VALUES, VALUE SYSTEMS, AND THE DEVELOPMENTAL STRUCTURE Thesis for the Degree of VIII. MICHIGAN STATE UNIVERSITY DAVID DANIEL McLELLAN 1970 . . , l . ....... ------ ...... . ..... "-'? ======= ['1 LIBRARY -" Michigan State University ' «ma-ht MSU I i LIBRARIES ' RETURNING MATERIALS: Place in book drofi_to remove this checkout from I Jul-I5IIIIL your record. FINES w111 7 be charged if book is I returned after the date stamped below. ”my I 4 ’7‘; ; Ina».— ABSTRACT VALUES, VALUE SYSTEMS, AND THE DEVELOPMENTAL STRUCTURE OF MORAL JUDGMENT By David Daniel McLellan Rokeach (1968) suggests that everyone who has undergone the pro- cess of socialization has acquired a set of beliefs about end-states of existence and modes of behavior which they consider personally and socially preferable to alternative end-states of existence or modes of behavior. The preferential end-states (terminal values) and preferential modes of behavior (instrumental values) are hierarchically organized into value systems. These values transcendentally guide behavior and judgments across specific objects and situations. Rokeach, however, does not directly consider the process of the development of values and value systems. Kohlberg (1964) has identified six distinguishable stages in the development of moral reasoning where this development is based on natural transformations of moral thought which reflect underlying cognitive pro- cesses. The stages, which reflect the formal, structural characteristics of the judgmental process, are: l. The punishment and obedience orientation 2. The instrumental relativist orientation 3. The interpersonal concordance orientation 4. The rigid rule orientation 2 . The social-contract legalistic orientation . The universal ethical principle orientation. David Daniel McLellan It was hypothesized that specific value (as conceptualized by Rokeach) differences, for individuals within a given culture, correspond to differences in their developmental levels of moral reasoning (as conceptualized by Kohlberg). To test whether the value concept as opera- tionalized in the Rokeach Value Survey actually is sensitive to develop- mental structural differences, 78 male gs from three grade levels (7th, 9th, & 11th) in a small suburban-rural public school were administered portions of the Kohlberg Moral Judgment Interview individually and were twice administered the value Survey in groups. The test-retest interval for the value Survey was three weeks. The major findings of this research were: ‘(a) the stability of terminal and instrumental value systems increased with age rather than with developmental level of moral reasoning; (b) Rokeach's "moral values" (a subset of instrumental values) were not found to differentiate better among §s at different moral levels than the non-moral instrumental values; (c) an overall measure of value system similarity did not reflect the moral development pattern; and (d) for specific individual values, there was a predictable pattern across stages of moral development. Two values, freedom and obedient, were found to discriminate strongly among moral stage-groups and to be predictable across moral stages within age levels. These two values were considered as values defining the structural variation in the development of moral reasoning and a single score utilizing both value ranks is discussed. value differences related to the age dimension were also discussed. It was concluded that the value Survey is sensitive to the structural differences in moral reasoning across the Kohlberg stages. However, it {II I I ‘ 1" I I'll. lu‘ l David Daniel McLellan was suggested that the current value Survey terms are best suited for adult §s and that different value terms be deve10ped for use with school age §_s . -/ ~ 0 Approved:/4ZLZ?§L- gtkfiéiéeeéxéf Date: 276/ £937 /7 70 VALUES, VALUE SYSTEMS, AND THE DEVELOPMENTAL STRUCTURE OF MORAL JUDGMENT By David Daniel McLellan A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1970 ’7f2“73f/ ’/¢ ”/./(3 To Zita ii ACKNOWLEDGMENTS I wish to acknowledge and to thank those individuals who aided in the completion of this Thesis: Committee members, Dr. Eugene Jacobson and Dr. James Phillips; committee chairman, Dr. Milton Rokeach, for his guidance and insightful comments; 'Miss Marge Hinze, for her secretarial assistance; Peg Stevenson and Ted Greenstein, for their intercession on my behalf with the computer; , the administration, faculty, parents, and students who assisted in my efforts to gather data; and my wife, for her critical aid in transcribing and scoring protocols and for her unfailing support. 111 TABLE OF CONTENTS DEDICATION ---------------------------------------------------- 11 ACKNOWLEDGMENTS ----------------------------------------------- 111 LIST OF TABLES ------------------------------------------------ vi LIST OF FIGURES ------------------------------------ _ ----------- ix INTRODUCTION -------------------------------------------------- l The Literature ------------------------------------------ 2 Value Definition ---------------------------------------- 4 value Development --------------------------------------- 7 Piaget -------------------------------------------------- 10 Kohlberg ------------------------------------------------ 14 values and Moral Reasoning ------------------------------ 19 Hypotheses ---------------------------------------------- 26 METHOD -------------------------------------------------------- 33 The Sample ---------------------------------------------- 34 Testing Procedures -------------------------------------- 37 Scoring the Kohlberg Moral Judgment Interview ----------- 39 Reliability of Scoring ---------------------------------- 41 RESULTS ------------------------------------------------------- 45 Moral Development, Cognitive Development, and Socio-economic Level ------------------------------- 49 value System Stability ---------------------------------- 50 value Structure and value Content ----------------------- 60 Age-related and Stage-related Value Differences --------- 83 Freedom and Obedient as Defining Moral Values ----------- 85 Rank-difference Score ----------------------------------- 90 Summary of Hypotheses ----------------------------------- 93 DISCUSSION ---------------------------------------------------- 95 FOOTNOTES ----------------------------------------------------- 100 REFERENCES ---------------------------------------------------- 105 iv APPENDICES ---------------------------------------------------- 109 A. B. C. D. E. F. G. H. Definition of Kohlberg's Moral Stages --------------- 109 Aspect List ----------------------------------------- 111 Selected Aspects by Stage of Development ------------ llS Parental Request Form Letter ------------------------ 118 Kohlberg Moral Judgment Interview ------------------- 119 Value Survey, Form D -------------------------------- 123 Global Rating Guide for Situation III of Kohlberg Moral Judgment Interview ----------------------- 126 Supplementary Tables -------------------------------- 131 8. 9. 10. ll. 12. 13. 14. LIST OF TABLES Predicted Value Peaks Across Stages of natal Development for Entire Sample (N578). Predicted Value Peaks Across Stages of Mbral Development for Each Grade Level. Number and Percent of Parental Permission Slips Returned by Male Students in Each Grade. £.tests for Mean Differences in Scholastic Achievement Scores for Sp vs, All Other Males in Each Grade. Mean C.A., PPVT Raw Score, and I.Q. for'gs in Each Grade. Percent Agreement Between Coders on Global Scores for 31 Ss. Percent Agreement Between Coders on Globaléclobal Scores for 31 Ss. Correlations Between.nbra1 Maturity Score Distributions of Two Coders. .g tests for Mean Differences in (weighted) Global Scores for Each Grade. Analysis of variance on Moral Ebturity Scores for Entire Sample by Grade and Socio-economic Level. Analysis of variance on Terminal value System Stability Coefficients (rho) for Entire Sample by Grade and Socio- economic Level. Analysis of variance on Instrumental Value System Stability Coefficients (rho) for Entire Sample by Grade and Socio- economic Level. Analysis of Variance on Terminal Value System Reliabilities by Stage (2, 3, G 4) and Grade. Analysis of Variance on Instrumental value System Reliabilities by Stage (2, 3, 5.4) and Grade. vi Page 30 31 35 36 38 42 43 46 51 52 53 57 58 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. Kendall Coefficient of Concordance, g, for Each Moral-stage Group at EAch Grade Level at time-1. Average Correlations (rho) of Time-l Terminal Value Systemm Between and Among So at Stages 1, 2, and 3 in Grade 7. Average Correlations (rho) of Time-l Terminal value Systems Between and Among gs at Stages 2, 3, and 4 in Grade 9. Average Correlations (rho) of Time-1 Terminal Value Systems Between and Among So at Stages 2, 3, 4, and 5 in Grade 11. Average Correlations (rho) of Time-l Instrumental Value Systemm Between and Among S9 at Stages 1, 2, and 3 in Grade 70 Average Correlations (rho) of Time-l Instrumental Value Systems Between and Among‘gs at Stages 2, 3, and 4 in Grade 9. Average Correlations (rho) of Time-l Instrumental value Systems between and Among gs at Stages 2, 3, 4, and 5 in Grade 11. Predicted value Peaks and Actual Value Peaks (time-l & time-2 combined) Across Stages of Moral Development for Entire Sample (N-78). Predicted value Peaks And Actual Value Peaks (time-l only) Across Stages of wral Development for Each Grade Level. Number of Correct value Peak Predictions in Table 22 and the Associated Binomial Probability Where the Probability of a correct prediction is 1/5. Number of Correct value Peak Predictions within Each Grade (Table 23) and the Associated Binomial Probabilities Mbdian Rank and Composite Rank Order of values Which Differentiate Best Among‘gs at Different Grade Levels. Median Rank and Composite Rank Order of values Which Differentiate Best Among So at Different Stages of Moral Development. Means and Standard Deviations of Rank-Difference Scores by Stage of Moral Development. vii 63 64 65 66 67 68 71 72 73 75 79 82 91 29. 1.3. 2.3. 3.H. 40n. 5.H. 6.H. Analysis of variance on Rank-Difference Scores by Stage of Moral Development and Grade Level. Mean Terminal Value System Stability Coefficients (rho) by Stage of Moral Development and Grade Level. Mean Instrumental value System Stability Coefficients (rho) by Stage of Mbral Development and Grade Level. Terminal Value Medians and Composite Rank Orders for gs at Each Grade Level (time-l & time-2 Value Survey data combined). - Instrumental Value Medians and Composite RAnk Orders for §_s at Each Grade Level (time-1 & time-2 Value Survey data combined). Terminal Value Medians and Composite Rank Orders for gs at Each Stage of Moral Development (time-l S time-2 Value Survey data Combined). Instrumental Value Medians and Composite Rank Orders for §_s at Each Stage of Moral Development (time-l and time-2 Value Data Combined). viii 131 132 133 134 135 136 1. 2a. LIST OF FIGURES Distributions of _S_s by Stage of Moral DevelOpment (Global- Global Score) Within Each Grade Level. Mean Percent of Total Moral Statements of each of 6 Moral Judgment Types at Three Ages (from Kohlberg, 1964, p. 403). Percent of Global Scores for §_s at Three Ages in This Study. Mean Value System Stability Coefficients (rho) by Grade Level. Mean Value System Stability Coefficients (rho) by Stage of Moral Development. Median Ranks for fleedom and Obedient by Stage of Moral Development (combined time-1 G: time-2 analysis). ix 48 48 55 56 87 INTRODUCTION values and morality have long received the attention of philosophers, theologians, and social scholars of various persuasions. As Kohlberg (1964) notes, morality was, for many generations, the central category for defining social relationships and development, and the social sciences were termed "the moral sciences." Morality and moral values have received only sporadic attention by the behavioral sciences in more recent years I but current trends would indicate that the pendulum is on an upward course. Rokeach (1968), for example, argues that values should replace attitudes as the central concern of social psychology. In the area of morality, Kohlberg has identified developmental stages in the structure of children's moral reasoning which have strong implications for the educational process. The importance of values, moral and otherwise, is twofold: for the individual and for the society. The sociologist, as Inkeles (1968) points out, in stating what it is that any society must have in order to survive is, in effect, specifying adult characteristics which must presumably be acquired by a significant portion of the population. Re- viewing Marion Levy's "functional requisites of any society," Inkeles argues that they are more a statement of the properties or qualities which individual members of a society must have if the society is to survive. Among these societal requisites are (a) a shared set of articulated goals, (b) regulation of choice of means, and (c) effective control of disruptive forms of behavior. Inkeles translates these