides to the attitude in its expression in appropriate corresponding behavior. It refers to the difficulty of carrying out the behavior alternatives appropriate to the attitude within the attitude context. Differential Facilitation. This concept refers to the vary— ing degrees to which the attitude context permits an atti- tude to be carried into appropriate corresponding behavior. It may be considered the varying degree of difficulty expe— rienced in carrying out the behavior alternatives appropriate to the attitude within the attitude context. Individual,Actor. These terms are defined as a person in active relation to the physical, social and psychological environment in which he is situated. Object. This term is defined as anything individuals cog— nize and define as real. Thus, it may be a tangible thing, 14 such as a person or a book, or an intangible thing, such as ghosts or angels. In the context of this thesis, then, it refers to things towards which individuals may have atti— tudes. Postulate. This term is defined as a general statement of relations assumed to be true on the basis of a logically developed theoretical framework, but without Specific proof. It is an unproved statement of relationship. It is consist- ent with Francis' definition that ”a postulate is an assump- tion which is not tested but is accepted as true.” Theory. This concept refers to a logical and consistent formulation, supported by various research findings, of seemingly apparent relationships which may be stated in the form of principles or postulates adequately defined. Thus, it is in operational terms a type of policy——in Conant‘s words, ”an economical and fruitful guide to action by sci- entific investigators.”3 2Francis, op. cit., p. 19. 3James B. Conant, Modern Science and Modern Man (New York: Doubleday Anchor Books, 1954), p. 97. 15 Theoretical Foundation and Conceptualizations Attitudes have been a central topic of social psycho— logical study for many years.4 Complex methods for measuring them have developed,5 their relationship to personality has been examined,6 techniques for changing them have been dis- covered,7 and theories in which they are but Special in— stances of more general cognitive phenomena have been pro— posed.8 Vast numbers of empirical studies of attitudes have also been published and it is quite apparent that the con- cept ”attitude” continues to hold a rather prominent place 4G. W. Allport, ”Attitudes,” A Handbook of Social Psychology, ed. C. C. Murchison (Worcester, Mass.: Clark University Press, 1935). 5A. L. Edwards, Techniques of Attitude Scale Con— struction (New York: Appleton-Century-Crofts, Inc., 1957). 6T. W. Adorno, et al., The Authoritarian Personality (New York: Harper, 1950); and Milton Rokeach, The Open and Closed Mind (New Yorkzr Basic Books, Inc., 1960)i 7C. I. Hovland, "Changes in Attitude Through Commu- nication,” Journal of Abnormal and Social Psychology, Vol. 46, 1951, pp. 424—437; I. L. Janis, et al., ”Effects of Preparatory Communication on Reactions to a Subsequent News Event," Public Opinion Quarterly, Vol. 15, 1951, pp. 488-518; and I. Sarnoff, et al., "The Motivational Basis of Attitude Change,” Journal of Abnormal and Social Psychology, Vol. 49, 1954, pp. 115-124. 8Fritz Heider, The Psychology of Interpersonal Rela- tions (New York: Wiley & Sons, 1958); Leon Festinger, A Theory of Cognitive Dissonance (Evanston, Illinois: Row, Peterson & Co., 1957); and T. Newcomb, Social Psychology (New York: Dryden Press, 1950). 16 in the contemporary literature of social psychology. Never- theless, it is at least partially true that the general relations between attitudes, behavior, and the context with- in which attitudes are carried into behavior, have not been developed. To the knowledge of this writer, only very re— cently have writers viewed the context within which atti— tudes must be carried into overt behavior as important. De Fleur and Westie Seem to suggest that more intensive inves- tigation of mediating or facilitation variables may explain in part why attitudes are not always consistent with overt action. The lack of a straight—line relationship between verbal attitudes and overt action behavior more likely may be explained in terms of some sort of social involvement of the subject in a system of social constraints, preventing him from acting (overtly) in the direction of his convictions, or otherw$se ”legitimizing” certain behavioral patterns. And more recently in an intensive review of the attitude concept, they contend the following. There are clearly situational factors such as group, norms, roles, definitions of situations, and other social constraints which materially "mediate” reSponses in situations involving either verbal behavior or overt action. Such “mediating” social constraints appear to explain lack of cor- respondence better than such a conception as 9M. L. De Fleur and F. R. Westie, ”Verbal Attitudes and Overt Acts: An Experiment on the Salience of Attitudes,” American Sociological Review, Vol. 23, December, 1958, pp. 667-673. l7 "true attitudes" conceived of as latent variables.10 Thus, it seems that there may be much gained by the theoret— ical development of at least some general postulates concern- ing the relationship between attitudes and behavior, with some systematic concern for the context within which this relationship exists. It is felt that research arising out of such postulates has a number of virtues which could con— tribute greatly to present sociological and social psycholog- ical thinking, both in regard to attitude theory and to more general theories and their verification. In spite of its prominent place in the literature, the concept ”attitude” is not defined uniformly by contem— porary writers.12 In view of the many definitions of atti— tudes which have been formulated over the years, a theoret- ically formal attempt at the definition of an attitude here would be ill—timed. Definitions are matters of convenience, and they attain great precision only in the advanced stages 10M. L. De Fleur and F. R. Westie, ”Attitude as a Scientific Concept," Social Forces, Vol. 42, October, 1963, pp. 17-31. 11T. Parsons and E. A. Shils, Toward a General Theory of Action (Cambridge, Mass.: Harvard University Press, 1951); Robert Merton, Social Theory and Social Structure (Glencoe, Illinois: The Free Press, 1957)? and Leon Festinger, op. cit. 12De Fleur and Westie, 1963, op. cit. 18 of a science. In time, as the science of social psychology advanceS, greater precision will develop. In the meantime it is felt that we may well use the knowledge we presently have in formulating a definition. The concept ”attitude” has often been used in such a way that it has a definite behavior potential associated with it.13 Haller indicates that many earlier definitions of attitude stressed the prob— ability that a behavior will occur, or they stressed a "tendency to behave in a certain way.”14 Recently, some social psychologists and others have come to distinguish Sharply between overt behavior and orientations to behavior. This approach defines attitudes as Special instances of orientations to behavior. Haller sums up the slightly vary— 15 ing forms of the new definitions, such as that of Edwards and Peak,16 by stating that an attitude is an affective l3Allport Speaks of ”a mental and neural state of readiness.” (Allport, op. cit., p. 310). Sherif and Cantril place attitudes ”among those components of the psychological makeup of the individual which determine that he shall react not in a passive or neutral way. . . .” (M. Sherif and H. Cantril, "The Psychology of Attitudes,” Psychological Review, Vol. 52, 1945, p. 300). 14A. O. Haller, ”Some Principles of Attitudes and Behavior," (unpublished working paper) 1957. It should be mentioned that many of the ideas presented in this chapter originated in, or are based upon, this paper. 15Edwards, op. cit., p. 2. 16Helen Peak, ”Attitude and Motivation,” Nebraska Symposium on Motivatiopj 1955, ed. M. R. Jones (Lincoln, Nebraska: The University of Nebraska Press, 1955), pp. 149- 189. 19 orientation to an object. This definition is used in the development of the postulates and the subsequent research presented in this thesis. In terms of behavior, the above definition of atti- tude implies that action may or may not result from an atti— tude.l7 Behavior may be narrowly and simply defined here as action of individuals.18 However, the definition of attitude also implies that alternative behaviors are pos— sible: behaviors following from attitudes will vary between persons, or from time to time in one person according to whether the affect is positive, zero or negative. Since behavior is a central concept in the development of the‘ postulates it must be given definite limitations in research; all behavior alternatives must be isolated. An attitude, then, may result in behavior toward an object. The object of an attitude can, of course, be a tangible thing such as a person or a chair, but it can be 17A basic and essential concept implied throughout this thesis is that of the ”actor.” The concept is broadly defined and used here as a person in active relation to the physical, social and psychological environment in which he is situated. 18Behavior is not used here in the broad sense of constituting any change in the state of the organism. We are concerned with the specific cases of behavior often called ”action,” that is, goal—directed or attitudinally— directed behavior of the individual. 20 more than this. Individuals can have attitudes toward any— thing they cognize: God, ghosts, demons, angels, Cuba and the like. Anything individuals define as real may be the object of an attitude 19 If an attitude may result in behav- ior toward an object, it may also result in behavior toward the means of fulfilling Specific object behavior. Since the latter is central to the problem of this research, it will be discussed further in a later section. The behavioral consequences of an attitude must always occur in Situations and thus attitudes are differen— tially facilitated in their expression in Specific behavior by the situations at hand, or the attitude contexts. The attitude context is the social and intrapersonal situation within which attitudes are carried into behavior. The rel— evant effect of the attitude context is the degree to which it facilitates the expression of the attitude in behavior. Thus far five basic concepts necessary in the devel- opment of the postulates have been presented (attitude, behavior, actor, object, and attitude context). In investi- gating these concepts more closely, it is necessary to de— rive additional concepts for the total development of the conceptual framework. lgMerton, o . cit., p. 42lff. 21 Since any extensive review of the relevant litera— ture will diSplay multiple variations in the interplay be- tween the terms attitude and value,20 the latter must be briefly discussed here. The concept value is related to our use of the concept attitude in two ways. Value may be used to indicate anything which has positive affect for the per— son or it may be used to indicate an orientation which is societally—defined and which holds that a certain behavior or object (or series of behaviors or objects) is inherently good. The latter relation is most applicable to the present concern. Attitudes may be assumed to arise out of either basic cultural values21 or specific individual and value- establishing experiences. These attitudes arising out of the basic cultural values are likely to be distinguishable by their arrangement within specific areas or phases of the individual‘s life style. In short, an individual‘s atti- tudes are likely to be grouped into attitude constellations 2OHenry J. Watts, Methodological Problems in the Measurement of Values (unpublished Ph.D. thesis, Michigan State University, 1962), Chapter III, p. 42ff. 21Cultural values are defined here as Social prod- ucts which members of any culture derive from the accumu- lated action and thought ways of the people responsible for their socialization. They are values based upon a complex of evaluations which constitute a design for living, i.e., as to what is good or bad, proper or improper, valuable or valueless, adequate or inadequate; and thus are accepted and used as criteria of worth. A cultural value is a value real— ity arising out of a cultural reality. 22 or attitude complexes.22 On the other hand, those attitudes arising out of Specific individual experiences are more likely to stand alone or in a conflicting position with other attitudes or attitude constellations. Thus, for example, a given individual’s attitude toward Negroes may be negative due to a Specific unpleasant experience(s) with a Negro or group of Negroes. A ”stereotyped” attitude may thus be developed and may not necessarily be consistent with other attitudes which the individual holds, at least not initially. It is conceivable that an independent attitude may be opposed to a cultural value and thus be in conflict with all other attitudes arising out of that value. However, if this attitude is inconsistent or incompatible with his 22These attitude complexes or attitude constellations may be considered similar to what Heider refers to as a ”unit formation." A unit formation designates that situation in which "separate entities are perceived as belonging together.” (In this context an attitude constellation designates a group of separate attitudes which are perceived as belonging to- gether because they generally are centered around a given activity or object.) Thus, a given individual's attitude toward church, toward school, toward leisure time, etc., may all conceivably be grouped into what might be labeled the person‘s "religious attitude constellation.” See Heider, op. cit., p. 176. 3An independent attitude is defined here as an atti- tude which tends in any actor to be independent of other attitudes or attitude constellations, and which may or may not be congruent with them, because it arises out of specific (and possibly unique) personal experiences of the actor. 23 other attitudes or attitude constellations the individual must attempt to bring his total attitude orientation into balance if he is to maintain a psychological or cognitive balance.24 Thus it seems that major concern should be given to those attitudes arising out of cultural values, and grouped into attitude constellations, when attempting to predict human behavior. However, the importance of inde- pendent attitudes in explaining the behavior of deviates in society might well be investigated further.25 It has been noted that an attitude may or may not be acted upon, and that behaviors following from attitudes may vary between persons, or from time to time in one person. Thus one must also be concerned with differential attitudes, or the difference in degree or direction of affect invested in an object. It has also been noted that individuals can have attitudes toward anything they cognize. However, 24 Again we may refer to Heider's model; this time to his concept of “balanced states.” The concept of balanced state designates a Situation in which the perceived units and the perceived sentiments co—exist without stress; there is no pressure toward changes, either in the cognitive organization or in the sentiment. See Heider, o . cit., p. 175. Thus, in this context, the person is likely to bring his independent attitudes and his attitude constella— tions into a balanced state, or a co—existence without stress. 251t may be found, in this theoretical framework, that some forms of deviant behavior are as psychologically normal as conformist behavior. Merton, op. cit., p. 132. 24 cognitions may also vary. Objects not cognized by a person at one time may be cognized at another time, and one person may cognize an object that another does not. Variations among individuals, then, are vastly significant in the pre— diction of behavior. And it follows from the above that differential attitudes exist when two or more people invest an object with a different degree or direction of affect, or when an object is uncognized by one or more individuals but is cognized and invested with affect by one or more others. Thus, for example, college students with different attitudes toward life occupations may be expected to behave in differ— ent ways academically and non—academically because of direc- tional or degree differences in the affect they give to the object of their attitudes, occupational goals or aspirations. To assess the importance of a differential attitude in giving rise to a particular behavior alternative, a specification must of course be made of other attitudes out of which the behavior alternative might also possibly have arisen. Un— less the differential attitude is the only or most signif- icant attitude logically related to the behavior alternative, the other possible significant attitudes must be assessed more carefully. It has been noted that different behaviors are pos— sible with reSpect to any object. These may be called behav- ior alternatives or differential behavior. For some obvious 25 examples: one can join a club or not join it, study for an exam or not study for it, drop out of school or continue, go to a meeting or stay away. Behavior alternatives may be much more complex. In education one may attend college for any number of years, from zero to eight or more; or he may earn any one of an infinite variety of grade point averages. Other complex behavior alternatives are possible, depending upon the structure of the object. Whatever the particular possible range of behavior alternatives, predicting or accounting for differential behavior with respect to an object requires the Specification of the behavior alterna— tives. Thus, if one wishes to predict success or failure at college on the basis of attitudes, or any other items, he must determine what behavior is considered successful and what is considered unsuccessful. For example, one would probably say that a high grade point average, persistence in difficult curricula, and completion of a degree are behavior alternatives indicating success in academic work. Dropping out, low grade point average, and various curricula with— drawals would probably number among the unsuccessful behav— ior alternatives. The behavioral consequences of attitudes may be Viewed in two different ways, as object behavior and means bEhavior. Object behavior is the individual‘s behavior with IESpect to the object of an attitude variable. If, for 26 example, the variable is attitude toward Negroes, then the object behavior will be actions favorable toward Negroes, unfavorable toward Negroes, or somewhere in between these extremes. Means behavior, on the other hand, denotes behav- ior toward something which is only a means to achieving some more distant end.26 For example, education may be inter- preted as a means for achieving high occupational goals. In such a case those with high occupational goals will tend to gain more education than will those without high occupational goals. Thus, it would Seem logical to assume that differen— tial attitudes toward high occupational goals are positively correlated with level of education attained and quality of performance in the educational structure. There is empirical . . 27 ev1dence to support this. When one focuses on the range of 26Two basic and essential concepts implied here are means and end. Means is defined here as any object, device, procedure, or activity that intervenes between a problem and its final adjustment (end) and which is directed by the actor toward such adjustment. It is thus any object, device, procedure or activity that brings the actor closer to a goal. End is defined as a desired result of striving, a goal or objective. It may also include the purpose implied in a goal or objective. In the Parsonian action frame of ref— erence, the means and ends are intricately related. See Talcott Parsons, The Structure of Social Action (New York: The Free Press of Glencoe, 1949), pp. 43ff. 27A. O. Haller and I.W. Miller, The Occupational Aspiration Scale: Theory, Structure and Correlates (East Lansing, Michigan: M.S.U. Agricultural Experiment Station Technical Bulletin 288, 1963), pp. 39ff. 27 consequences of one differential attitude, each possible means and its behavior alternatives must be Specified. When a supposed means is actually being used as such, then the attitude toward the object will be correlated with behavior toward its means. To predict or account for behavior toward one object, a Specification must be made of the variety of more distant objects for which the behavior toward the form— er is only a possible means, together with the behavior alternatives of each of the more distant objects. The atti— tudes toward the more distant objects will be correlated with the behaviors toward the Specific object to the extent that the latter is in fact viewed as means to the more dis- tant ends. Thus, a student who perceives a college educa— tion as a crucial means to a Specific occupational goal will have a positive attitude orientation toward his academic performance as well as toward the occupational goal. His study behavior, academic persistence, and grade point aver— age are likely to be favorable in accord with his attitude toward the occupation. It has been noted that the behavioral consequences of an attitude always occur in situations which may be called attitude contexts. These attitude contexts include the social context, which may in turn include other individ- uals, norms, and material objects, and the intrapersonal context, which may in turn include any or all aspects of the 28 personality other than the attitude itself, such as ability, self—control, self—concept, and the like.28 It was also pointed out that the relevant effect of the attitude context is the degree to which it facilitates the expression of the attitude in behavior. Sometimes the nature of the attitude context makes it difficult or impossible for the individual to carry his attitude into behavior, and sometimes the oppo— site occurs. One individual differs from the other in the degree to which his attitude context permits him to carry his attitude toward an object into behavior. Facilitation is seldom perfect for all. This is an important reason why attitudes are not always acted upon, and why they are often not highly correlated with overt behavior. It is illustrated by the well—known phenomenon in race relations: not all prejudice results in discrimination and not all discrimina— . 29 . . . . . tors are prejudiced. Fac1lltatlon is the a551stance or 28The treatment of both personality and situational variables (or interpersonal and intrapersonal variables) as members of one class, differential facilitation variables, is done because both are here considered to perform the common function of either assisting or hindering the expres— sion of an attitude in overt behavior. The grouping of per- sonality and situational variables under one rubric is gen- erally, however, not common innuunztheoretical formulations (i.e., symbolic interactionism). 29De Fleur and Westie, 1958, op. cit.; R. T. La Piere, "Attitudes vs. Action,” Social Forces, Vol. 13, 1934, pp. 230-237; and B. Kutner, C. Wilkins, and P. Yarrow, ”Verbal Attitudes and Overt Behavior Involving Racial Prejudice,” Journal of Abnormal and Social Psychology, Vol. 41, 1952, pp. 649-652. 29 hinderance which the context of the attitude provides to the attitude in its expression in overt behavior. Thus, differ— ential facilitation may be considered as the degree to which the context of the attitude permits its expression in corre— Sponding behavior. Measurements may be made on each of several variables possibly operating as differential facil— itators.3O The precise variables having such effects will vary according to the object and its behavior alternatives. Unfortunately, there are no rules which tell us exactly which variables are worthy of consideration. In general, however, they will frequently include such things as the degree of assistance or perceived assistance offered by other people, the ability the person has to perceive the steps needed to carry the attitude into behavior, his com- mand of physical assets, his ability, and his conceptions of that ability, to perform the necessary tasks. Thus, for example, in attempting to predict student success in college one would probably consider significant student reference groups (family, peers, teachers, etc.), the student's per- ception of the importance of education for attaining his 30We tend normally to think of positive facilitation only in this context. However, the possibility and probabil- ity of zero or negative facilitation cannot be ignored. This problem will be discussed further in a later section of the thesis related to measurement. 30 goals (importance of the means behavior), the student's socioeconomic situation, and his various unique intellectual characteristics, as well as other variables. It may thus be inferred that there is also differen— tial facilitation by differential attitudes. Attitudes vary in strength or degree of affect. One person‘s attitude strength may be high, but another's may be zero. Obviously any degree of facilitation, whether low or high, will make little difference in behavior for persons whose attitude strength is very low. Conversely, the degree of facilita- tion will make a great deal of difference in behavior among those whose attitude strength is high. Thus, a student who has a degree of attitude strength (level of occupational aspiration, in this context) toward a particular occupation— al goal but who has a low mental ability test pattern will differ in behavior (i.e., grade point average) from the stu- dent who has a similar degree of attitude strength but a high mental ability test pattern. The opposite relationship would also be possible. Facilitation refers to the degree to which the atti— tude context makes easy or difficult the expression of an attitude toward an object in overt behaviors. Facilitation may be objective or subjective. That is, not only do such objective factors as social pressures and abilities function as facilitators, but such subjective factors as perceptions 31 of difficulty, self concepts, and other attitudes also func— tion in the same way. Because of the latter fact, facilita- tion may also have an influence on attitudes. There is experimental evidence which Shows that failure reduces levels of aSpiration and success raises them,31 Assuming attitudes and levels of aspiration are isomorphic, levels of aSpiration being an attitude toward the object of the aSpiration, then of course level of aSpiration will behave like other attitudes. If this is so, then the experience of successfully carrying an attitude into behavior should in- crease the strength of the attitude, and the experience of unsuccessfully attempting to carry an attitude into behavior Should reduce the strength of the attitude.32 Of course, social psychologists are keenly aware that one's experience in a given area is not the only avenue to learning about it. For example, a person also learns from other people, and he may generalize from one area of experience to another. Thus, an important factor for a theoretical formulation is the influence of the facilitation as he has learned about it. Thus, an individual who perceives his parents as giving full 31K. Lewin, et al., ”Level of Aspiration,” in Personality and the Behavior Disorders, ed. J. McV. Hunt (New York: The Ronald Press Co., 1944), Vol. 1, pp. 333—378. 32 A. O. Haller and I. W. Miller, op. cit., p. 31. 32 support, both moral and financial if" necessary, to his occupational goal, and who perceives his ability as adequate to attain it, will have greater attitude strength (higher level of aspiration) than the individual who does not have this high degree of subjective facilitation. The Problem The general problem of research arising out of the theoretical and conceptual foundation developed here is that of attempting to predict differential behavior on the basis of differential attitudes. Since the prediction of behavior on the basis of Specific factors inevitably involves one in the problem of causation, and since it is well-recognized that causal relationships in the field of social psychology are extremely difficult to establish rigorously, one must focus Specific research upon patterns of association between two or more phenomena. One must attempt to isolate a number of probable causal or explanatory variables, which when used together may be more predictive of human behavior. (See Figure 1.) Assuming that causal relations in social psychology are primarily patterns of association between phenomena and that the designata of the social sciences are multivariate in nature, the general predictive problem is approached by asking the question, "What is the nature of the relationship 33 .uofi>mno£ com: poowwm “Hosp ca pcmoflwflcwfim pmoe omonp mo meow mum nonmomou was» Ge UoHHonwcoo pom powuoaom AmoHDMflum> pXowcoo unapfipgmv moanmflum> Hmpcoecoufi>cm on“ “any UoESmmm Hosanna we pH .ooooaoom one on oooooamos we pomdefl was“ “map can .moHDMMnm> >ufiamnomuom Honpo one paw moonpflwpm nonwo coo: mm Hams mm .oouSmmoE oBSHMpwm on“ com: pummefi cw Um: Ham o>m£ moaomfinm> HwHSHUDHDm Hmfi60m msofium> map “may oosdmmm we pH . XHOSOENHW HNGOH #mfiuvflw m cenuez moUSHEppm Heepsouowwfio mo mfiwwn on“ :o noe>mnon mcmos Hmfipcouow uwflo mo soapofiooum onp nwfiz we cuoonoo ocH :HcpflB xnoaoEmnw HmSonocoo pom Hmofipouoonp Hmnocom ogy m>mupuoa Emummflo 62H Astaoameoooo mo moouwoo woe -soe> can: soomv nofi>mnon mcwoz .m soa>eeoo ooomoo .< man 0 n.1 0+t 41c t a a r . {le uwm Cit o n 5.1 “doocoo mHom S A mo>flpoz Ammanmfium> “Kopcoo opsofluup ocm mpfimufi Ammanmflum> soapmpfiaflommv haeooa>aoea mooae> on“ mo Homecoufl>co Hm:0muomnmupxo pom Immune may moosaoch m663pepp< ZOHHmupuom Ewuwmflo onH .m madman m>pfiaflnm mo ammocoo waomv oocmEuowqu ofiEoUmom “mom .6 Ammoaaoo mo >pfimmooocv u0fi>mnon momma mo mocmwuomefl mo coeudoouom .o “HOQQSm macaw mucouowou oo>fioonom .o >pfiafinm Hmwcoz .m oocopmfiwuod omoaaoo .N moanmfium> Hm20muomnmnch .m mapmpm ofiEo:0660fioom .m owmuo>m “neon oomum owoaaoo .H oaomfium> HmGOmuomumupxm .H cofipmufidmm Hmnofipmdsooo .H ooaseeoo HmaoeosoMGSQAmlll.soaoeoeaaoom Heeoeosomwao.mlllihooooaopo Hmaooosommao AmoH> eoaoeoaaaoem HO HMHfiMECOHerfimV moH>2-2-<2z)21 where y = Score of first administration 2 = Score of second administration. 5Its application appears to be logically sound within the theoretical framework used here. See Calvin F. Schmid, ”Scaling Techniques in Sociological Research,” Chapter 14, in Pauline Young, Scientific Social Surveys and Research (New York: Prentice—Hall, Inc., 1949), p. 376. 66 father's occupation, the Duncan ”Socioeconomic Index for All Occupations" was used.36 The Duncan Index is particularly well-adapted to deriving a measure of socioeconomic status for individuals from documents listing only occupations, and excluding more detailed information about income and educa- tion. It makes possible the converting of occupation into an index of socioeconomic status, this also being the orig- inal motivation for the construction of the instrument. This index, developed by Otis Dudley Duncan on the basis of 1950 census occupations, is the most careful and elaborate classification of occupations by socioeconomic rank known to the writer and therefore was felt to be most applicable to the present study. Thus, the Duncan Index was used to differentiate among the members of the sample in terms of socioeconomic status. The Duncan Index is constructed from two factors, education and income. The suitability of these factors for the construction of the index is argued elsewhere and may be reviewed directly by the interested reader.37 The case for the validity and stability of the index is also presented by 36For a detailed description of this instrument, see Otis Dudley Duncan, ”A Socioeconomic Index for All Occupa- tions,” Chapter VI in Albert J. Reiss, Jr., Occupations and Social Status (New York: The Free Press of Glencoe, 1961), pp. 109ff (also see Chapter VII). 37 Ibid., pp. ll4ff. 67 Duncan and need not be repeated here.38 All occupations were scored by the writer thus insuring maximum reliability of scoring procedures. It may be concluded on the basis of the characteristics of the index, and the data available, that it is particularly suitable for use in the present study. Mental Ability The measure of mental ability used in the research is the College Qualification Tests (CQT), Form B, 1956.39 Three sub—tests provide a total score which constitutes a measure of general academic aptitude or mental ability for the individuals taking the test. The sub—tests consist of a verbal category which is a measure of vocabulary and reflects verbal abilities which are most significant in social sci- ences, literature, etc.; an information category which is a measure of general information in the social and natural sciences; and a numerical category which is a measure of abilities involved in quantitative thinking. Since the CQT is primarily a power test (students who have ability to answer the test items have the necessary time to do so), its 3§1219-, pp. 124—125, 151—154. 39For detailed information concerning these tests, see G. K. Bennett, M. G. Bennett, W. L. Wallace, and A. G. Wesman, College Qualification Tests Manual (New York: The Psychological Corporatian, 1961, revised), 61 pp. 68 reliability may properly be estimated through coefficients of internal consistency. The reliability coefficients of this type generally exceed .90 and the standard error of measurement is about 5.50.40 Since the tests were developed primarily to serve as predictors cd’ success in college per— formance, tests of its validity are best measured by its utility in predicting success of students in college, the criterion being grade point average. The overall picture of the validity of the CQT is very favorable. The majority of the coefficients found in the prediction of grade point average strongly indicate that the CQT is highly predictive of grade point average and college Success. Validity coef— ficients range generally between .50 and .70 for various four-year institutions.41 It must be remembered that the validity coefficients are likely to be larger for institu- tions which represent wide ranges of academic ability than for those which are more homogeneous in this respect. Also, the coefficients are likely to be higher if assigned grades are reliable, well-distributed among students, and based on actual achievement. If grades are unreliable it is futile to hope that any test, however well prepared, will succeed in forecasting those grades. 4OIbid., pp. 46—47 (Tables 31 and 32). 4llbid., pp. 52—53 (Table 35). 69 It is concluded that the CQT is a reliable and valid instrument, and its use in this research is adequately jus— tified. Perception of Importance of Means .Behavior (Necessity of College) The perception of the degree of importance of means behavior (college performance) variable is measured by just one item. The item permits the student to indicate how important a college education (means behavior) is for obtain- ing the occupational aSpiration (object behavior) as he sees it. The range extends from ”absolutely necessary” to ”not important at all.” To obtain the kind of job I wish to have in later life college education is absolutely necessary very important quite important of little importance not important at all. AAA/\A- vvvvvm This one item is used to measure the degree to which the student perceives the means behavior as essential for his job aSpiration. The answer is scored and tabulated into a numerical value and this is then considered the perception of necessity of means behavior (college performance). The reliability of this item was estimated by use of the test— retest technique which provides a reliability coefficient 70 42 . . . . of .63. The validity of the item cannot be estimated, however, the logical test of validity seems justified in 43 this operational framework. Past Academic Performance (Self Concept of Ability) The student's high school academic average is the item used in the measurement of the past academic performance variable. All high school averages were computed on the basis of the system used at Michigan State University. Stu— dents from high schools using other systems of grading were separated out and their averages were converted and re—com- puted on the basis of the common system. This method con— stituted the construction of an index with theoretical limits of 0 and 400 at the two extremes of the continuum. This index was used as the past academic performance (self concept 42The test-retest reliability coefficient was com- puted by the product-moment method. Equivalent forms of the ”About My College Plans” questionnaire (see Appendix C) were administered to a sample of 43 freshmen students taking a course in political science (PLS 100) during the Spring term of 1964. A three week interval occurred between the two administrations. The equation used in calculation is: niiyz - iiyigz r 2 YZ [11 £(y)2-( iy)2][n ((2)2—(22121 where y = score of first administration 2 = score of second administration. 43See Schmid, op. cit., p. 376. 71 of ability) measure and all individuals in the sample were thus placed and scored with respect to this variable. The reliability of the grades given in high school must be assumed to be adequate. Since past performance or self concept of ability indices are generally used as predictors of future performance, the predictive validity of the index used here may be estimated from the present research. The correlation coefficient of this index with college perform— ance is .57. General Facilitation A general facilitation index was additively con- structed by utilizing the five selected variables discussed above. (See Appendix E.) This relatively crude index was also used in the analysis of data as a type of summed meas— ure of facilitation. It should be noted that all of the facilitation indices described above are designed to measure the degree of facilitation in the positive direction only. Thus each student in the sample is positively facilitated to some degree (although this may be very low) to carry his attitude into behavior, or to carry his level of occupational aspira- tion into satisfactory and successful academic performance (means behavior) which will, in turn, make it possible for him to achieve his occupational aSpiration (object behavior). 