societal requisites into elements of the personal system as (a1) values, (b1) values plus conscious functions of ego or "social self," and (c1) modes of moral functioning. That these elements are imperatives for any system of child sociali- zation makes them of great interest not only to the psychologist but to all members of society in that the relatively enduring patterns of adult values and moral functioning are of considerable significance as inputs into the social process. This concern for the development of values increases if the commonly-held notion that values of the young are more malleable than those of adults is correct. The establishment of valid and reliable methods of assessing development in these areas is useful for establishing norms, for identifying adaptive and maladaptive shifts early in development, and for assessing the effects of experimental treatments or socialization efforts. It is toward this general goal which this paper is aimed; specifically, a preliminary assessment of the validity and reliability of the Rokeach value Survey as a measure of value and value system development. The Literature Research and theory in the area of moral and non-moral values have taken a number of distinctive paths. Pittel and Mendelsohn (1966) have reviewed much of the literature on these efforts to assess values and have considered these efforts within the context of behavioral considera- tions. They see the history of these attempts to assess values as having three distinct eras since 1900, with each era characterized by a particu- lar type of instrument. The first era was seen by Pittel and Mendelsohn as being character- ized by paper-and-pencil tests which sought to differentiate normal children from deviant children. Among such efforts, the work of Hartshorne and May (1928-1930) stands out. Hartshorne and his collaborators de- veloped a number of instruments designed to tap a child's moral knowledge. These Tests of Moral Knowledge, for example, asked children to pick one of four solutions to a social situation. Others used by Bartshorne, et a1. attempted to measure vocabulary of moral words or attitudes towards various acts of misconduct. These tests were of little use in discrimi- nating among children who exhibited differential resistance to temptation. The tests did, however, correlate highly with intelligence. The second era was characterized by a growth of theoretical orienta- tions and the consideration of moral values within these orientations. Pittel and Mendelsohn saw two major trends within the second era: (1) deveIOpment of interview techniques for assessing the formal structure of moral reasoning within a developmental cognitive framework; and (2) integration of the consideration of values within omnibus investigations of personality based primarily on psychoanalytic and behavioristic models. The first trend was begun by Piaget (1932), about whom I shall have more to say later. The second trend is exemplified by Murray (1938) and by Eavighurst and Tabs (1949). Murray designed items to tap what were called Superego Integration, Superego Conflict, and Sentiments of the Superego in his efforts to study the normal personality. Havighurst and Taba's work on the adolescent character and personality included scoring essays on such topics as "Where Do I Get My Ideals?" and questionnaires designed to tap such traits as "Honesty" and ”Moral Courage." Here, too, there was found little correlation between "values" and "behavior."1 The third, and current, era in the study of values makes particular use of projective techniques which "all seem to deal with the superego in operational termm which place emphasis on the tendency of subjects to take a moralistic stance in the consideration of violations of conven- tional prohibitions, to project guilt feelings onto characters who vio- late these standards, and to indicate by their responses that they characteristically deny or suppress impulses which lead to socially un- acceptable behavior (Pittel and Mendelsohn, 1966, p. 32)." Another overview of value studies by MacCurdy (1950) lists four general value measuring techniques: 1. self report (paper-and-pencil tests), 2. intensity of emotional reaction, 3. variations in observed moral judgment, and 4. choice of "fealty" (i.e., behavior choice). Homant (1967) points out that the last three techniques all require ob- servation of a subject's behavior, thus they are all based on the notion of behavior reflecting implicit values. The self-report technique is the only one which lends itself to tapping explicit value orientations. V These techniques have been summarized in Homant (1967) and Bollen (1967). Robinson and Shaver (1969) present brief treatments of a number of the current instruments. Value Definitiog The consideration of values and value orientations and efforts to measure them depend, to a large extent, on the researcher's conceptuali- zation of what a value or value orientation is. Dukes (1955), in reviewing value studies in psychology, admits that philosophical considera- tions, such as defining means and ends, are inextricably intertwined in most conceptions of values. The various definitions and conceptualiza- tions used and instruments used are selectively summarized in Homant (1967) and Hollen (1967). These two theses, while summarizing value research, also offer extensive discourses on the Rokeach value Survey which has been used in this study. One of the most extensive treatments of the problem of the definition aspects of values has been given by Kluckhohn (1959). Reading the voluminous, and often vague and diffuse, literature on the subject of the various fields of learn- ing, one finds values considered as attitudes, motivations, objects, measurable quantities, substantive areas of be- havior, affect-laden customs or traditions, and relation- ships such as those between individuals, groups, objects, events. The only general agreement is that values some- how have to do with normative as opposed to existential propositions. CKluckhohn, p. 390). Kluckhohn does, however, offer his definition and elaborate on it. For him, a value is an implicit or explicit conception (for an individual or group) of the desirable which influences the selection from available modes, means, and ends of action. The union of the cognitive (conception) and the affective (desirable) dimensions is necessary in his definitions of value for "if the rational... is omitted, we are left with something not very different from...'senti- ment.’ When the affective aspect is omitted, we have something resembling 'ethics plus aesthetic and other taste canons.' The elements of 'wish' and 'appraisal' are inextricably united in 'value' (Kluckhohn, p. 400)." A different treatment of the definition of value by Barton (1962) is a consideration of the different object frameworks in which value has been treated. In his discussion, he points to four major classifica- tions of the use of "value." These classifications are combinations of explicit-implicit and preferential-normative dimensions. Here, the explicit-implicit dimension refers to the inference of value from behavior, on the one hand, and the verbslization of values, on the other. The preferential-normative separation is somewhat more ambiguous. Preferential values are individual goals while normative values are qualities used for judging others or oneself. It is not difficult to see that a normative value used by an individual to judge others may become a preferred goal for himself, thus blurring the dis- tinction between the two. Rokeach (1968) has also addressed himself to the area of values. Rokeach assumes that everyone who has undergone the process of sociali- zation has learned a set of beliefs2 about modes of behavior and about end-states of existence which they consider to be personally and socially preferable to alternative modes of behavior or end-states of existence. As noted above, Rokeach argues these values should become the major focus of psychological research, replacing attitudes, because values occupy a more central and dynamic role within the individual's cognitive-affective system. 3 A value, for Rokeach, is differentiated from an attitude in a number of respects: While an attitude represents several beliefs focused on a specific object or situation, a value is a single belief that transcendentally guides actions and judgments across specific objects and situations, and beyond immediate goals to more ultimate end-states of existence. Moreover, a value, unlike an attitude, is an imperative to action, not only a belief about the preferable but also a preference for the preferable (Lovejoy, 1950). Finally, a value, unlike an attitude, is a standard or yardstick to guide actions, attitudes, comparisons, evaluations, and justi- fications of self and others. (1968, p. 160). These preferential end-states of existence (terminal values)‘and preferential modes of behavior (instrumental values) are conceptualized by Rokeach to exist in a hierarchical organization within each individual's belief system. That is, each individual is posited to have two distinct value systems, terminal and instrumental, each with a hierarchy of values. These value systems are considered to be functionally and cognitively connected with each other and with specific attitudes. While Rokeach does theoretically consider value change and concom- itant attitude and behavioral change, he does not directly consider the development of values and value systems. The socialization process which has resulted in the acquisition of terminal and instrumental values has not been specified nor have the resulting emergent value patterns for varying socialization processes. Value Developggnt I have, to a limited degree, already touched upon psychological explorations dealing with the area of the development of values and value orientations during childhood and a more extensive examination will re- veal that the conceptual definitional problems noted in the preceding section are enlarged somewhat by the addition of a developmental dimension. The preceding section dealing with value definition left us with a generalized notion of a value being a cognitive conception of the desir- able means and ends of action and the conceptualization of Rokeach was specifically elaborated. It is explicit in this orientation that the value is "internalized;" that is, the value is an integral aspect of the functional cognitive structure of the adult individual. A shift in focus from the adult value and value system to a de- velopmental dimension has meant, for most psychological conceptualiza- tions of this century, a major concern with the increasing internaliza- tion of values. That is, if adult values have been considered internal- ized preferences of means and ends, then how and why this internalization occurs has been the primary concern of the develOpmentalist. In reviews of the theories of the development of value orientations, both Rohlberg (1963a, 1964) and Maccoby (1968) have noted that the most prevalent conceptualization of moral development has been that of increasing internalization of basic cultural rules of social action. They also note that three different aspects of this internalization which have been stressed in the theoretical literature and research. These aspects are (a) moral behavior, (b) moral affect, and (c) moral judgment. The previously-cited Eartshorne and May (1928-1930) studies ex- emplify research in the moral behavior area. Here, internalization was considered to be intrinsically motivated conformity or resistance to temptation. Hartshorne and May defined moral character as a set of culturally defined virtues (honesty, service, self-control) which they felt would be translated into measurable traits through the use of temptation situations. As noted above, their Tests of Moral Knowledge correlated little with moral behavior. They also found there was little consistency of moral behavior from one situation to another. The affective criterion of the existence of internalized standards is that of guilt. Both learning theories and psychoanalytically-oriented theories focus on guilt or anxiety as a basic moral motive and focus on inhibition as the basic expression of morality. Rokeach (1969) also makes room for this notion in his conceptuali- zation of values: To my mind, the general concept of value is consider- ably brosder than the concept of moral value. For one thing, moral values refer only to modes of behavior, instrumental values, and not to end-states of existence, terminal values. For another, moral values refer mainly to those modes of behavior which, when violated, arouse pangs of conscience or feelings of guilt or wrong doing; they have an interpersonal focus (p. 6). In other words, Rokeach sees moral values as a sub-set of instru- mental values which have specific functions and are specifically related to the affective dimension. Finally, the judgmental or cognitive aspect of internalization suggests an understanding of a standard and a positive valuing of it. For Rohlberg (1964), "the internalization of a standard implies a capacity to make judgments in terms of that standard and to justify maintaining the standard to oneself and to others (p. 384)."4' However, Kohlberg (1960) notes elsewhere the developmental position on this internalization: The basic assumption of most...is that moral develop- ment is a matter of internalizing external cultural stan- dards through reinforcement or identification. The develop- mentalist would say in opposition that there are basic attributes of adult morality which represents the outcome of developmental transformations of earlier motives and conceptions. Moral deveIOpment is not a simple stamping in of external prohibitions and rules, on a childish tabula rasa mentality. It is morality in this more positive and developmental sense which was seen as the key problem of socialization by the founders of social psychology; McDougall, Dewey, Mead, and Baldwin (p. 1). 