72 The nature of the present research, or more specifically of the facilitation indices, assumes then some degree of posi— tive facilitation for all. There is theoretically, however, no reason to believe that one cannot be negatively facil— itated (positively impeded) or, at least, not facilitated at all in this respect (zero facilitation). It would appear entirely logical, although possibly somewhat improbable or infrequent, that this might be the case. For example, it is possible that all one's significant reference groups might discourage his going to college. This would, in fact, be a case of negative facilitation. The theoretical problem is not, however, restricted to indices which are used as meas- ures of facilitation variables as in the framework guiding this research. For example, all level of aspiration meas- ures known to the writer are so constructed that the re— Spondent obtains some sCore or value above some theoretical zero point, which is then interpreted as a positive index. Yet it does not seem empirically impossible for individuals to have no degree (zero) of aspiration toward given objects, or even negative aSpiration toward them. Certainly this is frequently the case concerning many attitude variables other than level of aspiration. Although resolution of the prob— lem is beyond the scope of this research, it is felt that it is significant enough from a theoretical perspective to be investigated further. It appears to constitute a major 73 theoretical problem of measurement in the social sciences, particularly social psychology. The Behavioral Variable The major behavioral variable (and in this theoret— ical framework the dependent variable), college performance, is measured by the official University cumulative grade point average for students of the sample. The student's grade point average is selected as the measure because it provides the widest selection of behavior alternatives and is a good measure of differential performance (behavior) in college. Grade point average in college is the basic crite— rion used by professors, administrators and students to evaluate a student's academic performance. Hence, it is fairly fundamental as a behavior variable in college. Although the merits of grade point average as a measure of performance quality are often debated in academic circles and its use for this purpose often criticized, it tends to maintain its position as a good measure. There is no empir- ical evidence known to the writer which can refute this position. In any event, it is the best measure available to the researcher in this field and no better one has been suggested or proposed. The cumulative grade point average used for the sam— ple is that on the student's official University record 74 during his seventh term at Michigan State University, which would normally be the first term of the junior year. For those students who have dropped out of school the official University cumulative grade point average at the time of dropout was used. This procedure was used primarily on the rationale that most students who are performing unsatisfac— torily in their academic work by the seventh term have dropped out of college voluntarily or have been compelled to leave the University. (Since students at Michigan State University must attain a 2.00 or better average to enter an upper degree-granting college at the junior level, most of the attrition has taken place by this time.) Dropout from college was also used in further analy- sis as a behavioral alternative to remaining in college.44 This type of analysis calls for a dichotomous measure of success—failure in the total means behavior. It provides a self-evident index of apparent failure (which may be more palatable to those who seriously question the ultimate sig— nificance of the differences in grade point averages) and provides insights into the differences between these two oft-discussed groups. 44A "dropout” was operationally defined as a student who was not enrolled in the University during the Fall term, 1962, which would normally be the seventh term or the begin— ning of the junior year. Students who had left the Univer- sity prior to this time, but who returned, were not defined as dropouts. 75 The Hypotheses From the postulates developed and stated in Chapter 11 one may logically construct a number of hypotheses which can be tested within the operational framework of the pres— ent research. We shall review briefly here the theoretical framework and the postulates as developed in Chapter II so that the reader may observe specifically the derivation of the hypotheses. It was noted that an attitude may result in behavior toward an object and that it may also, in turn, result in behavior toward means of fulfilling Specific object behavior. Because the behavioral consequences of an attitude always occur in situations, attitudes are differentially facil- itated in their expression in behavior by the situation at hand, or the attitude contexts. The postulates, therefore, are constructed by the use of three types of variables: attitudinal, facilitational, and behavioral. The postulates predict certain interrelationships between these variables. The postulates are stated as follows: 1. When individuals differ in their attitudes toward an object, and when the facilitation offered by the attitude context (or any one facilitation variable) is greater than zero, their attitudes will be positively correlated with their behavior with reSpect to the object, and with respect 76 to objects viewed as means for promoting their desired behavior toward the object of the attitude. a. When the degree of facilitation offered by the attitude context (or any one facilitation var- iable) is high, the positive correlation will be significantly higher than when the degree of facilitation offered by the attitude context (or any one facilitation variable) is low. 2. When individuals differ in the degree to which their attitude contexts facilitate the expression of an attitude in behavior, and when their attitude strengths are greater than zero, there will be a positive correlation between the facilitation offered by the attitude context (or any one facilitation variable) and their behavior with respect to the object, and with respect to objects viewed as means for promoting their desired behavior toward the object of the attitude. a. When the attitude strength is high, the positive correlation (between facilitation and behavior) will be significantly higher than when the attitude strength is low. 3. When the degree of facilitation offered by the atti— tude context (or any one facilitation variable) varies among individuals, there will be a positive correlation between the degree of facilitation offered by the attitude context (or any one facilitation variable) and the attitude strength. The theoretical focus of the hypotheses is thus the attitude—behavior relationship. The focus within the opera— tional framework, then, is the relationship between level of 77 occupational aspiration (attitude) and academic performance at college (means behavior). However, the facilitation variables outlined in the previous section on Measurement are felt to be significant in modifying this relationship and thus the inquiry considers their effects on this rela— tionship. The basic and operational question of the re- search is thus, "What is the nature of the relationship between level of occupational aspiration and level of per- formance in college when certain facilitation variables and varying conditions of facilitation are considered and con- trolled?" The specific hypotheses to be tested are stated below. 1. When individuals differ in their levels of occupa- tional aSpiration, and when the facilitation offered by the attitude context (or any one facilitation variable) is great- er than zero, their levels of occupational aspiration will be positively correlated with their levels of performance in college. (Because the facilitation offered by the attitude context is only minimally controlled in this central hypoth- esis, that is, it is simply above zero, it is expected that the positive correlation will be significant but relatively low.) a. When the degree of facilitation offered by the attitude context (or any one facilitation var— iable) is high, the positive correlation will be significantly higher than when the degree of facilitation offered by the attitude context (or any one facilitation variable) is low. l) 2) 4) 5) o) 78 When the mental ability is high, the corre— lation between occupational aspiration and grade point average will be Significantly higher than when the mental ability is low. When the perceived reference group support is high, the correlation between occupa- tional aSpiration and grade point average will be significantly higher than when the perceived reference group support is low. When the socioeconomic status is high, the correlation between occupational aspiration and grade point average will be significant— ly higher than when the socioeconomic status is low. When the perception of the necessity of college is high, the correlation between occupational aspiration and grade point average will be significantly higher than when the perception of the necessity of college is low. When the past performance is high, the cor- relation between occupational aspiration and grade point average will be significantly higher than when the past performance is low. When the general facilitation is high, the correlation between occupational aspiration and grade point average will be significant- ly higher than when the general facilitation is low. 2. When individuals differ in the degree of facilita- tion offered by their attitude context (or any one facil- itation variable), and when their levels of occupational aSpiration are greater than zero, there will be a positive correlation between the facilitation offered by the attitude context (or any one facilitation variable) and their levels 79 of performance in college. (Because the level of occupa— tional aSpiration is only minimally controlled in this central hypothesis, that is, it is simply above zero, it is expected that the positive correlations will be significant but relatively low.) a.“ When the level of occupational aSpiration is high, the positive correlation will be signif- icantly higher than when the level of occupa- tional aspiration is low. 1) When the occupational aSpiration is high, the correlation between mental ability and grade point average will be significantly higher than when the occupational aspiration is low. 2) When the occupational aspiration is high, the correlation between perceived reference group support and grade point average will be significantly higher than when the occupational aSpiration is low. 3) When the occupational aSpiration is high, the correlation between socioeconomic status and grade point average will be significantly higher than when the occupational aspiration is low. 4) When the occupational aSpiration is high, the correlation between perception of the neces- sity of college and grade point average will be significantly higher than when the occupa— tional aSpiration is low. 5) When the occupational aspiration is high, the correlation between past performance and grade point average will be significantly higher than when the occupational aspiration is low. 6) When the occupational aSpiration is high, the correlation between general facilitation and grade point average will be significantly higher than when the occupational aSpiration is low. 80 3. When the degree of facilitation offered by the attitude context (or any one facilitation variable) varies among individuals, there will be a positive correlation be- tween the degree of facilitation offered by the attitude context (or any one facilitation variable) and the level of occupational aspiration. 1) 2) 3) 4) 5) 6) Mental ability will be positively corre- lated with level of occupational aspira- tion. Perceived reference group support will be positively correlated with level of occupa— tional aspiration. Socioeconomic status will be positively correlated with level of occupational aSpiration. Perception of the necessity of college will be positively correlated with level of occupational aspiration. Past performance will be positively corre— lated with level of occupational aspiration. General facilitation will be positively correlated with level of occupational aspiration. These hypotheses will be discussed further and tested in Chapter IV. 81 The Sample In this section of the chapter only the most rel- evant and important characteristics of the sample, and the procedures utilized to select it, will be presented. A great deal of descriptive information does not appear directly relevant to the main thesis of the study. The subjects on whom data were gathered are 585 new freshmen entering Michigan State University in the fall term of 1960. This sample is approximately random and was selected from the total new entering freshman class which numbered 4,583.45 The M.S.U. Office of Evaluation Services was asked to administer the Occupational Aspiration Scale (see Appendix A) and the About My College Plans question— naire (see Appendix C) to 600 new freshmen students during the fall term orientation week.46 Since orientation tests are given to students in groups of various Sizes, the Office of Evaluation Services Selected several groups at random to which it administered the scale and questionnaire. Thus, 45These were students on the East Lansing campus only and were first-time college students. This figure does not include students entering other than regular degree pro- grams (i.e., one and two year terminal courses, etc.). 46Fifteen of the scales and questionnaires were ei- ther not completed by the student or were unidentifiable after administration. The remaining 585 students constitute the Sample for this research. 82 although the sample of students was not selected at random in the strict sense of the term, it is felt that the tech— nique used was sufficiently systematic in efforts to obtain randomness so that we may assume a random sample. Empirical evidence would also suggest that this assumption of random— ness is justified. Table 1 shows certain demographic data for the total population from which the sample was selected. Differences, where they exist, are extremely small and it is concluded that these are not significant for this research. The sample is thus confidently assumed to meet the qualifica- tion of randomness.47 The sample includes 305 men and 280 women students representatively located with respect to major field of study in the nine different colleges of the University at the time of entrance. (See Table 1.) Seven students in the sample were married. Exactly 82.7% of the sample was made up of 18 year olds at the time of entrance. Another 7.7% was 17 years old and 7.2% was 19 years old. Only two students were less than 17 years old and 12 were twenty years old or older. Of the total sample, 205 students had dropped out of school before the seventh term or what would normally be the beginning of the junior year. Thus, 35.1% of the sample 47On the basis of the evidence, we may reasonably generalize our findings to the entering freshman class of 1960 at Michigan State University. 83 Table 1. Comparison of Sample and Total Population on Certain Demographic Characteristics Sample N==585 Total Population N==4,583* N % N % Sex Male 305 52.1 2,450 53.4 Female 280 47.9 2,133 46.6 Residence Michigan 452 77.3 3,456 75.4 Other 133 22.7 1,127 24.6 College of Major Agriculture 31 5.3 208 4.6 Business 104 17.7 765 16.7 Com. Art 20 3.5 177 3.8 Education 59 10.1 460 10.0 Engineering 77 13.1 570 12.5 Home Economics 29 4.9 196 4.3 Sci. & A. 153 26.2 1,211 26.4 Vet. Med. 27 4.6 206 4.5 Univ. 85 14.6 790 17.2 (No pref.) * Data derived from Enrollment Report, Fall 1960 (East Lansing, Michigan: Michigan State UniverSity, Office of the Registrar), pp. 3-4 (Table 11). 84 may be considered "dropouts” in the present study. An additional 34 students were performing unsatisfactorily (below the 2.00 grade point average) but were continuing in an academic program at the University. Thus, a total of 239 students or 40.9% of the sample had either dropped out or were displaying unsatisfactory academic behavior at the time the behavioral variable was measured for this research (seventh term or beginning of junior year). Of the 205 dropouts, 157 had already dropped out during, or at the close of, the first academic year. Thus, 26.7% of the sam— ple dropped out in the first year. These data also compare favorably with the official records for the total population in the Office of the Registrar.48 p The research procedures after selection of the sam— ple may be outlined briefly here. For each student in the sample, additional data were necessary beyond the completed Occupational ASpiration Scale (see Appendix A) and the About My College Plans questionnaire (see Appendix C). Much of the additional necessary data was obtained from each stu- dent's application and credentials for admission to the 48Data in the Office of the Registrar Show 1,237 of 4,583 dropped out during the first year (27.0%), 1,852 of 4,583 dropped out during the first two years (40.4%) and 2,028 of 4,583 dropped out during the first three years (44.3%). These data also support the contention that most of the dropout occurs during the first two years. 85 University. Still other portions of necessary data were obtained from the official records of the Office of the Registrar and the Office of Evaluation Services at Michigan State University. All of the editing, scoring, coding and tabulating of the data was completed by the writer. All data were recorded on standard coding sheets and were punched on cards by the Data Processing Department (see Appendix B). Initial analysis was done by the writer on IBM counter—sorter equipment. However, the major statistical analysis was completed by the Control Data 3600 computer at Michigan State University. Limitations of the Sample The nature of the sample utilized to determine the adequacy of the postulates imposes limitations which Should be reviewed briefly at this point and possibly discussed further in the final chapter of this thesis, particularly as these limitations may affect the results obtained. The general limitation may be stated and explained simply as the relatively limited range of variability of the selected sample on several of the critical independent vari— ables used in the research. A rather limited actual range of variability of the sample on the attitudinal variable, level of occupational aspiration, and several facilitation variables, namely: mental ability, past academic performance, socioeconomic status and perception of the necessity of 86 college, is observed. In the case of each of these vari- ables there is a conspicuous absence of respondents at and toward the low extreme of the response continuum. The attitudinal variable, level of occupational aspiration, is probably most significant in this respect. There are no reSpondents with really low occupational aspira- tion in the sample; this may readily be explained, it seems, on the grounds that individuals with low aspiration would probably not make the necessary prior decision to go to college. Thus, the very nature of this new college student sample has doubtless eliminated the low—aSpiring persons. Furthermore, it must be noted that there is also a lack of reSpondents at the very high end of the occupational aSpira- tion continuum. Thus, the majority of the respondents varied within the "middle range" of the continuum. Similarly, and also quite significant in the present theoretical framework, there are no really low respondents on the scales measuring mental ability, socioeconomic status, past academic performance, and perception of the necessity of college. This also seems to be due to the nature of the sample. Selective admission procedures practiced by the University automatically impose the lack of low respondents on the mental ability and past academic performance scales. Again, the lack of respondents with very low perceptions of the necessity of college for attaining their desired job may be explained on the grounds that most students with low 87 perception of the necessity of college simply do not make the decision to go to college in the first place. The lack of reSpondents with low socioeconomic status may be explained on the basis of several arguments, each of which may be valid to varying degrees with different persons. Low socioeconomic status may make college attendance extremely difficult from a simple economic or financial standpoint. It may reflect a cultural phenomenon placing college attendance low on the value heirarchy for these groups. It may also be a lack of support from ”Significant others” for college attendance. In any event, it seems reasonable that these factors, along with others, may impose the scarcity of low socioeconomic status respondents in the sample. Thus, for all of the above variables, it may be contended that the very nature of a new college student sample imposes the lack of this ”low” category of reSpondent. The above limitation is illustrated in Figure 3. The implications of this limitation for the analysis may be reflected in somewhat lower correlation coefficients than what otherwise might be expected when utilizing these vari- ables. For example, a lower correlation between level of occupational aspiration (where the sample is characterized by clustering at the middle range) and grade point average 49For detailed data regarding this limitation, also see Appendix G. 88 (where the sample is well-distributed over the continuum) would be expected than if the level of occupational aspira— tion were also well-distributed. This may also be illus- trated (see Figure 4). If the limits of the sample on level of aspiration could be extended, the theoretical effect would be to raise the correlation by extension of the cases along the X axis with theoretically corresponding Y axis relationships. LOA MA SES RGS GPAh PNC H17\ ———————————————————— —""‘ ————— “wr— ----- -F- ————— -1- —————— fi_. Theoretical limits of the scale Loy. _____________________________ _.t_ ______________________ Figure 3. The diagram shows that the range of variability of the sample on almost all of the variables used in this research does not extend to the theoretical limits of the scales used, particularly the low theoretical limits. Furthermore, clustering with- in these ranges of variability also are observed (see Appendix G) and thus further complicate the nature of this limitation of the sample for the present research. 