10 Piaget This cognitive aspect of internalization received its modern pioneering work from Piaget (1932). It is important to note that the subject matter of this work is not separated in any important manner from Piaget's general cognitive theme at that time. Piaget's efforts in the moral judgment area can only be fully understood within the larger context within which he worked, i.e., the context of a more comprehensive effort to describe cognitive and logical development. Flavell (1963) notes that the important theoretical tie between Piaget's The Moral gudgggnt of the Child and his preceding works lies in understanding that the mechanism Piaget considers responsible for the development of rational morality is the same as for rationality in general. In addition to the developmental parallelism, Piaget saw an even deeper intrinsic connection between thought and morality: "Logic is the morality of thought just as morality is the logic of action (1932, p. 398)." Thus, to understand fully what Piaget is saying about moral develop- ment, we must first understand the larger, more inclusive concerns of his general theory. It is beyond the scope of this paper to provide comprehensive coverage of Piaget's general theory; therefore, I will not attempt to review Piaget's general theoretical foundations. Flavell (1963) details much of Piaget's work and summarizes his theory and Furth (1969) presents a reasonably concise summary of Piaget's theoretical foundations, emphasizing the biological and epistemological dimensions. Piaget's general approach to moral development is essentially a "stage" theory approach. That is, a child growing up in an environment 11 and interacting with the environment undergoes identifiable changes in his cognitive structure and functioning. These changes take place in a predetermined sequence in all children, yet not necessarily at the same time or at the same rate of change. Indeed, Piaget suggests that in certain environments development may stop at some particular stage. Use of the word "stage" is only a convenient method of condensing at certain points the characteristics of the changes which are taking place. Though we could not point to any stages properly so called, which followed one another in a necessary order, we were able to define processes whose final terms were quite distinct from one another. These processes might mingle and overlap more or less in the life of each child, but they marked nevertheless the broad divisions of moral development (Piaget, 1932, p. 175). The final result of Piaget's considerations of the child's moral reasoning is to identify two stages, or ideal types, in the early develop- ment of the child. The earlier of the two stages reflects the morality of constraint (the heteronomous stage) and the later stage reflects the morality of cooperation (the autonomous stage). As the developmental process is essentially concerned with the internalization of basic cultural rules of social action, the stages represent increasing internalization of rules by the child. Prior to the heteronomous stage, the child has not internalized rules at all; they are entirely external to himself. The heteronomous stage repre- sents a partial internalization of rules where the child feels an obliga- tion to conform even though he considers the sources of the rules to be external. The autonomous stage reflects the full internalization of the rules with the child feeling some control over them. 12 In the sample of children Piaget talked with, most children reached the autonomous stage by eleven years of age. That is, few children eleven years old or older made moral judgments using the immature aspects of the heteronomous stage. Children below seven years of age seldom characterized their moral judgments with aspects indicative of the higher autonomous stage. Kohlberg (1963a), in reviewing the research in the area of moral development, concluded that Piaget's generalized deveIOpmental view of moral judgment has received clear support in that there was cross- cultural evidence of age trends along several dimensions. However, many of the specifics of Piaget's theory, such as the two stages of develop- ment, have not been supported by research evidence. ’ At the time gfig ggral Jhdgggnt of the thld was published in 1932, Piaget had restricted himself primarily to verbal methods and it was not until some years later that he reached the perspective that the central mechanism of intelligence is found in the construction of operations which derive from the general coordinations of actions. Instead of "overcoming" egocentrism and realism, Piaget now writes of the develop- ment of the ability to carry out concrete operations. In his theory of cognitive development as currently viewed, the child moves from the sensori-motor stage, through the preoperational stage, to the stage of concrete operations and, finally, to the stage of formal operations. During the preoperational stage (roughly from 2 to 6 years of age) the child acquires the use of symbols but confuses means and ends. As the child moves into the stage of concrete operations, he begins to differ- entiate means and ends and begins to view means as instruments. This 13 stage of concrete operations, beginning about the seventh year and lasting until about the eleventh or twelth, roughly encompasses Piaget's autonomous stage of the development of moral judgment. Just as few children below seven years showed any use of the autonomous aspects of moral judgment, so do few children of that age show operational thinking. The strong emphasis of Piaget on the parallelism between general cognitive development and the development of moral judgment still holds, suggesting that the heteronomous and autonomous stages of moral develop- ment could be reformulated to fit into the contemporary theoretical structure. Piaget does allude to the continuing parallelism of cognitive and ‘moral development in The Growth of Logical Thinking from Childhood to Adolescence. In The floral ludgggnt of the ghild, adolescents were not considered and thus, development into the age range of formal operational thought was not then considered. In their book on adolescent thinking, Inhelder and Piaget (1958) were struck by the fact that feelings about ideals are _ practically nonexistent in the child. A study of the concept of nationality and the associated social attitudes ...has shown us that the child is sensitive to his family, to his place of residence, to his native language, to certain customs, etc., but that he preserves both an astonishing degree of ignorance and a striking insen- sitivity not only to his own designation or that of his associates as Swiss, French, etc., but toward his own country as a collective reality. This is to be expected, since, in the 7-ll-year-old child, logic is applied only to concrete or manipulable objects. ...The notions of humanity, social justice (in con- trast to interindividual justice which is deeply ex- perienced at the concrete level), freedom of conscience, civic or intellectual courage, and so forth, like the idea of nationality, are ideals which profoundly influence the adolescent's affective life; but with the child's mentality, except for certain individual glimpses, they can neither be understood nor felt. 14 In other words, the child does not experience as social feelings anything more than interindividual affects. Even moral sentiments are felt only as a function of uni- lateral respect (authority) or mutual respect. But, beginning at 13-15 years, feelings about ideals or ideas are added to the earlier ones, although, of course, they too subsist in the adolescent as well as the adult. Of course, an ideal always exists in a person and it does not st0p being an important interindividual element in the new class of feelings. The problem is to find out whether the idea is an object of affectivity because of the person or the person because of the idea. But whereas the child never gets out of this circle because his only ideals are people who are actually part of his surroundings, during adolescence the circle is broken because ideals become autonomous (pp. 348-349). Thus, moral development is not complete at the autonomous stage outlined in The Moral Judgggnt of the ghild. It is the development of formal operational thought during adolescence which is necessary for the formation of social ideals and principles. 5211223 The most extensive and significant reformulation of Piaget's theory of the development of moral judgment has been done by Kohlberg (1958 +). Essentially Kohlberg takes a quantitative approach to the general quali- tative material of Piaget. While Kohlberg (1963a) suggests that his work provides clear support for the general developmental view, it does not support Piaget specifically: As opposed to Piaget's view, the data suggest that the "natural" aspects of moral development are con- tinuous and a reaction to the whole social world rather than a product of a certain stage, a certain concept (reciprocity), or a certain type of social relations (peer relations). (pp. 322-323). Kohlberg's conceptualization of moral judgment suggests, as does Piaget's, that morality develops within a framework of general cognitive growth which imposes restrictions on the judgmental abilities of the child. 15 Within this generalized framework of developing cognitive abilities, Kohlberg has identified six distinct stages in the development of moral reasoning. These six stages (or ideal-types) which are thought to form an invariant, culturally-universal sequence are: I. Preconventional Level Stage 1: The punishment and obedience orientation. Stage 2: The instrumental relativist orientation. II. Conventional Level Stage 3: The interpersonal concordance or "good boy-nice girl" orientation. §§age 4: The rigid rule ("law and order") orientation. III. Postconventional, Autonomous, or Principled Level Stage 2: The social-contract, legalistic orientation. Stage 6: The universal ethical principle orientation. (See Appendix A for elaboration of stages) The research evidence to date indicates that the following four conditions, which Kohlberg considers necessary for validating the presence of true "stages", do hold for his typology: a. Regular age-related changes with lower stage judgments de- creasing and higher stage judgments increasing. b. Considerable generality across situations. c. Higher correlation of frequency of judgments in adjacent stages than with more distant stages. d. It should be easier to move a child up one stage in judgment than to produce any other change (summarized from.Maccoby, 1968). 16 Kohlberg (1958, 1963a, 1963b, 1964) has presented evidence to support his typology in both longitudinal and cross-sectional studies. Turiel (1965) demonstrated the invariance of the sequential progression and integration of lower stages into higher stages. Kramer (1968) in- vestigated the Kohlberg stages longitudinally and found support for earlier findings. Kohlberg (1968) also presents data supporting the cultural universality of his stages of moral development. Returning to the concept of internalization, Kohlberg concludes that moral internalization relates closely to the cognitive develop- ment of moral concepts. At the first two stages (stages summarized in Appendix A) standards of judgments and motivations are external to the child. The motivations are essentially external rewards and punishments. At the middle two stages, the standards the child uses are for the most part, external. However, he has internalized much of his motivation to conform*where he feels it necessary to maintain the expectations of the family, group, or nation. At the highest stages, the motivations and standards have become internal to the individual. It is at these stages that the individual becomes truly "moral" for Kohlberg. Each of these stages, for Kohlberg (1969), is a normative ethical theory. To define these stages, he has listed all the concerns on which a normative ethical theory must potentially take a stand in any given moral situation. These concerns, called Aspects or Categories (see Appendix B), are the basic units of moral judgments. At present, Hohlberg lists thirty Aspects which, in effect, each define a question, "What stand does your theory take on this area of concern?" The Aspects of Kohlberg's system are exhaustive but not necessarily mutually exclusive, 17 as one statement might bear on a number of Aspects at once. These areas of concern which the Aspects define must be culturally universal and universal across situations. That is, every culture or elaborated moral theory must focus to an extent on the Aspect. Also, it must be always logically possible to raise a question on any aspect in any moral situation. The structure of Kohlberg's typology is a 6 x 30 matrix: Stage x Aspect. Thus, for each Aspect or area of concern, there is a stage- typical orientation reflecting the developmental scheme. Each person, in making a moral judgment, logically can be called upon to take a stand on each one of the thirty Aspects at one of the six stages. Any given individual may be at different stages for different Aspects on a particu- lar moral situation. At the lower stages, some of the Aspects will not be differentiated from one another. The thirty Aspects are separated into three groupings: (a) the basic 22Q22.of normative moral judgment, (b) the basic principle; of normative moral judgment, and (c) the basic moral gplpgp,(see Appendix C). These groupings are subdivided into capital-letter modes and principles, each of which includes several Categories or Aspects. The Aspects which fall under (a) above may be seen as the different kinds of answers to questions as to gppp,is right or wrong, good or bad, Aspects under (b) are different kinds of moral reasons of answers to the question "Why is it right?" or "Why should someone do what you say is right?" The Aspects under (c) include such recurrent themes as Life, Property, Liberty or Autonomy, etc. Basically, the Aspects (c) are just applications of the other aspects to particular content areas. 18 Kohlberg's approach to moral development is essentially a descriptive one. He has identified stages in the development of moral judgment and provided some evidence to support the existence of the stages. However, he has yet to clearly confront the processes by which individuals pro- gress through the stages. He has cited evidence (Kohlberg, 1963) supporting, to an extent, a theory of developmental identification. It is through role-playing and identification at various levels that the individual develops and internalizes values. But, the variables and the mixture which facilitate moral development for Kohlberg are still not elaborated. In answering his own question about the interpretation and defini- tion of level of maturity of moral judgment, Kohlberg suggests: One general answer is that a more mature judgment is a more gpral judgment. This does not mean that a child who utters mature judgments is a more moral person, as judged by the standards of the community. It means that his judgments more closely correspond to genuine moral judgments as these have been defined by philosophers. While philosophers have been unable to agree upon any ultimate principle of the good which would define "correct" moral judgments, most philosophers agree upon the character- istics which make a judgment a genuine moral judgment... Unlike judgments of prudence or esthetics, moral judgments tend to be universal, inclusive, consistent, and to be grounded on objective, impersonal, or ideal grounds (1964, p. 405). Thus, the "goal" toward which moral development is continually striving may be structurally delineated and the progress of any child may be viewed relative to this goal. This goal, as represented by Kohlberg's postconventional stages, is represented best as a concept of justice. And, as Kohlberg (1968) notes, "The man who understands justice is more likely to practice it (p. 30)." This is not to suggest that there has been demonstrated a 19 one-to-one relationship between moral judgment and behavior. Kohlberg (1963a, 1964) considers at length the findings dealing with the corres- pondence between moral judgment and behavior. The findings suggest moderate correlations between stage of moral thinking and such behavioral measures as resistance to temptation.5 Basically, Kohlberg has found "that youths who understand justice act more justly, and the man who understands justice helps create a moral climate which goes far beyond his immediate and personal acts. The universal society is the beneficiary (1968, p. 30)." yelpep and ppral reasoning The research focus of this paper is the relationship between value survey responses, age, and developmental level of moral reasoning. Kohlberg has identified distinguishable stages in the development of moral reasoning where this development is based on natural transforma- tions of moral thought which reflect underlying cognitive processes. Kohlberg is concerned with the formal, structural characteristics of the judgmental process. That is, what are the formal criteria for action? This contrasts the formal structure of a judgment of right or wrong with the growth of moral knowledge or increased behavioral or verbal conformity to societal norms. This concept of the development of moral reasoning implies that, within a given cultural context, certain preferences about what are and are not desirable modes of behavior and end-states of existence will be shared by individuals at the same level of reasoning. Kohlberg has, in fact, built into his elaborated typology the generalized corres- pondence of the modes and principles of moral reasoning with specific 20 content areas such as Hupgn Life and Liberty or Aptonogy (see Appendix B for these basic values). Thus, for Kohlberg, values are an integral component of the developmental schema. For example, contrapt, promise, and nap-deception (a specific contept area) may be followed across the stages (structure). At stage 1, there are pp,reasons for maintaining trust; at stage 2, the reason for main- taining trust is self-interest; at stage 3, trust is maintained to avoid disappointing others; at stage 4, there is a categorical attitude about "keeping your word" with a sense of disappointment if you don't; at stage 5, there is a contractual conception of expectations somewhat more impersonal than 3 or 4 with an emphasis on the freedom not to enter into a contract when considering blame for violation of the contract; at stage 6, mutual trust is universalistic but more personal than at stage 5 as it is an act of faith going beyond simple contract--trust is also seen as a condition for the ideal society at this stage. In terms of a value hierarchy, contract, promise and non-deception as a value may be seen as becoming increasingly important within the individual's belief system.as he progresses through the six stages of moral development. For the two highest stages, maintenance of trust and honesty become preferred modes of behavior for ethical and moral reasons which transcent specific situational or societal variables. It will be noticed, that, in terms of a value hierarchy, cultural influences may reduce or increase the relative value distances among stages, thus making stage-related differences more or less measurable. In a society such as ours which places heavy emphasis on honesty, we would expect that even stage-1 individuals would rank hopesty highly but that the increasing internalization 21 of the value which occurs with the development of moral reasoning would cause honesty to be valued still higher by the later stages. ,%32 Thus, within a given cultural context, we would hypothesize that the increasing internalization of specific moral content which accompanies the development of moral reasoning in the Kohlberg schema and increasing ranking of that moral content (i.e., value) within the individual's value hierarchy as conceptualized by Rokeach are functionally equivalent. For Rokeach, the judgment of right or wrong is based on the value hierarchy within an individual's belief system. That is, a moral judgment (i.e., a judgment of action based on a set of factual and evaluative beliefs about a particular set of situations and individuals) will reflect the terminal and instrumental values of an individual. The particular set of values (value system) a person holds become "a standard or criterion for guiding action,...for morally judging self and others (Rokeach, 1968, p. 160)." These values, these preferential end-states and modes of behavior, are not situationally-bound but rather transcendentally guide actions and judgments across specific objects and situations. It is possible, then, for two individuals to value, as an end-state of existence, eguality more highly than all other alternative end-states. However, the reasons why they value equality so highly may be different. Person Y may value eguality because of his universal ethical principle orientation (Kohlberg's stage 6) which is based on the principles of the reciprocity and equality of human rights. The other individual, person Z, may value eguality highly because of a rigid rule-orientation (stage 4) which recognizes that the official doctrine of the society is to value 22 egualipy highly and maintenance of the social order demands it. It is apparent, however, that Y's value of eguality has validity and applica- tion apart from any societal norms while z's high value on eguality depends on the perceived social and legal norms. Likewise, let us compare three persons who all value pppgpp,most highly as a mode of behavior. Person A, when asked why he values honest, might say that it is because his mother and father will spank him and send him to bed without supper if he is not honest (stage 1). Person B might say he values honest because his family and friends value honest highly and he wouldn't want to disappoint them (stage 3). Person C might say he values pppgpp,highly because of the need for maintenance of trust which is the foundation of extra-legal moral relations within the society (stage 6). We can see that a particular value may be held for and mean differ- ent things to different people. However, within a given cultural milieu, we anticipate certain value differences among individuals who utilize differential reasoning for maintaining their values. For example, with the two individuals, Y and 2, who both valued eguality highly, we would anticipate that Y actually values eguality more highly than 2 because Y's reasoning represents self-chosen ethical principles appealing to logical comprehensiveness, universality, and consistency while 2 considers eguality valuable because of his identification with the social authority. To say that one "actually values" eguality more highly is to say that the value is internalized for Y, suggesting little cross-situation variability and a more consistent high ranking of the value when compared with conflicting values. Thus, if the social 23 authority were to de-emphasize eguality in order, say, to placate segments of an electorate, Z would be considered to have a higher prob- ability of changing his value of eguality downward than would Y whose values are not based on identification with the social system. Likewise, A, B, and C's high value of hppgpp,carry different impli- cations for comparison.with conflicting values and for the correlation with honesty-related behavior. Just as with earlier research on moral behavior, simple knowledge of the value of pppgpp_by adults or peers for those at moral stages 1 through 4, does not necessarily mean they will ppp_honest1y suggesting that other values such as succeps or pppggl recognition actually take precedence in real-life situations. 'However, those who have achieved moral stage 5 or 6 and who value honest highly would be expected to be paying more than lip service to the value. These hypothesized relationships between an individual's develop- mental stage of moral reasoning and his values are not thought just to evidence themselves in value rank-differences but in the stability of his value system as well. Where values are external to an individual, as they are at stage 1, his value system is more susceptible to external influences than is the value system of the individual whose values are rooted in self-chosen ethical principles of Kohlberg's stage 6. Thus, we would hypothesize that the higher an individual's development stage of moral reasoning, the more stabile will be his value system over time. A seemingly obvious point of contact for the value concept of Rokeach with the developmental dimension of moral judgment is Rokeach's concept of moral value, which was discussed earlier in the context of moral affect. However, when the concepts of moral value and moral 24 judgment are considered beyond their nominal similarity, this point of contact becomes less obvious. As has been noted, Rokeach (1969) defines moral values as a subset6 of instrumental values which have an interpersonal focus and violation of which results in affective arousal (e.g., guilt). The judgmental aspect of morality considers the ability to make judgments in terms of a standard and to justify maintaining that standard. The concept of "moral" in moral value is not isomorphic to "moral" in moral judgment. For Rokeach, a value is intrinsically moral in that its referent is a mode of behavior which has an interpersonal focus. This use of moral value suggests a rule--deontologica1 theory which holds that rules such as "honest" or "obedient" are valid standards of right and wrong apart from any consideration of the specific situation or the consequences of acting in accord with the moral value. This is contrasted with a teleological theory which is ultimately concerned with the comparative balance of good and evil which results from acting (Frankena, 1963, chapter 3 deals with these topics in some detail). The concept of moral value may also be contrasted, in a somewhat separate way, with the position of Potion (1968) on "what makes a moral situation moral?"7 The use of the standard in Hohlberg's concept of moral judgment encompasses not only the rules for behavior (moral values) but also the terminal values which are used for such purposes as justifying a particular mode of behavior in a given situation. Especially at the post conventional level of Kohlberg's typology, all three levels of reasoning in dealing with a moral situation suggested by Potion (1968) are found. These levels are: the descriptive or factual level, the l ‘1 '. 