50For further discussion of range of variability and its effects, see Hubert M. Blalock, Jr., Social Statistics (New York: McGraw-Hill, 1960), pp. 290-291. 89 ASpiration) 4.00 , I I " I I .f I . I .. I I (Grade --F----------3-3--4-- .. POint I .‘-.. .|"~.-' . . 1 \L —————— 1. ts f th Average) F , . . ... q.” , lml 0 e . -; ,“ -..1~ Sample on the :. I 'f..: .L2'1‘ LOA variable i', , . Z: Y I; .‘.'i- , -l .I. I' . '3 .I ' . . .I I ; l w , 'I .'c ...... .' I, I .0 ‘I ‘ . 'I . .I ' I . '0 .i I .I' ~ I a‘ ._J;_______________J__ -.‘. .I l . .I l I I Z . i 1 (Level of 0 d : X : 72 Occupational I I I I l I Figure 4. The diagram shows almost no relationship within a limited range of variation in the X variable but a positive relationship over a hypothetical total range of X. This limitation of the sample is described here so that the reader may be aware of it prior to reviewing the next chapter, and so that the results of the analysis may be interpreted with this forewarning. This limitation is not considered to be prohibitive from the standpoint of testing the hypotheses, however, it is felt to be important in evaluating the strength of the correlations which obtain. 90 Summary of Chapter This chapter has been focused upon presentation of the total research design. It has attempted to give the reader a complete description of the operational framework, the measurement techniques and procedures, the variables, the sample, and the Specific hypotheses to be tested. Several problems and limitations of parts of the total design were also discussed. The results of analysis and tests of hypotheses will now be discussed in the next chapter. CHAPTER IV THE RESULTS Introduction This chapter presents the results of the data analy- sis in relation to the theoretical and operational frameworks set forth and discussed in Chapters 11 and III. The analysis employed in this research utilizes the correlation coeffi— cient as the major measure of relationship. The postulates and hypotheses are stated in terms of correlation and, thus, this chapter will present primarily the correlational analy— sis of the data as guided by the stated hypotheses. This approach does not, of course, preclude or prevent the further analysis of the data by other statistical techniques. The presentation in this chapter, then, consists of the statistical testing of the twenty Specifically stated hypotheses derived from the postulates assumed in this re— search. Although there is some effort to interpret the Specific results in this chapter, a more general attempt to deal with them within the context of the theoretical framework will be delayed for the final chapter. 91 92 Tests of the Hypptheses The hypotheses guiding this study are so stated that the actual tests of them may be made by analysis of the cor- relations derived after appropriate programming and comput- ing by the Control Data 3600 computer used by the Michigan State University Computer Laboratory. Thus, the technique used is the correlation coefficient (r) which, if statisti- cally significant, indicates the existence, degree and direction of the relationship or association between two variables. Since the theoretical framework used in this study allows one to predict the direction of the relation— ship, a one—tailed test of the significance of the relation— ship may be applied.1 The ”t” test of significance is used in each case to test the null hypothesis of no relationship between the variables.2 The .05 probability level will be 1The formula for the computation of "r” is that used by the Michigan State University Computer Laboratory in its Control Data 3600 Statistical Program I (Identification code: Ll.01M). The correlation is calculated by the equation: : néixy —2{x:{y In ((X)2-(ZX)2][ni(y)2—(iy)2] rxy 2The formula for the computation of ”t” is that suggested by Edwards: . = ha See Allen L. Edwards, Statistical Methods for the Behavioral Sciences (New York: Rinehart and Co., 1954), p. 303. 93 the criterion used for the acceptance or rejection of the null hypothesis. Thus, if the size of the ”r” is found to be statistically significant, the correlation will be accepted as supporting the hypothesis in question. In those cases where the hypotheses predict a signif— icant difference between two correlation coefficients, the ”2” test for significance of the difference is applied. The null hypothesis of no difference between the coeffi- cients is thus tested, again using the .05 probability level as the criterion for acceptance or rejection (one—tailed tests are applied since direction is again predicted in each case). If the size of ”z” is found to be statistically sig- nificant, it will be accepted as supporting the hypothesis in question. Implications of Acceptance or Rejection The theoretical framework of the present study pro- vides general postulates which when stated operationally as general hypotheses are broken down into twenty Specific 3The formula for the computation of ”z” is that suggested by Edwards: 1 l Z1'22 fii‘Zi See Allen L. Edwards, op. cit., pp. 304—307. 94 hypotheses. The acceptance or rejection of these hypotheses must therefore have implications for either the theoretical framework or the operational framework, or possibly both. In short, if the hypotheses are all supported both the theoretical and the operational frameworks must be accepted as, at least, tentatively valid. If the hypotheses are not supported either one or both of the frameworks must be accepted as, at least, tentatively invalid. Furthermore, if some of the hypotheses are supported and others are not, at least tentative conclusions must be drawn about parts of either the theoretical or the operational framework, or both. These alternative conclusions may be briefly suggested here and discussed further in the final chapter if the results deem it necessary. If the expected relationship is not obtained in the testing of the first general hypothesis, attention must be focused upon both the total theoretical framework and the total research design and rational decisions made regarding possible weaknesses in each before at least tentatively rejecting either or both. If the expected relationships are not obtained in the testing of any or all of the sub- hypotheses, attention must first be focused upon the facil- itation variable in question within, and in relation to, the total research design. Because of the nature of the method utilized for selecting the facilitation variables and due to 95 certain limitations in the sample, it seems appropriate that only after intensive investigation of the variable and the research design should the theoretical framework be called into question on the basis of the rejection of any or all of the sub—hypotheses. If the expected relationship is not obtained in the testing of the second general hypothesis, attention must again be focused upon the total theoretical framework and the total research design. On the basis of logical and systematic decisions regarding the conceivable weaknesses in each, either or both must be at least tentatively reject— ed. If the expected relationships are not obtained in the testing of any or all of the sub-hypotheses, attention must be focused upon the facilitation variable in question, the attitude variable, and the total research design before making decisions in regard to the theoretical framework and its adequacy. As in the case of the Sub-hypotheses of Hypothesis 1, it is felt that this sequence of concern and analytical attention is necessitated by: (1) the nature of the method utilized for selecting the facilitation variables, and (2) certain limitations in the sample. If the expected relationship is not obtained in the testing of the third and final general hypothesis, attention must be focused upon that part of the theoretical framework which contends that degree of facilitation also influences 96 degree of attitude strength or, in operational terms, that level of occupational aspiration will be influenced in direct relation with the level of facilitation experienced or perceived for the attainment of the object of aspiration (occupation), or means objects (education).4 If the rela— tionship does not obtain for a given facilitation variable, however, that variable should first be analyzed in regard to its legitimacy within the research design. Since the results of analysis will determine, in part, how extensively these alternative conclusions must be discussed, it is not necessary to pursue them further here. Hypothesis 1 The first general hypothesis of this study is that level of occupational aSpiration is positively correlated with level of performance in college for individuals who obtain some degree of positive facilitation from the situa- tion or environment within which they are operating (atti— tude context). From this general hypothesis six sub-hypoth- eses are formulated which are tested by controlling the degree of facilitation offered by five selected variables which were isolated in this study as significant facilita— tion variables, and a general facilitation variable 4See Chapter II, pp. 29-32. 97 constructed by combining the five selected variables.5 Test— ing this general hypothesis and its six sub-hypotheses con— stitutes the test of Postulate l of the theoretical frame— work. Hypothesis 1 When individuals differ in their levels of occupa- tional aspiration, and when the facilitation offered by the attitude context (or any one facil— itation variable) is greater than zero, their levels of occupational aSpiration will be posi— tively correlated with their levels of performance in college. The test of this general hypothesis was completed by the correlation of just two variables for all cases in the sam- ple: level of occupational aspiration and level of perfor— mance in college (grade point average). The correlation coefficient obtained was .161. The null hypothesis of no relationship was tested by application of the ”t” test of I O 0 6 Significance. r=.l6l N=585 t=3.93 df=583 P<.Ol Since the ”t” value is statistically significant, the cor— relation of the two variables is significant and the null 5See Appendix E for detailed explanation of how this index was developed. ésee footnote 2 p. 92 of this thesis for the statis- tical formula used in computation. 98 hypothesis of no relationship may be rejected. As was ex- pected, the size of the correlation is quite low. The ex— pectation of a low correlation, as was indicated earlier, was based upon the minimal control of facilitation offered by the attitude context, that is, facilitation was simply above zero. Theoretically, facilitation is always greater than zero in this study since all instruments used were designed to measure only the degree of positive facilitation. The support of this hypothesis provides initial evi- dence to the theoretical position that attitudes are corre- lated with means behavior. Thus, it tends to support the position that the actor will behave in a consistent manner with regard to both objects toward which he has an attitude and objects viewed as means for promoting the desired behav— ior toward the object of the attitude. In the case of this operational framework, the actor behaves in an expected manner toward an academic program which is a means to attain- ing the desired occupational status toward which he aspires. Sub—hypothesis 1.1 When the mental ability is high, the correlation between occupational aspiration and grade point average will be Significantly higher than when mental ability is low. The test of this hypothesis was performed by dividing the sample into three groups by degree of facilitation on the mental ability variable. Thus, the sample was divided into "High," ”Medium,” and ”Low” facilitation groups with respect 99 to the mental ability variable.7 The ”High” and ”Low" groups were then compared to determine the significance of the difference between the correlation coefficients of occu- pational aSpiration and grade point average. The null hypothesis that the ”High” group would have a correlation coefficient between occupational aspiration and grade point average equal to or smaller than the ”Low" group was tested by application of the ”2” test for Significance of the difference. High r = .168 N = 195 2' = .1696 (Medium r=—.058 N=l95) Z=1.01 P<.16 Low r = .066 N = 194 Z' = .0661 The test of significance indicates that the difference is not statistically significant at the .05 level of probabil— ity and thus the null hypothesis that there is no difference cannot be rejected. These data tend, therefore, to reject the hypothesis that the correlation coefficient between level of occupational aspiration and grade point average is significantly higher for high mental ability students than for low ability students. 7This was done by dividing the index arbitrarily in- to three as "numerically equal" parts as was possible. The same procedure of allowing the data to determine the cate— gories is used for the other facilitation variables. See Appendix K for detailed information concerning the empirical basis of these categories. 8See footnote 3 p. 93 of this thesis for the statis- tical formula used in computation. 100 Nevertheless, several observations in the data are noteworthy. First, the difference is in the expected direc- tion, although it is not statistically Significant. Second, the "Medium" mental ability group shows a negative relation- ship between level of occupational aspiration and grade point average. Theoretically, the relationship would be expected to fall somewhere between the "High" and ”Low” groups. Thus, it deviates from the expected pattern. Third, the correla- tion coefficients for both the ”Medium” and ”Low” groups are not significantly different from zero and thus the values may be considered Spurious variations around the zero point. Finally, the correlation coefficient for the ”High” group is statistically significant from zero. These observations will be considered further after testing the remaining sub— hypotheses of Hypothesis 1. Sub—hypothesis 1.2 When the perceived reference group support is high, the correlation between occupational aspiration and grade point average will be significantly higher than when the perceived reference group support is low. The null hypothesis that the ”High” group would have a cor- relation coefficient between occupational aspiration and grade point average equal to or smaller than the "Low" group was tested using the same procedures as in sub-hypothesis 1.1. .191 N==227 z' .1934 .065 N=252) Z=l.82 P<.05 -.049 N== 80 z' -.0490 High r (Medium r Low r II 101 The "2” test of significance indicates that the difference is statistically significant at the .05 level of probability and therefore the null hypothesis can be rejected. Thus, the hypothesis that the correlation coefficient between level of occupational aSpiration and grade point average is significantly higher for students with high perceived refer- ence group support than for students with low perceived sup— port may be accepted. Within the theoretical framework utilized here, then, perceived reference group support operates as expected as a facilitation variable. Sub—hypothesis 1.3 When the socioeconomic status is high, the correla- tion between occupational aspiration and grade point average will be significantly higher than when the socioeconomic status is low. The "2" test of significance of difference was applied to the null hypothesis that the ”High” socioeconomic status group would have a correlation coefficient between occupa— tional aSpiration and grade point average equal to or smaller than the "Low" socioeconomic group. High r = .332 N = 171 z' = .3451 (Medium r = .180 N = 196) Z = 3.16 P‘<.001 Low r = .007 N = 185 z‘ = .0070 The statistic indicates that the null hypothesis may be re- jected; which means that the hypothesis is tenable as proposed, that is, that high socioeconomic status students have a signif- icantly higher correlation between their occupational aspira- tions and grade point averages than do low socioeconomic 102 status students. Thus, it may be contended that socio- economic status operates as expected as a facilitation vari— able in the theoretical framework guiding this research. Sub-hypothesis 1.4 When the perception of the necessity of college is high, the correlation between occupational aSpira- tion and grade point average will be significantly higher than when the perception of the necessity of college is low. The test of the null hypothesis that those students with a "High" perception of the necessity of college will have a correlation coefficient between occupational aspiration and grade point average equal to or smaller than students with "Low" perception of the necessity of college was accomplished by use of the ”2” value. High r =.160 N==357 z'==.16l4 (Medium r = .224 N = 135) Z = .606 P < .28 Low r=.070 N= 84 z'=.0701 The obtained statistic indicates that the null hypothesis of no difference cannot be rejected. Therefore, the hypothesis that the correlation between level of occupational aSpira- tion and grade point average will be significantly higher for students with a high perception of the necessity of college than for students with a low perception of the neces— sity of college must be rejected. Several observations in the data must, however, be noted. First, the difference (although not statistically 103 significant) is in the expected direction. Second, the group with "Medium" perception of the necessity of college shows a higher positive relationship between level of occupa- tional aSpiration and grade point average than the ”High” group. Theoretically, the relationship would be expected to fall somewhere between the ”High” and ”Low” groups. Thus, it deviates from the expected pattern. Third, the correla— tion coefficients for both the ”High” and ”Medium” groups are Significant from zero. Fourth, the correlation coeffi— cient for the ”Low“ group is not statistically significant from zero. Finally, a limitation of the sample noted earlier (Chapter III, p. 86) should be emphasized, that is, a severe lack of respondents who were at the low end of the perception of the necessity of college index.9 We shall consider these observations further after testing the remain— ing sub—hypotheses of Hypothesis 1. Sub—hypothesis 1.5 When the past performance is high, the correla- tion between occupational aspiration and grade point average will be significantly higher than when the past performance is low. The ”2” test was not applied to the null hypothesis that students with high past performance will have a correlation 9It is found, in fact, that less than 2% of the respondents actually perceive the necessity of college for attaining their occupational aspiration as less than "quite important.” More than 90% of the reSpondents perceived it as ”very important" or “absolutely necessary." 