25 rule level (where rules are equivalent to Rokeach's instrumental values), and the value level (equivalent to Rokeach's terminal values). Potion sees that the basic factual elements of a moral situation lead to a consideration of what rules for behavior are relevant in that situation which lead, in turn, to a consideration of principles or judgments of individual or societal goals which are relevant. In a somewhat different context, Frankena also acknowledges a similar sentiment: I propose therefore that we regard the morality of‘ principles and the morality of traits of character, or doing and being, not as rival kinds of morality between.which we must choose, but as two complimentary aspects of the same morality (p. 53). The use of both terminal and instrumental values as defined by Rokeach would thus be anticipated in making moral judgments. Thereby, developmental differences in the structure of moral reasoning could result in both terminal and instrumental value differences, rather than in just moral value differences as might be suggested by equating the concept Vmoral" in moral values and moral reasoning. My value data will give us the opportunity to determine the rela- tionship of the moral and nonmoral instrumental values to the moral judgmental dimension in order to test whether the moral values differ- entiate among individuals at different stages of moral reasoning better than the nonmoral values. The overall relationship between the terminal values and the developmental dimension of moral reasoning will also become more clear. Finally, based on the Kohlberg typology, there are specific values which are conceptually linked to the moral development dimension and these specific relationships will be detailed in the following section. 26 Hypotheses We have already (p. 21) stated the general hypothesis which we wish to test: specific value (as conceptualized by Rokeach) differences for a given cultural context, correspond to differences in the develop- mental level of an individual's moral reasoning ability (as conceptualized by Kohlberg). In the preceding section, the stability of value systems was hypothe- sized to be related to stage of moral development. Specifically, the hypothesis to test is: H1: The stability of terminal and instrumental value systems will increase as pppgg.of development of moral reasoning increases, with age held constant. Within a given culture, specific value content, as measured by the Value Survey, is thought to correspond to the developmental ptrpcture of moral reasoning. This general hypothesis may be operationally tested both by a global measure of value similarity and by specific value .differences predicted from the developmental stages of moral reasoning. The global value similarity notion yields this hypothesis: H4.1: Holding age constant, §s at a given pppgg,of moral development will show more value system similarity when compared to one another than when compared to §s at different stages; this pattern of value system similarity among the developmental stages will reflect the develop- mental pattern. Ill‘l‘ 27 If our hypothesis that the ethical theories which the Kohlberg stages represent have something to say about values is correct, we have only to select those values from.the value Survey with which the moral judgmental dimension deals directly, predict value differences consistent with the developmental structural differences, and, finally, assess the accuracy of our predictions. The nature of the Value Survey, however, limits the type of value differences which can reasonably be predicted. The relative ranking procedure makes any absolute value rank prediction or absolute rank differences prediction futile. The best that we can say is that when a stage-typical ethical theory emphasizes a value more than the other stage-typical theories, that value should, on the average, be ranked higher by gs at that stage of reasoning than by g; at the other stages. Thus, our next hypothesis will be in the form.of a list of values which will be ranked highest on the average ("peak"8) by gs at one or two specified stages of moral development, if the values are related to the moral dimension. These predictions are not thought to be related to age differences, so all 78 g; may be considered as one group. However, since age and stage of moral develop- ment are correlated, each value prediction will also be expected to hold within each grade level separately. Where there are too few g; at the moral level where a value is predicted to peak, that prediction is altered to conform to the stages available within the grade level. This somewhat lengthy hypothesis is: H4.2: The following values will be ranked highest on the average ("peak") by S; at the indicated stage(s) of moral reasoning. ll. .1 I {full ll 28 I. Terminal values 1. A pomfortable life will peak at stage 2. The stage 2 instrumental hedonists of the Kohlberg typology are preoccupied with a concern for their own comfort and well-being rather than for others. It would be at this stage where a comfortable life would most be utilized in making moral judgments. 2. Eguality will peak at stage 6. It is at this stage where eguality becomes a defining element of the moral . reasoning. Lacking any stage-6 g; in our sample, 1 ,m however, eguality will peak at stage 5, where eguality LS”5 and brotherhood are also among the defining elements of moral thought. In each grade separately, eguality will peak at the highest stage present. 3. Eagily pepppity will peak at stage 1 and at stage 3. The stage-l moralist defers unquestionly to the pre- ,, vailing power which, for our §s, is primarily found in the family. This deference to the parental authority reflects a high regard for the family structure. In a much different sense, the stage-3 moralist also has a high regard for the family structure, but this regard is reflecting an identification with the family rather than the simple deference to it as with the stage-l moralists. I ~ 4. Ereedom will peak at stage 2 and stage 5. The hedonistic relativists at stage 2 feel that anyone may do what they wish with their own lives or property, reflecting a con- cern for their own freedom. The stage-5 moralists share this concern for individual freedom with the stage-2 moralists in that the concept of social-contract empha- sizes free agreement outside the legal realm. 5. National security will peak at stage 4. The stage-4 moralists consider the honor and welfare of the state as an ultimate concern; something which should be defended at all cost. 6. Pleasure will peak at stage 2. The hedonistic overtones of the stage-2 morality reflect a prime concern with the self and the tendency is to make judgments on the basis of what pleasure or good an act brings to the self. 7. Salvation will peak at stage 1 and at stage 4. The moralities at these two stages represent the most socially conserving moralities and the most absolutist positions in terms of religious imperatives. As Rokeach (1969) has shown, salvation is highly correlated with religiousness and social conservatism. 29 II. Instrumental values 8. ppoadminded will peak at stage 5 and stage 6. It is at these stages of moral reasoning that broad- mindedness is strongly implied in that openminded- ness and flexibility are prerequisites for making moral judgments. 9. {pggiyppggwill peak at stage 6. It is in the stage-6 morality where the moral principles of justice are principles of obligation and not principles requiring blame or justifying blame of others on self. In our sample, stage-5‘§s will share this outlook most. Where no stage-5 §s are present in a particular grade, the morality of stage-3, with its concern with being nice and maintaining pleasant relationships, would lead to higher ranking of this value. 10. Helpful will peak at stage 3 and at stage 5. The good, for the stage-3 moralist, is often defined as helping others. For stage-5 moralists, the welfare of the community is an ultimate criteria of the consequences of action so that behavior which leads to this is highly valued. ll. Loving will peak at stage 3. The stage-3 moralist, in defining the good, often refers to this value. 12. Obedient will peak at stage 1 and at stage 4. At stage I, respect is defined as obedience to the superior power. At stage 4, right behavior consists, in part, of obedience to and respect for the social authority. 13. Responsible will peak at stage 5 and stage 6. The moralists at these stages consider themoelves responsible for all con- sequences of their own action or inaction to which moral principles apply (this is true more so for the stage-6 moralists). 14. Self-controlled will peak at stage 4 and stage 5. The stage-4 moralist is concerned with rigid maintenance of rules in order to avoid guilt. The stage-5 moralist, with a greater awareness of his own responsibility, demands a greater measure of self-discipline. These value peak predictions for the entire sample are summarized in Table l. The predictions within each grade level are summarized in Table 2. To specifically deal with the issue of moral values which was dealt with in the preceding section, we will make two additional hypotheses 30 Table 1. Predicted Value Peaks* Across Stages of Moral Development for Entire Sample (fi-78). Stage Terminal _1__L_3__g__5_ A Comfortable Life * Equality * Family Security * * Freedom * * National Security * Pleasure * Salvation * * Instrumental Broadminded * Forgiving * Helpful * * Loving * Obedient * * Responsible * Self-controlled * * 1A peak is defined as the highest median ra‘nk order among the groups for each value. 31 Table 2. Predicted value Peaks* Across Stages 0f Moral Development for Each Grade Level. Grade Level 7 9 11 Stage Stage Stage Terminal 1 2 PL 2 3 4 2 3 4 5 A Comfortable Lif W WT”— T 'fi Equalit * * * Family Securit * e e * Freed * * * a National Securit * e * Pleasur * * * Salvatio * b * * Instrumental Broadminde * * * Porgivin * * * Helpfu * e e * Lovin * * * Obedien * a * Responsible * * * Self-controlle * a s * *A peak is defined as the highest median rank order among the stage-level groups within each grade level for each value. 