104 coefficient between occupational aspiration and grade point average equal to or Smaller than students with low past performance because the observed values of ”r” show that a Significance of difference will not obtain, and that the very slight difference is opposite the prediction. High r = .137 N = 206 (Medium r =.088 N==192) Low r = .145 N = 184 Since it is apparent that the correlation coefficients are not different, the hypothesis must be rejected. Nevertheless, several possible sources of this rejection may be suggested. First, it was argued in Chapter III that past performance constitutes an indirect self con— cept of ability measure, since the student tends to use it as a primary objective measure of ability. There is some 10 If this argument is empirical evidence to support this. valid, it is possible and highly probable that the student uses this self concept of ability to adjust his aspirations in a consistent manner or direction, including his occupa- tional aspiration. Thus it may be suggested, on the basis of the obtained statistics and the rationale presented, that the person adjusts his occupational aspiration and self 10See page 55 of this thesis; also see footnote 26 of Chapter III. 105 concept of ability consistently, or in like manner. This is also consistent with the notion of ”cognitive balance” discussed in Chapter II.11 The student, in order to main— tain a cognitive balance and prevent a condition of cogni- tive stress (dissonance) between aSpiration and self concept, must bring his occupational aspiration and self concept of ability into a consistent cognitive co—existence without stress. If this is so, the observed values of "r” are reasonable Since differential self—concept of ability would not require Significantly different relations between aspira- tion and grade point average. Second, it is possible that past academic performance as an indirect self concept of ability measure, particularly when it refers to high school performance, undergoes considerable and significant change during the first year of college and thus may be considered as a possible source for the observed rejection of the hypothesis. For example, a considerable number of students who performed satisfactorily or even outstandingly in their high school environment (and thus their community environ- ment)may not do so (or necessarily be expected to do so) under the conditions of the university. For example, it is highly probable that quite different standards of performance, l o . cit. 1See pages 21—23 of this thesis; also see Heider, 106 different referents for comparison and evaluation of perfor— mance, and different conditions of and requirements for the performance process prevail for the student in the univer- sity environment as compared to his high school environment. Thus, between the time of measurement of this facilitation variable and the consequent behavior variable (approximately two and one-half years), the self concept of ability may undergo drastic changes due to these vast changes in the student's environment, both intra— and extra-personal. If this is so, the past academic performance variable may very well have other than the expected impact as a facilitator upon the relationship between the attitude and behavior variables. Third, another possible although improbable source contributing to the rejection of the hypothesis may be sought in the fact that the reliability of the past academic performance index (the individual‘s high school grade point average) cannot be adequately determined and the validity can only be estimated by its relationship with the behavioral variable (the individual's college grade point average) in this research. It must, of course, be recog- nized that this problem is not new or unique; it must be faced whenever grade point average is used as an independent variable index. Thus it is cited here only so that the reader may evaluate the observed rejection with this aware- ness. 107 The foregoing discussion cites several possible sources of the rejection of this sub-hypothesis. The valid- ity of any or all of the rationale must, of course, be dem- onstrated empirically. Sub-hypothesis 1.6 When general facilitation is high, the correlation between occupational aspiration and grade point average will be significantly higher than when general facilitation is low. In testing the null hypothesis that the ”High” facilitation group would have a correlation coefficient between level of occupational aSpiration and grade point average equal to or smaller than the ”Low" facilitation group, the following results obtained. High r = .097 N = 169 z' = .0973 (Medium r= .035 N=l86) Z=l.74 P< .05 Low r=—.095 N=159 z'=-.0953 The ”2” test of significance of difference indicates that the null hypothesis may be rejected. Therefore, the hypoth- esis that high facilitation students have a Significantly higher correlation between their occupational aspirations and grade point averages than do low facilitation students is tenable as proposed. This finding, then, tends to sup— port the fundamental position that the degree to which level of occupational aspiration and grade point average are re— lated varies directly (or consistently) with the degree of facilitation offered. In theoretical terms, it supports the 108 notion that the degree to which attitudes and behaviors are related varies directly with the degree of facilitation offered by the attitude context. In Spite of the tenability of this sub—hypothesis, the negative correlation which obtained between level of occupational aSpiration and grade point average for the "Low" facilitation group is worthy of consideration. Al- though the correlation for this low group was expected to be zero, or extremely close to zero, the possibility of a nega- tive correlation was not anticipated. An explanation, based in part upon the theoretical framework and in part upon the nature of the sample, may be ventured. It is possible and quite probable that students with generally low facilitation for college attendance and academic performance, but who have relatively high occupational aspirations, experience rather extreme frustrations and cognitive dissonances with some regularity during their stay at college. Thus, their academic performance is hindered, or at least it is not con— sistent with their relatively high occupational aspirations.12 On the other hand, students with a degree of general facil— itation more consistent with their relatively high 12It was noted that even the “Low" group of students on level of occupational aspiration have relatively high degrees of attitude strength, that is, level of occupational aSpiration. There were no really low aspirers in the sample (see pages 86-88, also Appendix G). 109 occupational aSpirations are not experiencing the frustra- tions of the former group, at least not frustrations of the same type or degree of intensity and regularity. Thus, they are not hindered as greatly in their academic performance, or in any event, their academic performance is likely to be more consistent with their occupational aSpirations. If this in fact occurs, a negative correlation between level of occupational aspiration and grade point average for the "Low” facilitation group would be expected. This explanation should, of course, be pursued in further research more Specifically directed at this problem before it is even ten- tatively accepted as valid. However, the theoretical and practical implications of such a finding would be signif— icant. The tests of the foregoing sub—hypotheses do not conclusively confirm the general hypothesis, although the bulk of the evidence tends to support it. Certainly, the evidence is not entirely clear. Two of the individual facil— itation variables (1.2 and 1.3) and the general facilitation variable (1.6) supported the general hypothesis by statisti- cally significant differences which were expected. Three of the individual facilitation variables (1.1, 1.4, 1.5) did not Show statistically significant differences which were expected in support of the general hypothesis, although two of them (1.1 and 1.4) Showed differences in the expected 110 direction. Thus, the expected significant difference which obtained when the general facilitation variable (an additive combination of the five individual facilitation variables—- see Appendix E) was controlled is apparently a function of the two individual variables showing statistically signif- icant differences and the two other individual variables showing non—significant differences in the expected direc— tion. Two general observations derived from the tests of these sub—hypotheses must be reviewed and their importance evaluated. First, it was noted that in the case of the two facilitation variables where non—statistically significant differences in the expected direction were observed, one (1.1) showed a correlation coefficient statistically signif- icant from zero for only the ”High" facilitation group, the other (1.4) for both ”High” and "Medium” facilitation groups. Furthermore, if we observe the two facilitation variables where statistically significant differences in the expected direction were observed, we also find that one (1.2) showed a correlation coefficient statistically significant from zero for only the ”High” facilitation group, the other (1.3) for both the ”High” and ”Medium” facilitation groups. These observations suggest that the facilitation variables uti— lized here vary in reSpect to the level or threshold at which they tend to have an impact upon the relationship between level of occupational aspiration and grade point 111 average. In other words, this tends to suggest that facil— itation variables vary in reSpect to the point or points on any index measuring them at which they ”make a difference” in the relationship between the attitude and behavior. For example, on a mental ability index the evidence leads one to posit that possibly only variations above or below a certain point make a considerable difference in the relationship; or possibly that variations between two points make little dif- ference but that variations either above or below these points make considerable differences in the relationship. In specific, this suggests that a simple linear relationship does not exist between these variables and, thus, that sta- tistical techniques which assume linearity are inadequate for observing the relationships which do exist. What this implies, then, is that more complex, non—linear relation— ships may be occurring between these variables which are not made clearly evident from the correlational analysis used here. If this is so, it suggests that a modified multiple cut—off method must be combined with a type of non-linear multiple regression technique to obtain the optimum effect of a combination of facilitation variables upon an attitude- behavior relationship. Further consideration to this prob- lem and its implications for the theoretical framework guiding this study will be given in a later section of this chapter. 112 Second, the negative correlation observed for the "Low" facilitation group in sub-hypothesis 1.6 and the rationale provided for it are suggestive of another phenom- enon with possible theoretical implications. The theoret- ical framework utilized presently does not lead one to expect a negative relationship between level of occupational aSpiration and grade point average, that is, between atti— tudes and behavior. It is possible, however, that under certain conditions a negative relationship should be ex- pected. For example, a student with a relatively high level of occupational aSpiration but who has a low degree of facil- itation for college performance (i.e., low mental ability, little reference group support, relatively low socioeconomic status, etc.) probably experiences a degree of cognitive dis- sonance and, furthermore, severe frustrations in his efforts toward satisfactory college performance. These frustrations and cognitive dissonances may, as separate mediating vari— ables, further hinder his academic performance to the extent that it is negatively correlated with his level of occupa- tional aSpiration; at least to the extent that it is not consistently related to either the level of occupational aspiration or the controlled facilitation variables. If this is so, the implications for the theoretical framework are obvious and significant. We shall further consider this problem and its possible theoretical implications in a later section of this chapter. 113 In summary, it is felt that on the basis of the evidence the general hypothesis from which the six sub— hypotheses were derived is supported. It was found that the test of the general hypothesis was statistically significant in the expected direction. Tests of three of the sub—hypoth- eses, including the general facilitation variable which constituted an index constructed by utilizing the other five individual facilitation variables, also showed statistical significance in the direction which was expected. Although three of the sub—hypotheses did not Show differences which were statistically significant, two of them showed differ- ences in the expected direction. Only one sub-hypothesis did not show an expected difference. In this case, and in the two cases where the null hypotheses could not be re— jected, the necessarily tentative rationale (subject, of course, to empirical verification) which was presented for the obtained observations is accepted as adequate. Thus, the general hypothesis and, in turn, Postulate l of the theoretical framework, are acceptable as tenable. This decision must, of course, be made tentatively and cautiously. The implications of this decision will be discussed further in a later section of this chapter. 114 Hypothesis 2 The second general hypothesis of the study is that the degree of facilitation offered by the attitude context is positively correlated with level of performance in col— lege for individuals who have some degree of positive occupa- tional aSpiration. Six sub—hypotheses are formulated from this hypothesis which, in turn, are tested by controlling the level of occupational aspiration. The testing of the general hypothesis and the sub—hypotheses constitute the test of Postulate 2 of the theoretical framework. Hypothesis 2 When individuals differ in the degree of facilita— tion offered by their attitude contexts (or any one facilitation variable), and when their levels of occupational aSpiration are greater than zero, there will be a positive correlation between the facilitation offered by the attitude context (or any one facilitation variable) and their levels of performance in college. The test of this general hypothesis was completed by corre- lating each of the selected facilitation variables with level of performance in college (grade point average)° A general index of facilitation was also constructed from the five variables and correlated with level of performance in college.13 In each case the null hypothesis of no relation— ship was tested by application of the ”t” test of 13See Appendix E. 115 . . . l4 Significance. Mental Ability r=.542 N=584 t=15.56 df=582 P<.01 Perceived Reference Group Support r=.323 N=559 t=8.04 df=557 P<.01 Socioeconomic Status r=.062 N=552 t=l.45 df=550 P<.05 Perception of the Necessity of College r=.OO6 N=546 1:30.14 df=544 P LOAIn Fm > LOAl F1 is statistically significant at the .001 level of probabil- ity. The realization of this expected relationship tends to substantiate the utility of the variables used in this thesis. In regard to persistence in college, assuming the hypotheses of this thesis are tenable, one may make the pre- dictive statement that students with low occupational aspira- tion and low facilitation will drop out of college more fre- quently than will students with high occupational aspiration and high facilitation. More generally, it may be predicted that the rate of dropout will vary inversely with the com- bined level of occupational aspiration and degree of facil— itation; or conversely, that the rate of persistence in col- l€ge will vary directly with the combined level of occupa- tional aSpiration and degree of facilitation. Table 4 shows Iabl C) ?‘ O M (‘0 :3‘ (I) Tim U) (‘2) 145 Table 4. Percent of Dropout and Non—Dropout for Major Com— bined Categories of Occupational ASpiration and General Facilitation Dropout Status Low OA and GF Medium OA and GF High OA and GF Non Dropout 42.9% 67.9% 81.2% Dropout 57.1% 32.1% 18.8% Number of cases 77 81 80 x2 = 25.92 df = 2 P< .001 the data obtained which are also consistent with this predic- tion. Note, for example, that only 18.8% of the "High” group dropped out of college while 57.1% of the ”Low” group dropped out. The significance of the differences shown in Table 4 was determined by application of the Chi—square (X2) test.27 A test of the null hypothesis of no difference was completed which allows rejection at the .001 level of probability. Thus, it is obServed that the expected relationship with re— gard to persistence in college is also statistically signif— icant. It also, therefore, lends support to the utility of the variables used in this thesis. 27The formula for the computation of Chi-square is that presented by Blalock: 2 2 -.:E;(fO-fe) 4’ “ f. See Hubert M. Blalock, Jr., op. cit., p. 213. 146 The behavioral consequences of other combinations of degrees of occupational aSpiration and facilitation may be observed in Appendix 1. Except in two instances, all empir- ical behavioral differences tend to be in the direction ex- pected. In the two instances cited, however, the variations from the expected are not statistically significant. In one instance, that of the "Medium” facilitation group, where the mean grade point average would be expected to vary directly with the level of occupational aSpiration, it is found that the mean grade point average is highest for the “Medium” rather than for the ”High” occupational aspiration group. Since the differences are extremely small, the probability of this occurring by chance appears to be relatively high, and thus the variation is considered to be spurious. The other instance of variation from the expected direction, although not statistically significant, may be more seriously consid- ered° In the case of the ”Low” facilitation group, the mean grade point average varies inversely with the level of occupa- tional aspiration rather than directly as would be expected theoretically. Two explanations for this variation may be ventured. First, since it does not approach statistical Sig- nificance, the possibility of a Spurious variation must be considered. Second, it may be proposed that when occupation— al aspiration is relatively high but facilitation is low, goal—frustrations and dissonance-producing experiences are 147 both frequent and intense and thus operate as major deterrents to high level of college performance as well as enhance the rate of withdrawal from the situation (dropout rate). Since persons with low levels of occupational aspiration would not have these experiences, at least not with the frequency and intensity of the medium and high aspiration persons, their levels of performance are likely to be equal to or higher, and their rate of withdrawal equal to or lower than the higher aSpiration individuals. Thus, it is found that the mean grade point average for the higher aspiration but low facilitation group is in fact somewhat lower than for the low aspiration and low facilitation group. Likewise, it is found that the rate of dropout for the high aspiration but low facilitation group is slightly higher (61.3%) than for the low aspiration and low facilitation group (57.1%). The feasibility of this latter explanation should of course be pursued in further research, since the theoretical implications are both obvious and Significant. (It should be noted that these data and the accompanying argument are similar to, consistent with, and supportive of the proposed rationale for the negative correla- tion coefficient which obtained in sub-hypothesis l.6--See pp. 108-112 of this thesis.) In Summary, the analysis of behavioral consequences demonstrates empirically the expected differential behavior which may logically be predicted from the variables utilized 148 in testing the hypotheses of this research. In brief, the expected differential behaviors did obtain from the appro- priate differential levels of occupational aspiration and degrees of facilitation. Thus, the variables employed in the operational framework of this thesis should also have contin- ued utility in the further pursuit of a multivariate non— linear statistical model (which was suggested by testing the hypotheses of this thesis) for the optimum prediction of dif- ferential human behavior from differential attitudes and facil— itation. CHAPTER V SUMMARY AND CONCLUSIONS This chapter of the thesis summarizes the observa— tions of this research within the theoretical framework upon which the problem is based. It will include a general sum— mary and evaluation of the empirical findings, some limita— tions of the study and suggestions for future research. Finally, some probable or at least tentative theoretical contributions and practical implications will conclude the chapter and the present study. Summary This thesis began with a theoretical framework from which three basic postulates regarding the relationship be— tween attitudes and behavior were obtained and stated (Chap- ter 11). Three general and testable hypotheses were derived from the postulates to be tested within the operational frame- work and research design of this study (Chapter 111). These general hypotheses were then expanded to include several more specific sub-hypotheses dealing with very Specific variables. Tests of the hypotheses thus constituted preliminary tests of the adequacy of the postulates. 149 150 The hypotheses and sub—hypotheses were tested against data collected from 585 new first-time freshmen stu- dents at Michigan State University in the fall of 1960. Additional data about these students were subsequently col- lected in the winter term of 1963. By the use of several Specific instruments designed to elicit appropriate informa- tion from each student, and with access to official Univer— sity records for obtaining other necessary information, in— sights into the nature of the relationship between level of occupational aSpiration and level of performance in college were pursued. In brief restatement, it was theoretically posited that an attitude may result in behavior toward an object and that it may also, in turn, result in behavior toward means of fulfilling specific object behavior. For a specific example, level of occupational aspiration (attitude) would be expected to result in occupational achievement (behavior) with reSpect to the occupational prestige heirarchy (object); and it also would be expected to result in educational achieve- ment (means behavior) in the educational structure (means structure essential for placement in the occupational hier- archy). Because the behavioral consequences of an attitude always occur in situations, attitudes are differentially facilitated in their expression in behavior by the context of the attitude. Again, for a specific example, level of 151 occupational aspiration (attitude) is differentially facil- itated in expression in educational behavior (means) by contextual factors such as perceived reference group support, mental ability, self-concept and the like. The postulates of the theoretical framework are, therefore, structured around three types of variables: attitudinal, facilitational, and behavioral. They predict certain interrelationships be- tween these variables. The postulates are: 1. When individuals differ in their attitudes toward an object, and when the facilitation offered by the attitude context (or any one facilitation variable) is greater than zero, their attitudes will be positively correlated with their behavior with reSpect to the object and with reSpect to objects viewed as means for promoting their desired behavior toward the object of the attitude. a. When the degree of facilitation offered by the atti— tude context (or any one facilitation variable) is high, the positive correlation will be significantly higher than when the degree of facilitation offered by the attitude context (or any one facilitation variable) is low. 2. When individuals differ in the degree to which their attitude contexts facilitate the expression of an attitude in behavior, and when their attitude strengths are greater than zero, there will be a positive correlation between the facilitation offered by the attitude context (or any one facilitation variable) and their behavior with re- spect to the object, and with the reSpect to objects viewed as means for promoting their desired behavior toward the object of the attitude. a. When the attitude strength is high, the positive correlation will be Significantly higher than when the attitude strength is low. 3. When the degree of facilitation offered by the attitude context (or any one facilitation variable) varies among in- dividuals, there will be a positive correlation between the degree of facilitation offered by the attitude context (or any one facilitatidn variable) and the attitude strength. 152 The empirical findings of the present study appear, on the whole, to substantiate the adequacy of the postulates. The three general hypotheses, derived from the postulates, were found to be tenable as stated and thus constitute a test of the adequacy of the postulates. The data show that Hypothesis 1 is tenable, and it has been tentatively accepted. These data indicate that level of occupational aspiration is positively correlated with level of performance in college for individuals who in some degree are positively facilitated by the environment within which they operate physically, socially and psycholog- ically. The evidence from the tests of the sub—hypotheses is not entirely clear and conclusive; however, the bulk of the evidence tends to support them. The tenability of the sub- hypotheses tends to suggest that the degree to which level of occupational aspiration and level of performance in college are correlated varies directly with the degree of facilitation obtained. In theoretical terms, then, the data indicate that attitudes are positively correlated with means behavior and that the degree of correlation tends to be directly related to the degree of facilitation provided by the physical, social and psychological environment of the individual and his attitude. This observation provides direct support for Postulate l and constitutes a test of its adequacy. It must, of course, be emphasized that these data can only partially 153 test the adequacy of the postulate because the present re- search design deals with means behavior, and presumes a similar relationship of attitude with object behavior. Although there is substantial evidence to suggest that this presumed relationship does exist, empirical data are neces- sary to substantiate it and to complete the test of adequacy of the postulate. Obviously, if the expected relationship of attitude and object behavior does not obtain empirically, the present theoretical framework must be thrown into Serious question. Hypothesis 2 is similar to the first hypothesis but is focused upon the relationship of facilitation and behavior. The data again indicate that it is tenable, and it has been tentatively accepted. The data show that the degree of facilitation is positively correlated with level of perfor- mance in college for individuals who have some degree of positive occupational aspiration. Furthermore, the general support of the sub—hypotheses indicates that the degree to which these variables are correlated varies directly with the level of occupational aspiration. In theoretical terms, the data indicate that means behavior is positively correlated with facilitation and that the degree of correlation tends to be directly related to the strength of the attitude toward the object. This finding Supports directly Postulate 2 and constitutes a test of its adequacy. For the same reasons as 154 indicated in Hypothesis 1 above, however, these data consti- tute in fact only a partial test of the adequacy of the postulate. The third hypothesis, focusing upon the relationship of facilitation and attitude, was also found to be tenable by the present data and is accepted on that basis. The data indicate that a positive correlation between degree of facil- itation and level of occupational aspiration does exist for individuals who are differentially facilitated to some degree. In theoretical terms, facilitation is positively correlated with attitude strength for individuals with some degree of facilitation. This observation provides direct support for Postulate 3 and constitutes a test of its adequacy. As in the previous hypotheses, this conStitutes in fact only a partial test of the postulate since our present research design deals only with means behavior, and theoretically assumes that a direct ends—means relationship does exist between the occupational prestige (or socioeconomic) hier- archy and the formal educational structure in our society. The latter assumption is made, of course, in light of consid— erable and substantial evidence in support of it. But empir- ical data in direct Support of it are still needed. The present study, in its support of the above postu- lates through establishing the tenability of the derived - hypotheses, substantiates the theoretical position that there 155 is considerable interaction and interrelationship between the three types of variables essential to it. The observa— tions of this study provide empirical support for the exis- tence of these interrelationships and preliminary insights into the more specific nature of them. Obviously, a con- siderable amount of additional research is needed to further clarify some aspects of these interrelationships, and thus eventually raise the Optimum predictability of human behavior in this framework. The relatively low correlation coefficients observed in this study may require discussion. There are several reasons why high correlation coefficients were not to be anticipated. The first reason was pointed out in some detail in an earlier section of the thesis.28 It may be briefly stated as a limitation of the sample evidenced by a relatively limited range of variability on several critical independent variables. The nature of the sample, a college group, was such that this limited range of variability could be logically expected, and thus also the lower correlation coefficients than might be expected from a sample with a broad range of . . . 29 . . . . variability. It was also to this limitation of the sample 28See Chapter III, ”Limitations of the Sample,” pp. 85—90 of this thesis. 29See Blalock, op. cit., pp. 290-291. 156 that several statistically non—significant results were attributed. Second, the fact that the research design used in this study utilizes a means behavior variable rather than the object variable makes it only logical to expect somewhat lower correlation coefficients than if the latter had been used. The logic of this expectation must of course be found in the imperfections of the means—ends relationship between educational achievement and occupational achievement. Even though this means—ends relationship in our society is ex- tremely rigid and tenacious, there is obviously some degree of freedom within the means-and object—structure relationship for selected individuals. Thus, the necessarily imperfect nature of the means—ends relationship probably tends to deflate the correlation coefficients obtained. Third, the lapse of time between measurement of the attitude variable (level of occupational aspiration) and the behavioral variable (grade point average) was considerable (about two and one-half years), thus allowing for consider— able change or adjustment both in the attitude and in the conditions of facilitation. An almost infinite number of reasonable examples of this type of change could be cited. A few may suffice. A student quite conceivably may change some of his reference groups while in college, or at least additional ones are added, or their hierarchy of importance in his perception may be restructured. High school peers, teachers, and counselors probably become less important while college groups and associations become more important as ref— erence groups and individuals. Also, the first term or sev— eral terms of academic performance may change the student's self concept of ability; a change in conditions of facilita- tion which may have an effect upon the attitude, that is, the level of occupational aSpiration. Furthermore, new associa- tions and new exposures to information about the occupational world provided on the college campus may cause changes and adjustments in the previously measured attitude. In any event, it is obvious from these few examples that changes in attitude and conditions of facilitation must function to reduce the size of the correlation coefficients obtained. Fourth, there is an almost infinite number of pos— sible facilitation variables which may be operative and func- tional. Only several of these were measured and controlled in this study. The failure to measure the full set of facil- itators probably also tends to deflate the correlation coeffi- cients. For the above reasons, then, it would have been logically unrealistic to expect high correlation coefficients in this study. Finally, in addition, the tests of the sub—hypotheses of Hypothesis 1 in the present study have presented some evidence to at least tentatively suggest that statistical 158 techniques which assume linear relationships (thus including the correlation coefficient) between the variables are prob— ably not adequate to observe the relationships which do in fact exist between them. Of course, if this is so, it would constitute another reason for low correlations to obtain between the variables when utilizing this theoretical frame- work. From the theoretical position of this Study, it would be contended that except for differentially facilitating factors in the attitude context, an attitude will always be carried into behavior. The critical role of the attitude context, and thus the facilitation variables, in the present theoretical framework is then obvious. The data of this study tend to suggest that the Optimum control of facilita- tion is necessary if the attitude-behavior relationship is to be scientifically pursued and eventually understood more clearly. The optimum control of facilitation, however, ap- pears to be a most difficult and complex task in the design of research. We may briefly evaluate the present study in this regard and possibly gain some insight into the problem. The facilitation variables used in this study were selected on the basis of their hypothesized relevance as differential facilitators to behavior in respect to the means structure of the object of the attitude. In Specific, they were selected on the basis of their hypothesized relevance 159 for differentially facilitating level of performance in college. Unfortunately, the logical basis of such hypothe— ses is unclear. We generally select such variables on the basis of ”sociological experience.” We do not have a set of general rules for selecting all possibly relevant facil— itation variables worthy of consideration in any given prob- lem. The primary reason this is so appears to be the fact that facilitation variables relevant for one action variable are not necessarily relevant for another. Thus, the best method now known seems to be the selection of facilitation variables on the basis of their logical relevance to the object of the attitude or its means structure, depending upon the Specific problem. The ”logical relevance" method for the selection of facilitation variables was relatively successful in the present study. Empirically it was found that four of the five selected facilitation variables were worthy of consid— eration in the operational framework used in this study. The perception of the necessity of college variable was found to be inadequate for two reasons. First, the apparent homogeneity of the sample in regard to its perception of the necessity of college caused all reSpondents to score high on the instrument used to measure this variable. Second, and probably a result of the first reason, failure of the instru- ment to empirically discriminate between the respondents of 160 the sample gave this variable minimal utility. The other four facilitation variables-—socioeconomic status, mental ability, past academic performance (self concept of ability), and the perception of reference group support--tended to behave predictably as differential facilitators in the atti- tude context and should be considered in future research with a similar operational framework. However, several problems in regard to the attitude context must be resolved if future research is to make continued profitable inroads into the better understanding of the attitude-behavior rela- tionship. These may be discussed both as limitations of the present study and as Suggestions for future research. Limitations of the Study and Suggestions For Future Research A number of limitations of the present study and possibilities for future research have already suggested themselves or have been mentioned previously in the study. However, several Specific suggestions arising out of these limitations may be discussed briefly. These constitute logical extensions of the present study. First, the present study leads one to contend that the attitude context must be more rigidly controlled in future experimental designs. It seems extremely necessary that we place much more rigid controls on the attitude con- text so that it becomes much more feasible to isolate all or 161 most of (l) the relevant facilitation variables and (2) the factors possibly effecting change in the attitude variable prior to measurement of the behavioral variable. This is a difficult task. In the present study, one may contend that countless other possible facilitation variables could have, and possibly should have, been controlled. It is obviously true that many others could have been Operative, even though not logically relevant and thus difficult to isolate and deem worthy of control when structuring the research design. Similarly, one may contend that the present study allowed for many factors of change in both facilitation and attitude to be operative between the time of attitude measurement and behavior measurement. This also is obviously true, even though we have no empirical measure of this kind of change and what effect it may have had upon the results. In any event, both of these contentions reflect less control Of the attitude context than would have been desirable. A more parsimonious research design allowing for more rigid control Of the attitude context should contribute additional in- sights into the nature of the attitude—behavior relationship. Second, once the facilitation variables worthy of consideration in a Specific research design are selected and isolated, a more complete analysis Of the nature of the rela— tionship between them is required so that a more satisfactory means of combining the effects of each in a prediction equa- tion may be developed. The present study utilized a simple 162 additive combination of these variables in the development of a general facilitation index. Although this is a satis- factory preliminary technique, there are at least two reasons why it lacks the potential of producing optimum predictive efficiency. One, there is good reason to believe that not all facilitation variables have equal importance as facilitators of a given attitude toward its appropriate behavioral response. In fact, some minimum degree of facil- itation from a given variable may be absolutely necessary if the appropriate behavioral response is to occur.30 Two, the present study suggests that facilitation variables may vary in reSpect to the level or threshold at which they ”make a difference” in the relationship between the attitude and behavior.31 Thus, linear relationships probably do not exist between facilitation variables, or between their 0For example, mental ability may be crucial for academic or occupational performance; without a certain amount of it, persons could not be able to perform; but, beyond a certain point, additional increments of mental ability may not be associated with similar increments in performance. If this is so, a multiple cut—off method similar to that used for occupational norms on test bat— teries by the U.S. Employment Service Should be used. A minimum or critical score is established on each Significant aptitude. See Donald L. Grant (ed.), "Validity Information Exchange," Personnel Psychology, 1955, Vol. 8, p. 105. 