32 based on the Rokeach conception of moral values. These hypotheses are designed to test whether the moral values as defined by Rokeach are operationally related to the moral judgmental dimension to a greater extent than are the terminal values or the non-moral instrumental values. Under H1, we hypothesized both terminal and instrumental value system reliabilities would correlate with stage of development. If the moral values are more intrinsically related to the moral judgmental dimension, then the instrumental value system stabilities should show a stronger correlation with that dimension than will the terminal stabilities. Formally stated, this hypothesis is: H2: Instrumental value system stability will correlate more highly with pppgg of moral development than will terminal value system stability. The related hypothesis is: H3: The "moral values" will differentiate among gs at different moral ppgggp more than will the non- moral instrumental values. METHOD Research designs in developmental areas, according to Kessen (1960), are of two generalized types: longitudinal and cross-sectional. The longitudinal designs make use of repeated measures of the same subjects at different points along the time scale. The cross-sectional designs utilize single measures of individuals who are, at that moment, at different points along the time scale. There are benefits to be derived from each approach: the repeated measures design gives more sensitive estimates of small, reliable changes which occur while the cross- sectional approach must rely primarily on group means and the like for analysis. Also the cross-sectional design fails to handle cultural variance at the different ages as does the longitudinal design. However, the cross-sectional design has the undeniable virtue of being quicker and cheaper. For this study, a cross-sectional design was chosen. Economic con- cerns were primarily responsible but it was also felt that the results from a cross-sectional analysis would, in the event hypotheses were confirmed, make results much more compelling. This is because larger value differences between groups would have to occur in order for significant results to be found. As Rokeach and Parker (1970) note, "we would expect value differ- ences to be associated with differences in subcultural membership, sex, religion, age, race, ethnic identification, life style, socio-economic 33 34 status, child-rearing practices, intelligence, authoritarianism, etc." Since our interest here is to determine what, if any, value differences are related to structural differences in moral judgment, we wish to eliminate all extraneous variables (other than age) when selecting our sample so that developmental difference is, to what extent possible, the single independent variable. To achieve this end of selecting a homogeneous sample, we drew our subjects from the public school system of a small suburban-rural community a few miles from Lansing,'Hichigan. The school and community are virtually one-hundred per cent white, predominately middle to lower- middle class. The Sample gs (NI78) were white males from the seventh, ninth, and eleventh grades. At the request of the school administration, positive parental approval for each potential subject had to be acquired (Appendix D con- tains a sample parental request form). This procedure reduced the 9 Table 3 indicates the varying available pool from which to draw Sp. return rates. The procedure also resulted in a somewhat biased sample, as Table 4 indicates. In all three grades, the mean scholastic achieve- ment scores for g; and non-subjects were compared. In all cases, the mean scores were higher for the Sp than for other male students in the same grade who were not S; in this study. In grades seven and nine, the differences are quite marked. Thus, the required procedure resulted in a more scholastically advanced sample than.we would have anticipated had we drawn our gs at random from each grade rather than from those students for whom we had parental permission. 35 Table 3. number and Percent of Parental Permission Slips Returned by Male Students in Each Grade. Grade 7 9 11 Total n n n N 1 Number of Male Students 79 68 55 202 Affirmative Parental Responses 39 (49) 38 (56) 32 (58), 109 (54) Negative Parental Responses 11 (14) 7 (10) 5 ( 9) 23 (ll) Total Parental Responses 50 (63) 45 (66) 37 (67) 132 (65) 36 Table 4. 3; tests for Mean Differences in Scholastic Achievement Scoresa for Se ypp4All Other Males in Each Grade. X' (72 n df t p Grade 7 Othegplales 2 g 382:: Z: 68 2°°33 ('05 Grade 9 Otheg’uales :4 2 223:: g; 56 1‘977 4<.10 e... u 0.1.33.1... i 2 32:? $3 55 0-388 mo aScores not available for all students. from pupil files were used for this analysis: - Composite percentile score on Iowa Test of Basic Grade 7 Grade 9 Skills The following test results - Composite G.E. on Iowa Test of Basic Skills Grade 11 - Verbal Reasoning raw score + Numerical Ability raw score on Differential Appitude Test. 37 Testing Procedures Each §.was first met by §,in an individual 50-minute session in the school. At this session, §.and §,chatted for a few minutes to get at ease and then §.explained the general purpose and nature of the testing to §, §.assured §Dthat the procedure would be harmless and that §fs responses would be confidential. §_was assured that no one other than ngould see his responses in any but anonymous form. '§.also reiterated that he was not connected with the school, that the school had only allowed him to use its facilities. §_was then asked if he wanted to ‘ participate. He was assured that he was free to return to class and not participate for any reason. No student chose not to participate. The testing began with asking age, parental occupation and education, and family size. §_was then given the Peabody Picture Vocabulary Test (PPVT) (Dunn, 1965) to quickly assess MA and IQ. The PPVT was given prior to the Rohlberg Moral Judgment Interview because the game-like quality of the PPVT is useful for establishing rapport with §.and getting him to respond freely. The mean C.A., PPVT raw score, and IQ for each grade are presented in Table 5. .The mean 10's for gs at each grade level are reasonably equivalent in that an ANOVA yielded a non-significant g, The bias effect of the sampling procedure noted in Table 4 has resulted in higher mean IQs for the seventh and ninth grade, but not significantly higher. The increasing mean PPVT scores, which are equivalent to.MA, show the predictable in- crease with grade. §.was then read four hypothetical moral dilemmas from the Kohlberg Moral Judgment Interview (DUI) and was asked to answer questions about 38 Table 5. Eben C.A., PPVT Raw Score, and I.Q. for Se in Each Grade Grade 7 9 ll C.A. 12.9 15.0 16.8 PPVT raw score 102.6 112.1 116.9 I.Q. 115.3 116.7 112.2 39 the stories (see Appendix E for the four stories and questions). st responses were tape-recorded and later typed up for scoring purposes. After all Sp had completed this initial individual testing, they were twice brought together in groups to complete the Rokeach value Survey--Form D (see Appendix F). The test-retest interval for all S9 was three weeks.10 Scorin the Rohlber Moral ud nt Interview I The global rating method (Rohlberg, 1958 & 1969) was used in scoring the st responses on the RHJI. (See Appendix G for a sample Global- Rating Guide). Each situation is scored separately for all Sp. With the Global-Rating Guide, each st responses are given a major code indi- cating the predominant stage of moral reasoning used by the subject in making judgments about the situation. If one stage is not clearly pre- dominant, the scorer may also assign a minor code indicating a secondary stage of reasoning used by S, This major code alone or major and minor code is the global score for that S.for that situation. When all proto- cols have been scored, each S_has four global scores indicating the stage of moral reasoning used in making judgments about each situation. The scorer then, on the basis of the four global scores, assigns each subject to one of the six stages of moral reasoning which best typifies that st level of moral reasoning. This assignation to one of the ideal- type stages is the st global-global (GG) score and is said to represent that st developmental stage of moral reasoning. Although the Global Rating method is somewhat less precise than the detailed scoring method first elaborated by Kohlberg (1958) or his newest, not-yet-completed, Aspect Scoring method (1969), it has the benefit of 40 being much faster. However, despite the method used to score each situation, there has, as of yet, never been developed a clearly de- fined method for "averaging" the global scores in order to arrive at the global-global score or "pure" stage-type. Haan, Smith, and Block (1968) used the following method for assigning college students to a final "pure" type: Each global score was given a weight of 3 (major code only-3; if major and minor, major code weight-2, minor code weight-l) and the weights at each stage were summed across all situa- tions. If, for two judges, the summed weight of the highest weighted stage was at least twice the summed weight of any other stage, the subject was assigned to that stage. This is a fairly rigorous method, as Haan, et. a1. point out even when the detailed scoring method is used. In this study, assignment of Sp to a final "pure" type was also based on the global scores, but every S.was required to be assigned so any method which eliminated S9 for failure to meet a criterion was not useful. It was thus decided to sum the weighted global scores and the largest sum would determine the global-global score for that S. For example, subject 30 had the following global scores on the situations and was determined to be a "stage 2" moralist on the basis of the weighting: Situation: III IV I VII Global Score: 2 (4) 1(3) 2 2(3) 2 l 2 2 u 2 2 3 2 (D g! l l 2 a; 4 l l 5 O 6 0 41 'Where ties occurred, the scorer re-evaluated st protocol to make a judgment as to which stage best represented the reasoning of S, Reliability of Scoping The outlined procedure for assigning Sp to "pure" types requires considerable training and practice to achieve reliable results. Two coders were trained in scoring the KMJI and the protocols £0 31 Se were randomly selected for scoring by both (the remainder were scored only by one coder, the experimenter) coders. For reliability purposes, global scores of the two coders were considered to "agree" if any of the following three conditions were met: a) both major and minor codes were identical, b) major and minor code were just reversed (e.g., 4(3) and 3(4)). or c) major codes were identical regardless of any minor codes (e.g., 4(2) and 4(3)). For global-global scores, agreement is only when "pure" type scores are identical as there are no minor codes. The percent agreement between the two coders on global scores for each situation and grade are shown in Table 6 and the agreement on global-global designations, in Table 7. An additional measure of reliability is the correlation between the moral maturity (HMO scores given by each coder.11 Table 8 shows these correlations. These tables indicate fairly good agreement between the two coders. One additional reliability measure is the percent agreement of the two coders with a set of test protocols used by Rohlberg in training. Of 10 situations, Coder A's (experimenter) global scores agreed 1001, Coder B's agreed 901. 42 Table 6. Percent Agreement Between Coders on Global Scores for 31 Ss. Grade 7 (n-10) Grade 9 (n-ll) Grade 11 (n-IO) Total KHJI Situation III IV I VII 501 802 801 601 1001 731 911 371 801 1001 802 802 771 842 841 581 Total 681 802 851 43 Table 7. Percent Agreement8 Between Coders on Global-Global Scores for 31 Ss. Grade 7 9 11 Percent Agreement 901 82% 802 aWhere agreement was not reached, the global-global scores given by the two coders were always in adjacent stages. 44 Table 8. Correlations Between Moral Maturity Score Distributions of Two Coders Grade 7 9 ll Combined (n-10) (n-ll) (n-lO) (n-31) r- .932 1 .867 T .903 .903 RESULTS The distribution of moral types within and between grades reflects the development dimension in that the lower stages tend to become less frequent and the higher stages more frequent as age increases (see Figure 1). For example, 231 of the seventh graders are at stage 1 while there are no eleventh graders at that stage. The mean global-global scores show an increasing trend as age increases. When weighted global scores for each situation are used as the unit of analysis, the pattern is essentially the same. An analysis of these means (Table 9) indicates that the dis- tributions of global scores are significantly different for the three age groups, as would be expected. The distribution of moral types is somewhat more positively skewed than we had anticipated, especially in the eleventh grade sample. Kohlberg's (1964) chart of the age trends across the six stages is in Figure 2a. Note that use of stage-1 and stage-2 reasoning decreases with age, use of stage-3 and stage-4 reasoning increases until age 13 and then stabilizes, and the use of the highest two stages increases from age 13 to 16. Figure 2b shows that percentages of global scores for our ages show roughly a similar pattern but, in the later stages, the absolute percentages of our sample lags behind that of the Kohlberg sample. For example our thirteen-year-olds exhibit a pattern close to the pattern of ten-year-olds in the Rohlberg sample. At seventeen, in our sample, stage-2 reasoning has not declined nearly to the extent it has in the 45 46 Table 9. p.tests for Mean Differences in (Weighted) Global Scores for Each Grade. x 0'2 n“ df c p Grade 7 2.31 1.09 104 206 2.63 <.Ol Grade 9 2.68 0.94 104 206 2 .25 <.05 Grade 11 3.02 1.43 104 a104 equals number of S; (26) times number of KHJI situations used (4). 10 0 ml “4 0 ;§ 6 6 ‘ 3 1 1 2 3 4 5 6 Stage Grade 7 ($2.35) Figure 1. 47 12 ”EFT 6 67; - 5 5 2 1 2 3 4 5 6 1 “if I. 5 r Stage Stage Grade 9 Grade 11 (332.61) (xx-3.27) Distributions of Se by Stage of Moral Development (Global- Global Score) Within Each Grade Level. 48 30“ 20-1 I we / bé—x—o 1 Percent of Total floral Statements I -----06 5 d,,/"' 10 f3 16 Age Figure 2a. Mean Percent of Total Moral Statements of each of 6 floral Judgment Types at Three Ages. (From Kohlberg, 1964, p. 403) m 3 o 040—1 a: H a .o .9. g, 30‘ 'o m U .m .3? o 20" 3 In o ‘3 m 10- u H m a. \ ‘.\ \ \ ‘02 3 1 4 3 \ \ \ s 93’ l \‘o Figure 2b. '6 “- ’,--—-'“.6 I I {7 Percent of Global Scores for Se at Three Ages in This Study. 