31For example, on a mental ability index probably only variations above or below a certain point make a con— siderable difference in the occupational aSpiration-perfor- mance in college relationship; or possibly variations between two points make little difference but variations either above or below these points make considerable differences in this relationship. 163 combined effects and the attitude. The Simple additive combination of facilitation variables into a general index takes neither of these facts into account and thus may lose much precision in the act Of combining. However, on the other hand, there are also at least three reasons to suggest that a simple multiple regression technique for combining facilitation variables is not satisfactory in this framework. One, multiple regression yields a composite index based on a combination of several variables and con- sequently also throws away much information by reducing several facilitation dimensions to one.32 Two, multiple regression weights permit a high degree of one facilitation variable to compensate for a low degree of another possibly more critical facilitation variable. Three, the multiple regression method assumes linearity and takes no account Of the possible optimum level of each of the variables. An equation utilizing the multiple cut—Off method,33 or a modification thereof, in combination with a non-linear multiple regression technique should be developed in an attempt to maximize the impact or effect of facilitators upon the attitude—behavior relationship. A computer program 32See Eugene L. Gaier and Marilyn C. Lee, "Pattern Analysis: The Configural Approach to Predictive Measurement," Psychological Bulletin, 1953, Vol. 50, pp. 140-148. 33Beatrice J. Dvorak, ”Advantages of the Multiple Cut—Off Method," Personnel Psychology, Vol. 9, 1956, pp. 45-47. 164 could conceivably be develOped to do this, but it should do more than this. One additional task should be incorporated. In addition, the computer program should be SO developed that a predictive equation in which the Optimum predictive efficiency of behavior on the basis of some combination of attitude and facilitation is Obtained. Presently one would expect this latter combination to be accomplished by some type of non-linear correlation technique. In summary, then, a computer program should be developed to simultaneously (1) determine scale points or scale areas on the index meas— uring each controlled facilitation variable at which it has optimum impact upon the relationship between the attitude (independent variable) and the behavior (dependent variable), (2) determine the relative importance of each of the con- trolled facilitation variables as opposed to the others and assign appropriate weights to each for optimum combined im— pact Of the facilitators upon the relationship between the attitude and the behavior, and (3) relate the attitude vari— able and the combined facilitators to Obtain the optimum predictive efficiency Of behavior. The development of such a computer program.and its subsequent utilization could con— tribute many new insights into our present problem. Thus it would also contribute much to our present knowledge of the nature of the attitude—behavior relationship. 165 Finally, measurement techniques must continually be improved so that the sophistication of the data analysis may be enhanced. This must be the Objective of the researcher in this field, and of social psychologists generally, if valid contributions to knowledge are desired. In our pres- ent state of knowledge, limitations of measurement must always be considered in evaluating data analysis. The pres— ent study has utilized several rather crude measurement devices, particularly in the measurement of facilitation variables, which should be considered in the evaluation of the results of this study and be improved by future research— ers utilizing these or similar variables. This problem of measurement, of course, is not unique to this study. Since measurement lies at the very heart of social psychology, continued improvement in measurement techniques is imperative. In Spite of the several limitations of this study, it is felt that it has demonstrated generally that the predicted relationships between attitudes and behaviors, particularly means behaviors, do exist and that the theoretical framework guiding this research is worthy of continued consideration in the development of further empirical pursuits and theoret— ical schemas. Although a considerable amount of additional research is needed to clarify some of the aspects of the relationships Observed, it is felt that this initial study has a number of contributions and implications for social 166 psychologists, and behavioral scientists generally, who are concerned with the analysis and understanding Of this rela- tionship between attitudes and behaviors. Theoretical Contributions It is felt that there are several theoretical con- tributions made by the present study. These contributions are, Of course, tentative insofar as much additional evi— dence is necessary. But it can be legitimately posited, again emphasizing that the findings of this study must be interpreted with particular caution, that the following contributions hold potential validity for the eventual devel— Opment of a more comprehensive theory Of differential behav— ior. The most general theoretical contribution is that the present study appears to provide empirical evidence for the theoretical framework upon which it is based. Thus, the major part of Chapter II of this thesis constitutes a con- siderable contribution. However, more Specific contribu- tions to this general theoretical framework can be isolated on the basis of the data obtained and analyzed in the study. First, the present study emphasizes and apparently substantiates the very critical role which the attitude con— text plays in the fundamental attitude—behavior relationship. The evidence presented in this research suggests that the facilitation variables operative within the attitude context 167 are major factors to be considered when explaining why atti— tudes are not always consistent with overt action behavior. It thus suggests that systematic concern with the attitude context is essential if new insights into the nature of the attitude—behavior relationship are to be Obtained and if significant strides toward the development of a valid theory of differential action are to be made. This systematic con— cern with the attitude in relation to the context in which it is enacted in overt behavior should also provide addition— al knowledge which could change some past as well as current conceptions and theories regarding attitudes generally.34 This, in turn, would have considerable impact upon present sociological and social psychological thinking. Second, the present study provides some tentative suggestions regarding how the two key types of variables (attitudinal and facilitational) of the theoretical frame- work may be combined for optimum prediction of differential behavior. Although future research must determine the util— ity of it, the present study tends to suggest a general "formula.for combination Of attitudinal and facilitational variables for optimum predictive efficiency of overt behav- II ior. The generalfformula may be stated as follows: 34 . For a reView of some of these, see De Fleur and Westie, 1963, op. cit. 168 B = (f)A[Flwl + sz2 + F3w3 ... ann] where B = overt behavior toward the object or means object of the attitude A = attitude toward the object F = each facilitation variable "optimally scaled" in respect to its impact on the attitude-behavior relationship w = the "weighted importance” of each facilitation variable in comparison with others in respect to the impact it has upon the attitude-behavior relationship and where each F is optimally scaled. Hopefully, the exact nature of the ”formula” should be better determined on the basis of future research utilizing a computer program similar to that suggested and discussed earlier. Fundamentally, how- ever, the general ”formula" hypothesizes non—linear relation- ships between the variables, and thus implies that non—linear statistical techniques, rather than the more traditional linear techniques, must be developed and utilized in empirical research if theoretical advances of significant consequence are to be accomplished. Third, this research provides empirical support, although tentative, for the existence of the means—ends relationship of various objects in respect to given attitudes. The data suggest that there is a means-ends relationship be- tween educational achievement and occupational achievement in‘ American society on the basis of the observation that differ- ential academic performance may be generally predicted from 169 level of occupational aspiration (attitude toward the occupa— tional hierarchy) used together with facilitation variables. This, in turn, supports the notion that the educational and occupational structures in American society stand in a means- ends relationship in which persons’ positions in one struc- ture differentially determine or allocate their positions in the other. Theoretically, then, these data tentatively posit that where a means-ends relationship does exist between two or more objects, overt action behavior toward any one may be differentially predicted on the basis of attitudes (used to— gether with facilitation variables) regarding any one. Finally, the data of this study validate the notion that there is a considerable amount of interaction between the three types of variables basic to the theoretical frame— work: attitudinal, facilitational, and behavioral. Although the present study only begins to describe the nature of these interrelationships, it seems to establish the fact that they do exist and that a great deal more must be learned about them if the theoretical framework is to provide maximum utility in future empirical and theoretical pursuits. Par— ticularly in this regard, the observations of the present study should be hypothesis-producing for future studies. 170 Practical Implications The theoretical observations of the present study suggest several practical implications which may be briefly mentioned here. The data utilized in this study relate directly to such practical concerns as student guidance, counseling and selection for college. Thus, the practical implications are oriented in the direction of these facets of the American educational system. The present study lends support to the recent inter— est in the predictive value of ”non-intellectual” factors in academic achievement, particularly at the college level.35 The apparent role of occupational aspiration in determining differential academic performance, portrayed in this study, is evidence of one such factor. Furthermore, the signif- icant role of the perception of reference group support as a facilitating variable in this study further suggests the relative importance of sociopsychological factors as differ— ential determinants of academic performance. The continued use of these types of factors in research should enhance our 35See Donald R. Brown, ”Personality, College Environ— ment, and Academic Productivity,” (Chapter 16) and Joshua A. Fishman, ”Some Social-Psychological Theory for Selecting and Guiding College Students," (Chapter 20) in The American College, (edited by Nevitt Sanford), op. cit. 171 present ability to predict success or failure in college and eventually lead to more practical application of this knowl— edge in college and university selection practices and pro- cedures as well as in junior high school and high school counseling and guidance programs. Selection practices in most colleges and univer- sities today consist Of the consideration of some combina- tion of intellectual factors (aptitude scores, achievement scores, and high school rank and grades) and personal assess- ments or appraisals (recommendations, personality factor rat— ings, and personal interviews) by several persons (high school principal, high school counselor, admissions officers, etc.) who are supposedly knowledgeable about either the stu— dent or some clinical methods of appraisal. It is interest— ing to note that these methods, although varying from time to time and from one institution to another, have changed basically not at all over the past half—century. Further- more, literally thousands of studies (mostly under the rubric of ”predictive” studies) have been completed using these vari— ables, particularly the intellectual variables. Again, it is interesting to note that the extent of the relationship between these variables and college performance has remained essentially the same for all institutions for the past 50 years (correlations generally between .45 and .60). In this light, then, it is not astounding to note that large numbers 172 of freshmen who enter colleges and universities with the intention of graduating do not graduate-~somewhere around 50 per cent for the entire nation. IrreSpective of the multiple reasons given for this, the fact remains that this attrition rate represents a considerable loss Of human re- sources and a far from optimum efficiency in the functioning of colleges and universities. From a practical perspective, then, the development of instruments and other techniques for the evaluation of these ”non—intellectual” and primarily social psychological factors, which are currently proving themselves more worthy of consideration in predictive ef- forts, should be actively sought. Furthermore, since most of these factors are social psychological in nature, it would seem logical that much could also be done to instill the desirable aspects of each of these isolated factors in adolescents during the social- ization process; particularly during the junior high school and high school years. For example, the use of the Occupa~ tional Aspiration Scale as an attitude measure in this study suggests that much more must be learned about how occupation- al aSpiration develops as an attitude in the individual and what can be done to engineer that development. Again, the means-ends relationship which apparently exists between the educational and occupational structures in American society, which relationship is substantiated in the present study, 173 suggests that if we could learn how to change levels of aSpiration and levels of facilitation through various coun— seling and guidance techniques we could deliberately deter— mine to some extent a youngster's relative success in either the means or ends structure, that is, in either educational or occupational achievement. Obviously, behavioral scien- tists and educators have only begun to empirically develop these areas Of concern and thus the practical applications will not be immediate—-but the little evidence which does exist looks promising. In any event, the practical implica- tions of this study, and many other studies dealing with Similar variables, for the applied fields of counseling and guidance are both numerous and obvious. Thus, the applied researcher may also derive from the present study some prac- tical hypotheses for further testing. REFERENCES Adorno, T. W., et al. The Authoritarian Personality. New York: Harper, 1950. Allport, G. W. ”The Historical Background of Modern Social Psychology,” in Handbook of Social Psychology, Vol. 1 (edited by Gardner Lindzey). Cambridge, Mass.: Addison— Wesley, 1954, pp. 3—56 (Chapter 1). . ”Attitudes,” in Handbook of Social Psychology (edited by C. A. Murchison). Worcester, Mass.: Clark University Press, 1935, pp. 798-844. Beardslee, David C., and O'Dowd, Donald D. ”Students and the Occupational World,” in The American Collegg (edited by Nevitt Sanford). New York: Wiley and Sons, 1962, pp. 599—626 (Chapter 18). Bennett, G. K., Bennett, M. G., Wallace, W. L., and Wesman, A. G. College Qualification Test Manual. New York: The Psychological Corporation, 1961 (revised). Blalock, Hubert M., Jr. Social Statistics. New York: McGraw— Hill, 1960. Borgatta, E. F., and Meyer, H. J. (ed.), Sociological Theory. New York: Alfred Knopf, Inc., 1956. Brookover, Wilbur B., Paterson, Ann and Thomas, Shailer. Self Concgpt of Ability and School Achievement. (Final report of Cooperative Research Project No. 845-- mimeographed), East Lansing, Michigan: College of Educa- tion, Office Of Research and Publications, Michigan State University, 1962. Brown, Donald R. ”Personality, College Environment, and Academic Productivity,” in The American College (edited by Nevitt Sanford). New York: Wiley and Sons, 1962, pp. 536-562 (Chapter 16). Caplow, Theodore. The Sociology of Work. Minneapolis: University of Minnesota Press, 1954. 174 175 Chapmann, Dwight W., and Volkmann, John. "A Social Determi- nant of the Level of Aspiration,” in Readings in Social Psychology (edited by E. E. Maccoby, T. M. Newcomb, and E. L. Hartley). New York: Henry Holt, 1958, pp. 281-290. Conant, James B. Modern Science and Modern Man. New York: Doubleday Anchor Books, 1954. De Fleur, M. L., and Westie, F. R. ”Verbal Attitudes and Overt Acts: An Experiment on the Salience of Attitudes,” American Sociological Review, Vol. 23, December, 1958, pp. 667-673. . ”Attitude as a Scientific Concept,” Sgcial Forces, Vol. 42, October, 1963, pp. 17—31. Duncan, Otis Dudley. ”A Socioeconomic Index for All Occupa— tions," in Albert J. Reiss, Jr., Occupations and Social Status. New York: The Free Press offiGlencoe, Inc., 1961, pp. 109—161 (Chapter VI and VII). Durkheim, Emile. "The Occupational Group,” in Sociological Theory (edited by E. F. Borgatta and H. J. Meyer). New York: Alfred Knopf, 1956, pp. 287-289. Dvorak, Beatrice J. ”Advantages of the Multiple Cut—Off Method,” Personnel Psychology, Vol. 9, 1956, pp. 45—47. Edwards, Allen L. Statistical Methods for the Behavioral Sciences. New York: Rinehart and Co., 1954. . Techniques of Attitude Scale Construction. New York: Appleton-Century-Crofts, Inc., 1957. Enrollment Report, Fall 1960, East Lansing, Michigan: Michigan State University, Office Of the Registrar, 1960. Festinger, Leon. A Theory Of Cognitive Dissonance. Evanston. Illinois: Row Peterson and Company, 1957. Fishman, Joshua A. ”Some Social—Psychological Theory for Selecting and Guiding College Students,” in The American College (edited by Nevitt Sanford). New York: Wiley and Sons, 1962, pp. 666-689 (Chapter 20). Francis, Roy G. The Rhetoric of Science. Minneapolis: The University Of Minnesota Press, 1961. 176 Gaier, Eugene L., and Lee, Marilyn C. ”Pattern Analysis: The Configural Approach to Predictive Measurement,” Psychological Bulletin, 1953, Vol. 50, pp. 140—148. Grant, Donald L. (ed.). "Validity Information Exchange,” Personnel Psychology, 1955, Vol. 8, p. 105. Green, Bert F. ”Attitude Measurement,” in Handbook of Social Psychology. Vol. I (edited by Gardner Lindzey). Cambridge, Mass.: Addison-Wesley, 1954, pp. 335—369 (Chapter 9). Haller, A. 0. "Some Principles of Attitudes and Behavior,” unpublished working paper, 1957, Michigan State University, East Lansing, Michigan. . ”The Scope and Nature of Social Psychology,” unpublished paper presented at the 1964 Annual Meeting of the Ohio Valley Sociological Society, Columbus, Ohio, May 1, 1964. Haller, A. O., and Miller, I. W. The Occupational Aspiration Scale: Theory, Structure and Correlates. East Lansing, Michigan: M.S.U. Agricultural Experiment Station Technical Bulletin 288, 1963. Haller. A. 0. Occupational Aspiration Scale. East Lansing, Michigan: Michigan State University, 1957. Heider, Fritz. The Psychology of Interpersonal Relations. New York: John Wiley and Sons, Inc., 1958. Hovland, C. 1. ”Changes in Attitude Through Communication,“ Journal of Abnormal and Social Psychology, Vol. 46. 1951, pp. 424—437. Iffert, Earl. "The Student Retention and Withdrawal Study," College and University, Vol. 30, Summer 1955, pp. 406- 411. Janis, I. L., et al. ”Effects of Preparatory Communications on Reaction to a Subsequent News Event,” Public Opinion Quarterly, Vol. 15, 1951, pp. 488-518. Kluckhohn, Clyde. ”Culture and Behavior,” in Handbook of Social Psychology, Vol. II (edited by Gardner Lindzey). Cambridge, Mass.: Addison—Wesley, 1954, pp. 921—969. Koelsche, C. L. ”A Study of the Student Drop—Out Problem at Indiana University,” unpublished Ed. D. Thesis, Indiana University, Bloomington, Indiana, 1953. 177 Kutner, B., and Wilkins, C., and Yarrow, P. "Verbal Atti- tudes and Overt Behavior Involving Racial Prejudice," Journal of Abnormal and Social Psychology, Vol. 47, 1952, pp. 649—652. La Piere, R. T. ”Attitudes vs. Action,” Social Forces, Vol. 13, 1934, pp. 230—237. Lenski, Gerhard. The Religious Factor. New York: Doubleday, 1961. Lewin, K., Dembo, Tamara, Festinger, L., and Sears, Pauline S° ”Level of Aspiration,” in Personality and the Behavior Disorders (edited by J. Mc V. Hunt). New York: The Ronald Press Co., 1944, Vol. 1, pp. 333—378. Lindzey, Gardner (ed.). Handbook Of Social Psychology. Cambridge, Mass.: Addison—Wesley, 1954. Lipset, S. M., and Bendix, R. Social Mobility in Industrial Society. Berkeley, California: University of California Press, 1959. Loomis, Charles P., and Zona, K. Modern Social Theories. Princeton, New Jersey: D. Van Nostrand Company, Inc., 1961. Maccoby, E. E., Newcomb, T. M., and Hartley, E. L. (ed.). Readings in Social Psyphology. New York: Henry Holt, Inc., 1958. Merton, Robert K. Social Theory and Social Structure. Glencoe, Illinois: The Free Press, 1957. Munger, P. F. ”Factors Related to Persistence in College of Students Who Were Admitted to the University of Toledo from the Lower Third of Their Respective High School Classes,” unpublished Ph. D. thesis, University of Michigan, Ann Arbor, Michigan, 1954. National Opinion Research Center. ”Jobs and Occupation: A Popular Evaluation,” Opinion News, Vol. 9, 1947, pp. 3—13. Newcomb, Theodore M. ”Attitude Development As a Function of Reference Groups: The Bennington Study,” in Readings in Social ngchology (edited by E. E. Maccoby, T. M. Newcomb, and E. L. Hartley). New York: Henry Holt, 1958, pp. 265- 275. 178 Newcomb, Theodore M. Social Psychology. New York: Dryden Press, 1950. Parsons, Talcott. The Structure of Social Action. New York: The Free Press of Glencoe, 1949. Parsons, T., and Shils, E. A. Toward a General Theory of Action. Cambridge, Mass.: Harvard University Press, 1951. Parsons, T., and Bales, R. F. Family, Socialization and Interaction Process. Glencoe, Illinois: The Free Press, 1955. Peak, Helen. ”Attitude and Motivation,” in Nebraska Symposium on Motivation, 1955 (edited by M. R. Jones). Lincoln, Nebraska: The University of Nebraska Press, 1955, pp. 149—189. Reiss, Albert, Jr. Occupations and Social Status. New York: The Free Press Of Glencoe, 1961. Rokeach, Milton. The Open and Closed Mind. New York: Basic Books, Inc , 1960. Sanford, Nevitt (ed.). The American College. New York: Wiley and Sons, 1962. Sarnoff, 1.. et al. ”The Motivational Basis of Attitude Change,” Journal of Abnormal and Social Psychology, Vol. 49, 1954, pp. 115-124. Schmid, Calvin F. ”Scaling Techniques in Sociological Research,” in Pauline V. Young, Scientific Social Surveys and Research. New York: Prentice—Hall, Inc., 1949 (Chapter 14). Selltiz, Claire, et a1. Research Methods in Social Relations. New York: Henry Holt and Co., Inc., 1959. Sherif, M., and Cantril, H. “The Psychology of Attitudes,” Psychological Review, Vol. 52, 1945, pp. 295-319. Slocum, W. L. ”Social Factors Involved in Academic Mortality, College and University, Vol. 32, Fall 1956, pp. 53—64. Snyder, L. M. ”Why Do They Leave?” Journal of Higher Education, Vol. II, 1940, pp. 26— 32. H 179 Stouffer, S. A., et al. The American Soldier, Vol. 1. Princeton: Princeton University Press, 1949. Summerskill, J., and Darling, C. D. ”A Progress Report on the Student Stability Studies,” Student Medicine, Vol. 3, 1955, pp. 85-91. Watts, Henry J. ”Methodological Problems in the Measurement of Values,” unpublished Ph. D. dissertation, Michigan State University, East Lansing, Michigan, 1962. Williams, Robin M., Jr. American Society. New York: Alfred A. Knopf, 1960. Young, Pauline. Scientific Social Surveys and Research. New York: Prentice—Hall, Inc., 1949. Zetterberg, Hans L. On Theory and Verification in Sociology. Totowa, New Jersey: The Bedminster Press, 1963. APPENDIX A OCCUPATIONAL ASPIRATION SCALE AND CODE 181 Scoring Instructions Occupational ASpiration Scale All eight questions are scored the same. There are ten alternatives for each question, and only one alternative may be checked. The scores for each alternative are as follows: Alternative S2233 l 7 2 4 3 8 4 2 5 9 6 0 7 6 8 3 9 5 10 l The total score is the sum of the scores for each of the eight questions. . INSTRUCTIONS: This set of eight questions concerns your feelings about 182 I cccuprncnsx. :sspmge'rlon SC!.LE—] Copyright 1957 Form X By Archie C. :ialler YOUR N AME: Your Student Number: different kinds of jobs. Each one asks you to choose ONE job out of ten presented. This is not a test. There are no right or wrong answers ex- cept what is right or wrong for YOU. Indicate your choice by marking an] X! in the square next to the job. Read each question CAREFULLY. They are all DIFFERENT. Answer each one the best you can. DO NOT OMIT ANY. Be sure your name is on the top of this pge. QUESTION 1: Of the jobs listed in this question, which is the BEST ONE you are REALLY SURE YOU CAN GET when your SC HOG-LING IS OVER ? 1.1 l: Lawyer l. 2 E] Welfare worker for a city government 1. 3 D United States representative in Congress 1.4 [:1 Corporal in the Army 1. 5 D United States Supreme Court Justice 1. 6 [3 Night watchman 1. 7 E] Sociologist l. 8 [:3 Policeman 1.9 I I County agricultural agent 1.10 I: Filling station attendant J11.” \ 4 ... . .... 4 . . I . . e f I: i . . c. . v u. 1 . . .0 1. . . 1. \i _ .. 'Ill .14 .. . .r : .f . . - . A... ; a . . I 4 K I. «l . .l i . ,_ 1 183 -2; - QUESTION 2: Of the jobs listed in this question, Which ONE would you choose if you were FREE TO CHOOSE ANY of them you wished when your SC HOOLING IS OVER ? QUESTION 3: 2.1 I: 2.2 [:3 2.3 [:1 2.4 [:1 2.5 [:1 Member Of the boardbf directors of a large corporation Undertaker I Banker Machine Operator in a factory Physician (doctor) Clothes presser in a laundry Accountant for a large business Railroad conductor Railroad engineer Singer in a night club Of the jobs listed in this question which is the BEST ONE you are REALLY SURE YOU CAN GET when your SCHOOLING IS OVER ? 3.1 3.2 3. DDUDDDDUU 5" .5 f" a so .1 3.8 C» C 9° pa G Nuclear physicist Reporter for a daily newspaper County judge Barber State governor Soda fountain clerk Biologist Mail carrier Official of an international labor union Farm Hand .o.. I «u... o . u a .. \ . f 4 I . v I c. . . 0.» a .0 - .(l 4. 4 --...- .. A. w. w... .. . 5» t. I... . (Ir ...). . r .. w oil \ _ ... . ll. .... u .. .. Iv. .J z ..1. u a: l .01 . . .. .. . . 11’ l . ’i . . r . o .. ‘ d V L . l. _. .. J O O . ', h .L I .. a. ,. [J .70. . ) . a r 1 v... I s. c ‘1 ... r \J fl .3. . .-. .4. ‘1‘ I... 7': '1 , ... . l l I .‘_ .. . ..l\ . --.-... c. ..---- .. .....v- ....— 184 -3- QUESTION 4: Of the jobs listed in this question, which ONE would you choose if you were FREE TO CHOOSE ANY of them you wished when your SCHOOLING IS OVER ? QUESTION 5: 4. 4. 1 2 .9 C] .101: Psychologist Manager of a small store in a city Head of a department in state government Clerk in a store Cabinet member in the federal government Janitor Musician in a symphony orchestra Carpenter Radio announcer Coal miner Of the jobs listed in this question, which is the BEST ONE you are REALLY SURE YOU CAN HAVE by the time you are 30 YEARS OLD? 5. 5. l 2 JUDGE] .10 (:1 Civil engineer Bookkeeper Minister or Priest Streetcar motorman or city bus driver Diplomat in the United States Foreign Service Share crapper (One who owns no livestock or farm machinery, and does not manage the farm) Author of novels Plumber Newspaper columnist Taxi driver QUESTION 6: QUESTION 7: 185 -4- Of the jobs listed in this question, which ONE would you choose to have when you are 30 YEARS OLD, if you were FREE TO HAVE ANY of them you wished? 6.1 E] 6.2 T 6.. :1 6.10 [:1 Airline pilot Insurance agent Architect Milk route man ...“.dé. Mayor of a large city,__,-....~ ' Garbagecollector Captain .in-the.army ' ' ’ ' Garage mechanic Owner-operator of a printing shop Railroad section hand Of the jobs listed in this question, which is the BEST ONE you are REALLY SURE YOU CAN HAVE by the time you are 30 YEARS OLD? 7.1 7.2 ...) ...-I O DUNDEEDUDU Artist who paints pictures that are exhibited in galleries Traveling salesman for a wholesale concern Chemist Truck driver College professor Street sweeper Building contractor Local official of a labor union Electrician Restaurant waiter 186 -5- QUESTION 8: Of the jobs listed in this question, which ONE would you choose to have when you are 30 YE ARS C'LD, if you were FREE TO +1 AVE AN‘.’ of them you wished? 6.1 C] 6.2 [:1 8.3 [:3 6.4 D 8.5 [:1 6.6 E; 8.7 D 6.6 C] 8.9 [._. 8.10:] (.wner of a factory that employs about 100 people Playground director Dentist lumberjack Scientist Shoeshiner Public sc..-ool teacher Cwner—operator of a lunch stand Trained machinist )ock worker )(x NCT WRITE BELC“ .3 T'iIS LINE 1+3 (;—-3) 2 +4 (B) 5 +7 (C) 6 +8'(D) ES = (:. +3) 30 = (C + D) R = (1') +C) I = (B +0) TOTAL C (".3 SCCRE = (R +1) APPENDIX B DATA CARD FORMAT AND CODE 34 188 Card Layout Column 1 2 3 Student number 5 6 7 Age 8 Marital status 9 Hometown address 10 Stated occupational aspiration (s-e) 11 . . . 12 SOCioeconomic index of stated OA 13 14 Blank 15 Academic program 16 Father's occupational status (s—e) 17 . .. , 18 SOCioeconomic index of father 5 OS 19 . . . . 20 Perception of difficulty of means behaVior 21 Father's education 22 Mother’s occupational status (s—e) 23 . . . , 24 SOCioeconomic index of mother 5 OS 25 Mother's education 26 High school class rank 27 H-S. principal prediction of success 28 29 MSU GPA at end of last term of lst year 30 31 MSU GPA at end of last term, lst year (category) 32 33 MSU GPA at end of second year Column 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 189 MSU GPA at end of second year (category) OAS realistic choice level total score OAS idealistic choice level total score OAS level of aSpiration (total) Parental support (encouragement) Parental expectations (not encouragement) High school counselor support High school teacher support High school peer support Greatest source of support Personal reason for going to college Peer support (active) Perception of college as essential means Female aSpirations Persistence in college . . . arents Significance of peers reference groups Sounselor in college decision teacher High school GPA Dropout status (lst year) Dropout status (2nd year) Reason given for dropout College Qualification Test raw Score CQT percentile score CQT percentile category MSU Reading test raw score Column 71 72 73 74 190 MSU Reading test percentile score MSU English Placement test raw score MSU English Placement test percentile score Sex Combined reference group support Combined facilitation index score 191 CODING KEY ENTERING FRESHMEN AT MICHIGAN STATE UNIVERSITY, FALL 1960 Card 1 Column No. 1-6 Person identification (student number) Age at college entrance 0 — under 16 years of age 1 - 16 years 2 — 17 years 3 — 18 years 4 — 19 years 5 — 20 years 6 — 21 years 7 — over 21 years Y — no answer 8 Marital status at college entrance 0 - single 1 — married Y — no answer 9 Hometown address (geographical area) 0 — within 100 miles from MSU l — Michigan (other than 100 miles radius) 2 — Midwest (other than Michigan) 3 — South 4 — East 5 - West 6 — Foreign Y — no answer 10 Stated occupational aspiration (socioeconomic) 0 — professionals, technical, and kindred workers 1 — farmers and farm managers 2 — managers, officials and proprietors 3 - clerical and kindred workers 4 — sales workers 5 - craftsmen, formen, and kindred workers 6 — operatives and kindred workers Column No. Kxooow 192 private household workers service workers, except (7) above farm laborers and foremen no answer 11—12 Socioeconomic index of stated occupational aspiration 00 — 96 — YY — 13—14 Blank lowest possible index highest possible index no answer 15 Academic program Kcn~1©tn¢>wtO»ao I 16 Father' t<<3m‘dOWfl£>wEOP‘O I no preference College of Business and Public Service College of Science and Arts College of Education College of Agriculture College of Veterinary Medicine College of Communication Arts College of Engineering College of Home Economics no answer 5 occupational status (socioeconomic) professionals, technical, and kindred workers farmers and farm managers managers, officials and proprietors clerical and kindred workers sales workers craftsmen, foremen, and kindred workers Operatives and kindred workers private household workers service workers, except (8) above farm laborers and foremen no answer l7—18 Socioeconomic index of father's occupational status (socioeconomic) 00 — 96 - YY - lowest possible index highest possible index no answer 193 Column No. 19—20 Perception of difficulty of Means Behavior 0 - zero points (low difficulty) 31 - thirty-one points (high difficulty) X - negative figures 21 Father's education 0 — less than grade school (8 grades) 1 - grade school completion (8 grades) 2 — partial high school (9-11 grades) 3 — high school completion (12 grades) 4 - one year college 5 — two years college 6 - three years college 7 - college graduate 8 — beyond four years college (graduate work, law school, etc.) Y — no answer 22 Mother‘s occupational status (socioeconomic) - professionals, technical and kindred workers — farmers, and farm managers — managers, Officials, and proprietors - clerical and kindred workers — sales workers craftsmen, foremen, and kindred workers - Operatives and kindred workers - private household workers — service workers, except (7) above — farm laborers and foremen — no answer v<<>m~qown£>erFao I 23—24 Socioeconomic index of mother‘s occupational status (socioeconomic) 00 - lowest possible index 96 - highest possible index YY — no answer Column No. 194 25 Mother‘s education 26 27 28-30 31 DONOm-DWNF-‘O K1 less than grade school (8 grades) grade school completion (8 grades) partial high school (9—11 grades) high school completion (l2 grades) one year college two years college three years college college graduate beyond four years college (graduate work, law school, etc.) no answer High school class rank waNl-‘O bottom 50% upper 5% upper 5-15% upper 16-25% upper 26—50% no answer High school principal (knowledgeable) prediction of success <0JNHO below 1.00 1.01 to 1.99 2.00 to 2.50 2.51 to 2.99 3.00 to 3.50 3.51 to 4-00 no answer 195 Column No. 32—34 35 36-37 38—39 40—41 Michigan State University GPA at end of second year (Jr. standing) 000 - lowest GPA 400 — highest GPA YYY - no answer Michigan State University GPA category at end of second year (Jr. standing) 0 — below 1.00 l - 1.01 to 1.99 2 — 2.00 to 2.50 3 — 2.51 to 2.99 4 - 3.00 to 3.50 5 - 3.51 to 4.00 Y - no answer OAS realistic choice level total score 00 - zero points 01 — one point 36 — thirty-six points YY - no answer OAS idealistic choice level total scores 00 — zero points 01 - one point 36 — thirty—six points YY - no answer OAS level of occupational aspiration (realistic and idealistic) 00 - zero points 01 - one point 72 - seventy—two points YY - no answer Column No. 42 43 44 45 46 47 196 Parental support (encouragement) *<[\)l-‘O high degree of support moderate degree of support no support no answer Parental expectations (not encouragement) w10kao occupational world understanding scholar good time work avoidance other no answer Peer support (active) *oaNHO I 69—70 MSU reading test raw score 00 — lowest possible score 42 - highest possible score YY - no answer 71-72 MSU reading test percentile score 00 — no Score (lowest) 99 — ninety-nine (highest) YY - no answer 73—74 MSU English placement test raw score 00 - lowest possible score 35 - highest possible score YY - no answer 200 75-76 MSU English placement test percentile score 00 - no score (lowest) 99 — ninety—nine (highest) YY - no answer 77 Sex 0 — male 1 — female 78 Combined perceived reference group support - high — quite high - high moderate low moderate - quite low — low — no answer FHwfimom mum .OocmOMmficmfim mo “mow poafimunozp m How >pfiafinmnoud mo HO>OH mo. one we HGMOflwficmfim >Ham6flpmflpmwm mum m>onm no wwo. mo mOQOAOflmmooo coflpmaouuoo loam Apv Op Amv ”moanmfium> GOAHmpfiHfiomm AOV Op nmv coo. qu. Nam. Hmm. mum. 00H. mmm. mum. moo. Hoa. o nofi>mnon ANV ”Adv mo mom>p ”passase> messaep< AHV Anv oocmEuowuom owoaaoo on owoaaoo mo >uemmooo: mo :ofipmoouom Amv Spaaaps Hesse: Apv seem AOV Honomofi ADV hoaomcsoo Amv mucoumm Adv phOQddm dsoum mucouowou Om>floosom MOHV OHH. £ MOH>wo mo nonesz mpsomouaucoz mwsoaoua Unmocmpm .ufimmmomz asouw uuomumm ofieocoom >6666n< umpfiafiomm cofipmuflam< mo GOMquouwm museumMQm pmmm nofloom 666:6: 6666260 Hac06wmQ30oO mmqm ZOHH