49 Kohlberg sample at age sixteen. This suggests that either the measurements differ or the samples differ, or both. Although the percentages referred to in Figures 2a and 2b are not precisely equiva- lent, they substantially represent the same thing: percent of stage- typical moral reasoning at a given age for a given sample. If we assume that the differences between the two samples are actual develop- mental differences and not measurement error, It is apparent that our sample is somewhat slow to develop in the moral sphere. upgal deyplopmentI cognitive development and socio-economic level. The question of the relationship between cognitive development and moral development is answered by Kohlberg in his suggestion that level of cognitive development is a necessary but not sufficient criterion for attainment of moral stages. That is, a given level of cognitive ability is required for reasoning at a particular moral level but the ability to reason at that given level of abstraction does not automa- tically mean that the individual will attain the moral level: Other social-situational factors are required. In our sample, the correlation between level of cognitive development (as measured by the raw PPVT score which is roughly equivalent to MEA.) and level of moral develop- ment (Moral Maturity score) is .58. This general size of correlation is found within each grade level as well. The correlations between cognitive levels and moral levels for the 7th, 9th, and 11th grade Sp are .58, .49, and .66, respectively, thus confirming a moderate trend for attainment of higher moral levels with higher cognitive levels. However, as Haccoby (1968) notes, certain social-structural variables are also related to progression through the stages of moral 50 development. The suggestion in Rohlberg's theory of identification is that the major difference between children from different social classes is not that they acquire different values but that the rate of progress through the stages differs. Kramer (1968) found this to be the case in his study of adolescents and young adults. When Se in our sample are dichotomized in High and Low groups based on parents' jobs and educational levels (socio-economic levellz), a two-way ANOVA on Moral Maturity scores (Table 10) confirms main effects for both grade (age) and socio-economic level without any significant inter- action. In all cases, the mean Moral Maturity score for the High socio-economic Se is higher than that for the Low socio-economic Ss. value system stability The Rokeach Value Survey was administered twice to each S_at a three week interval to ascertain the stability of each st value rankings. For both the terminal and instrumental values, Spearman rho correlations were computed between the time-one and time-two rankings. For the entire sample (H-78), the median stability coefficients (rho) are .712 for terminal value systems and .673.for the instrumentals.l3 The correlation.(p) between terminal and instrumental value system stability is .477. As with Mbral'Maturity scores, mean value system stability coefficients were compared across grade levels and socio-economic levels. These ANOVA's are summarized in Table 11 (terminal stabilities) and Table 12 (instru- mental stabilities), indicating that grade level (age) has a significant effect on both terminal and instrumental value system stability. The mean stability coefficients for both terminal and instrumental value 51 Table 10. Analysis of Variance on Moral Maturity Scores for Entire Sample by Grade and Socio-economic Level. Source df MS F p Socio-economic level (A) 1 1030.4 11.88 <.005 Grade (B) 2 500.9 5.78 <.005 A X B 2 136.6 1.57 n.s. Within cell 72 86.7 - - 52 Table 11. Analysis of variance on Terminal Value System Stability Coefficients (rho) for Entire Sample by Grade and Socio- economic Level. Source df MS F p Socio-economic Level (A) l .0025 .06 n.s. Grade (B) 2 .1341 3.31 <.05 A X B 2 .1109 2.73 n.s. Within cell 72 .0405 - - 53 Table 12. Analysis of variance on Instrumental Value System Stability Coefficients (rho) for Entire Sample by Grade and Socio- economic Level. Source df MS F p Socio-economic level (A) l .056 1.29 n.s. Grade (B) 2 .192 4.46 <.05 A X B 2 .001 0.02 n.s. Within cell 72 .043 - - ‘ \: 54 systems show consistent increases as grade level increases (Figure 3). The main effect of socio-econoudc level and the interaction effect were nonsignificant for both terminal and instrumental stabilities. (Tables 1.H.& 2.H in Appendix H list mean stability coefficients for all cells in a Stage X Grade table). In H1, I predicted that value system stabilities would increase with increasing moral development. Figure 3 indicates stability increases with age. Figure 4 indicates a similar pattern of increasing stability over stages of moral development. The general trend is for higher value system stability to be related to higher stages of moral reason- ing. However, since the higher moral stages occur more frequently at the higher grade levels, a two-way ANOVA was run to determine the main effects of stage and grade and interaction effects, if any. For these analyses, only Sp at stages 2, 3 and 4 were retained in order to eliminate empty cells (there are no stage-1 Se in the eleventh grade and no stage-5 Se in the ninth grade). Tables 13 and 14 summarize the ANOVA's for the terminal and instrumental stabilities. As may be seen with this restricted sample, the main effect of grade on the terminal stabilities is no longer significant (as it was in Table 11 which utilized all 78 Se). More importantly for H1, the main effect of stage of development is non- significant for both terminal and instrumental value systems. This tends to disconfirm H1 and to rather suggest that value system stability is most strongly related to differences in chronological age which, in turn, suggests that such factors as reading ability, vocabulary size, familiarity with value terms, and the like may be influencing the stability of the value rankings rather than any underlying intrinsic value instability 55 3 .8- I: 0 q-l 1'3 Terminal $2! .7__ Instrumental :9 e. H ...g :1 a ,9 .6 Q U cn o m :13 .51L 1;I l I I 7 9 11 Grade Figure 3. Mean value System Stability Coefficients (rho) by Grade Level. 56 .8" u Terminal : £3 .7“ .3 0 Instrumental 3:.“ U 0 ° .6-+ >. U 9H H mil '3 a; .5“ 6 fl :2! .4'- If I T i l 1 1 2 3 4 5 Stage of Moral Development Figure 4. Mean value System Stability Coefficients (rho) by Stage of Moral Development. 57 Table 13. Analysis of variance on Terminal Value System Reliabilities by Stage (2, 3, & 4) and Grade. Source df MS F p Stage (A) 2 .068 1.86 n.s. Grade (B) 2 .063 1.73 n.s. A X B 4 .019 0.52 n.s. Within cell 55 .036 - - 58 Table 14. Analysis of variance on Instrumental value System Reliabilities by Stage (2, 3, 64) and Grade. Source df MS F p Stage (A) 2 .051 1.41 n.s. Grade (B) 2 .123 3.37 <~05 A X B 4 .014 0.38 n.s. Within cell 55 .036 - - 59 specifically related to developmental stage of moral reasoning. Likewise, H2 is disconfirmed in that the relationship between stage of moral development and value system stability is not stronger for the instrumental value system'with its moral values than for the terminal value system. This suggests that the moral values of the Value Survey are not being differentially responded to at the different stages of moral development. In H3, it was predicted that these moral values would differen- tiate among individuals at the different moral stages more strongly than would the non-moral instrumental values. As a measure of a value's ability to differentiate among various stage groups, the Xruskal- Wallis S statistic was used as a descriptive index of the degree to which the moral groups differentially ranked the value. The more stage-groups differed in their rankings, the larger S,becomes. If these S's for each instrumental value are then rank ordered from largest to smallest, a Mann-Whitney S’statistic may be derived to test the null hypothesis that the rank sums of the moral and non-moral value Sp do not differ. If the moral values are differentially ranked by the stage-groups more so than the non-moral values, the S's for the moral values should be consistently larger, resulting in a signi- ficant n. For the entire sample and for each grade individually a S was derived. In none of these four tests was the S’significant (the S_for the largest difference in the predicted direction was 25, which is not significant at the .05 level) and the rank sum differences were in the predicted direction only in two of the four cases. Thus H3 was not confirmed. 60 Value Structure and Value Sontent The structure of an individual's level of moral reasoning has implications for the content of the individual's values. Thus, we predicted that the content of our culturally homogeneous sample would be related to stage of moral reasoning and that, in some cases, value peaks could be predicted. The more general relationship between structure and content is predicted in H4.1, which suggests that for individuals who share a common culture, there will be more value similarity within moral development stages than between stages and that the pattern of value similarity will reflect the moral development sequence. To test this hypothesis, it is necessary to determine the extent of value similarity between each and every S_and then to determine the average value similarity among Sp within each stage and between stages. If rho correlations are used as indices of terminal and instrumental value system similarity between Sp, we would expect highest correlations be- tween S9 at the same stage of development and increasingly lower correlations between Sp at increasingly distant stages. Thus, a matrix of the average value system (either terminal or instrumental) correla- tions (rho) between and among S9 at various stages of development should show decreasing correlations as we move away from the diagonal. Since our interest is in stage-related value similarity as opposed to age- or grade-related similarity, we will test H4.l at each of the three grade levels independently. The first question to answer is whether the S9 at each moral stage are ranking the values independently or whether they are using similar standards to rank the values. A Kendall Coefficient of Concordance (E) 61 was computed as a measure of both terminal and instrumental value system similarity for each moral stage group at each grade level. Each Liwas tested under the null hypothesis that the S’sets of value rankings were independent. A significant fl.indicates that the Se are, to an extent, using the same standard in ranking the 18 values. Table 15 summarizes this analysis. Tables l6, l7, and 18 contain the matrices of average terminal value system correlations between and among S9 at the various stages of moral development for each of the three grade levels. Likewise, Tables 19, 20, and 21 show average instrumental value system correla- tions. The significance tests in Table 15 indicate that Ss at each stage (exceptions in eleventh grade sample noted) are applying similar standards in the ranking of the terminal and instrumental values. However, our specific hypothesis related to the matrix patterns is strongly disconfirmed. Not only are the values of Se at a given stage often more similar to the values of S9 at different stages than to the values of same-stage fellows, but also this similarity often incpeases as the distance between stages increases. The hypothesized pattern is evident only for the instrumental values among the seventh grade Sp, (Table 19) with only one cell not in the predicted direction. However, we must conclude that the developmental dimension of moral reasoning is not evident in the global measure of value system content similarity derived from the value Survey. That is to say that Ss in our sample do not show a reliable indication of value similarity based on stage of moral development. 62 Table 15. Kendall Coefficient of Concordance, S3 for Each Moral- stage group at Each Grade Level at time-l. Stage 1 2 3 4 5 Terminal - .24:1' .37**:* .23 .40“ ll Instrumental - .23 .24 .28 .24 0 m 'k *‘k '6 Terminal - .32 .30 .34 - g 9 Instrumental - . 19** .32** .33** Terminal .34: .303 17:“ - - 7 Instrumental .33 .27 .47 - - aSis a linear function of the average rho correlation among each set of rankings. *p<.05, **p<.01, ***p<.001. 63 Table 16. Average Correlations (rho) of Time-1 Terminal Value Systems Between and Among S9 at Stages 1, 2, and 3 in Grade 7. Stage 2 n- 6 10 1 .21 8’0 m 2 .24 .22 U a: 3 .22 .18 .13 Table 17. Average Correlations (rho) of Time-1 Terminal Value Systems Between and Among Ss at Stages 2, 3, and 4 in Grade 9. Stage 2 3 4 n- 12 6 6 2 .26 o a“ 3 .15 .16 U en 4 .07 .24 .21 65 Table 18. Average Correlations (rho) of Time-l Terminal Value Systems Between and Among S9 at Stages 2, 3, 4, and 5 in Grade 11. Stage 2 3 4 5 NI 8 8 5 5 2 .13 3'0 3 .17 .28 3 m 4 .14 .08 .04 5 .20 .24 .07 .25 66 Table 19. Average Correlations (rho) of Time-1 Instrumental Value Systems Between and Among S9 at Stages 1, 2, and 3 in Grade 7. Stage 1 2 3 n- 6 10 6 1 .20 0 m1 3 2 .11 .19 to 3 .06 .25 .36 67 Table 20. Average Correlations (rho) of Time-l Instrumental Value Systems Between and Among S9 at Stages 2, 3, and 4 in Grade 9. Stage 2 3 4 n- 12 6 6 2 .12 o co 3 3 .19 .18 a: 4 .13 .22 .20 68 Table 21. Average Correlations (rho) of Time-l Instrumental value Systems Between and Among Ss at Stages 2, 3, 4, and 5 in Grade 11. Stage 2 3 4 5 n- —'8 8 5 5 2 .12 o 3 .17 .13 3.“ a: 4 .14 .16 .10 5 .14 .18 .22 .08 69 This failure of H4.1 to be confirmed may be attributable to a number of possible factors. One possibility is that the S9 are responding to the value Survey not according to their own value preferences but rather according to what they think they ppgpp_to value. That is, the values may be being perceived by the Ss as cultural norms or impera- tives and they are ranking them in order of perceived social desirability. Then, to the extent the Ss perceive the same cultural value norms, they will tend to rank the value labels in the same way. Another possibility is that the structural aspects of moral reasoning are not, contrary to Kohlberg's implication, strongly related to specific value content even in a homogeneous group. A third possibility is found in Rokeach's notion that values related to morality are but a subset of the instru- mental values and thus similarity in value content among the stages is related only to this subset. In this case, the global measure of value system similarity used in the preceding analysis may be "washing out" moral value similarity with the inclusion of the additional, unrelated values. We found previously that the moral values specified by Rokeach do not, in fact, seem to be discriminating among the stages better than the non-moral values, suggesting that an analysis of the moral values alone would not result in a different conclusion. However, value pre- dictions under H4.2 include seven terminal and seven instrumental out of the 36 values, suggesting a subset of morally related values but a subset different from those suggested by Rokeach. If the notion that the stage-typical ethical theories of the Kohlberg typology speak only to a subset of the values in the value Survey, then the ineffectiveness of the global similarity measure is not inconsistent. However, the specific predictions of H4.2 must be 70 demonstrably accurate if we are to conclude that value content is, indeed, predictable from the underlying structural variation. In H4.2, I specifically predicted at what stage or stages a particu- lar value would be most highly ranked. Only fourteen values (7 terminal and 7 instrumental) from the Value Survey were directly related to the moral judgment concept of Kohlberg. Table 18 and Table 19 contain the value peak predictions (originally given in Table l and Table 2) as well as the actual value peaks as they occurred in my sample. The correctness of each value peak predic- tion is contained in Tables 22 and 23. The more general question which needs answering is that of the overall success of my predictions. If we assume that each prediction is independent of the others and that there is an equal probability associated with guessing a peak correctly by chance, we may determine the probability associated with making as many or more correct predictions by chance.14 If this probability is small, we may reject the null hypothesis that chance alone is working and feel reasonably confident that, given knowledge of moral stages, our ability to predict value peaks is enhanced. For purposes of getting the best estimate of value medians for the entire sample, time-1 and time-2 value rankings were combined and grand median ranks computed for each value for each stage-typical group. Table 24 presents the data on correct peak predictions and the probabilities associated with getting as many or more correct by chance if p(correct) - .2. These probabilities indicate that we may reject the hypothesis that p(correct) - .2 in favor of the alternative, p(correct))».z. That is, our ability to predict value peaks given stage of moral development is considerably better than chance. 71 Table 22. Predicted value Peaks and Actual value Peaks (time-1 & time-2 combined) Across Stages of Moral Development for Entire Sample (u-78). Termina 1 Fl_ A Comfortable Life Equality Family Security Q * Freedom 69 National Security * Pleasure Sa 1va tion @ e [ oeoeju W Broadminded Forgiving Helpful 8 8 * e G *1- Loving Obedient Responsible Self-controlled |®® pe * - predicted value peak(s). C) - actual value peak(s). 72 Table 23. Predicted Value Peaks and Actual Value Peaks (time-l only) Across Stages of Dbral Development for EAch Grade Level. Grade Level C / A. \.- \17 \ 7 9 11 Stage Stage Stage Terminal 2 r_3y‘ 2 3 4 p _L 3 FILL A Comfortable Life m 0 lav—F— * O Equality @ 9 O * Family Securit @ Q Q 0 * Freed O * Q Q G National Securit G O * O * Pleasure Q Q 9 mm» Le _Le._oem W _. . Broadminde Try “'6' O * Forgivin * ® 3 Helpfulo * O * O * ® Lovin O * Q 0 * ObedientQ ' 8 O * G Respons ib le 0 * self-concroue L—L...1. ...Qim. 9.9.13.4}. * - predicted value peak(s). 0- actual value peak(s). 73 Table 24. Number of Correct value Peak Predictions in Table 22 and the Associated Binomial Probability Where the Probability of a correct prediction is 1/5. Number of Number Prediction Predictions Correct p Terminal j q 7 4 l 4.05 I Instrumental Combined 74 Here, as before, an age factor is correlated with stage of development so that our stage predictions are, to a degree, age pre- dictions. For this reason, peaks were predicted within each grade level separately in Table 2. Table 25 summarizes the probabilities associated with the correct predictions in Table 23. When age is controlled, it is apparent that our predictive ability is somewhat diminished. The major predictions hold up well for the ninth grade Sp, somewhat less well for the seventh grade Se, and almost not at all for the eleventh grade Ss. However, H4.2 is generally supported by the data. The failure of our predictions to hold up strongly in the eleventh grade sample may be partly explained in Tables 18 and 21. Mote that the three concordance coefficients which failed to achieve significance are in the eleventh grade sample (stage 4 terminal and instrumental and stage 5 instrumental). These low average intercorrelations were not sufficiently large enough to allow us to reject the hypothesis of independent value ranking for the Sp at these stages. Obviously, as group value similarity decreases, it becomes more difficult to predict accurately. In a similar vein, the partial failure of our predictions in the seventh grade instrumental values may well be the result of the fact that five of the seven major peak predictions are for the highest stage (3). It should be apparent that the ranking procedure used in the value Survey does not allow for measures of absolute value but rather for relative rank order. Thus, if some values are to be ranked higher, other values must assume the lower ranks. So, perhaps it would be unlikely that stage 3 S9 would rank all the predicted values higher than the other stages. 75 Table 25. Number of Correct value Peak Predictions within Each Grade (Table 23) and the Associated Binomial Probabilities. Grade 7 9 ll p(correct) - 1/3 1/3 1/4 Number of Number Number Number Prediction Predictions Correct Correct p Correct Terminal 7 5 <.05 6 <.05 3 n.s. Instrumental 7 3 n.s. 5 <=.05 2 n.s. Combined 76 The support received for H4.2 suggests that the Value Survey is, in fact, tapping value content differences related to the underlying structure of an individual's development stage of moral reasoning. This support of H4.2, while H4.l was not supported, lends credence to the idea that moral judgments are related to a particular subset of values. However, as has already been noted, it is not just the moral instrumental values (which Rokeach cites) which are strongly related to moral judgmental dimensions. To further pursue this question of the moral values and to gain more insights into the relationship between values and moral develop- ment, additional analysis of the value data was carried out. This further analysis relates to the ability of the Value Survey to discrimi- nate among individuals at the various stages of moral development. We have seen that relative value peaks can be predicted among the stages, but this says nothing about the absolute value differences between individuals at the various stages. Given the relatively small sample and the cultural homogenity of the sample, we would anticipate a high degree of value similarity. However, the individual differences implied by the moral judgmental differences require the value Survey be able to discriminate among the various stages if it is, in fact, tapping structural developmental differences and not just knowledge of cultural value labels. At this point, it is important to recall the general exploratory nature of this work so that our following analysis of the value Survey responses are not misinterpreted. As Bakan (1966) relates, Berkson suggested that, given a sufficiently large S5 almost any division of 77 subjects will result in chi-squares sufficiently large so that the prob- ability will be small beyond any usual limit of significance. Thus, in this work, given enough subjects, we would find that each and every value discriminates among S9 at different stages or grades at a statistically significant level. For this reason, my intent is not to "prove" or "disprove" that the value Survey discriminates among the S9 but rather my purpose is to determine how well the value Survey discriminates among various age and developmental levels. The specific generalizability of the finding in this study is limited by the restricted nature of the sample. However, this restric- tion also has the effect of minimizing value differences among the Se, thereby making the Value Survey's task more difficult. Thus, the significant findings are indicating robust relationships within our sample which would be magnified by a more heterogeneous sample. Also age-related differences which are found would be more readily inter- pretable as age-related in that cultural and intellectual variation is well matched among the three age groups. To measure the ability of the value Survey to discriminate among the age and stage groups, the Kruskal-Wallis analysis of variance (Siegal, 1956) was used to indicate what values were being ranked in systematically different ways across ages and stages. It is my intent to use the S_statistic and its associated probability as a descriptive statistic since as the rank order differences between groups become larger, the Siincreases and p_decreases. Although a strict usage of the probability level would allow only the making of inferences to the populations from which our samples are drawn, the relationship between 78 the probability and sample differences makes it a useful descriptive statistic. Thus, a smaller p_for value X than for value Y indicates a greater amount of systematic rank order differences among the groups for value X than for value Y. For descriptive purposes, it was decided that a p‘<.10 indicated a difference in rank order of sufficient size to suggest that strong group differences in value rank actually existed in our sample. To evaluate value differences among the three age levels, it was decided that there were systematic value rank differences in our sample if the Kruskal-Wallis S for any value had a p<.10 for 3393!; time 1 299 time 2 25’s combined analysis resulted in p4(.05. A difference which occurred once might or might not be due to chance, so the retest with the value Survey allowed us to check if time 1 differences were repli- cable. Thus, large differences (p‘<.10) which occurred at both time 1 and time 2 are less likely to be random sample differences. Also a reliable difference in value rankings might, by chance, fail to meet our criterion of p-<.10 at one of the administrations. So that a value was not overlooked for such a reason, a probability less than .05 associated with value differences in a combined analysis was taken to indicate non-random robust value differences. If the combined analysis resulted in p‘<.01, this large group difference was considered at least as strong a discrimination as the criterion of p‘<}10 over both administrations. Given the non- independence of the two administrations, this criterion actually is more stringent. Table 26 indicates those terminal and instrumental values which best differentiate among the Se in the three grades. 79 .emmhamcm mocuoaoo um .ummhamcm mocqoaoo mow nonmascm «usage a armada soon you .udumflmcm «usamu u Haoamu nc.uVm .¢56z¢ healoeo madamznamxnamx wow ”nosam> sumo nosed o 8.Vm no ca.v.a .<>oz< harness saaasz-asxssnm see smu=u<> mass «mean a vocmoaoo any scum one summed scum summonses use names summer a ma a.aa a n.e m a.» r H.e ea r.aa ma n.sa as a.oa o m.a a A.“ e N.» ha m.aa ea a.~a ma n.sa n a.e e n.e n a.a e a.» as o.ea N a.e N n.e m ~.m e «.5 Ha a.o~ ma n.~a a n.a ea o.ma am m.ea ea m.~a no n.aa N ~.e «a o.aa r n.r as H.HH r o.a ea ~.ea ea m.~a as m.na as m.oa r n.ea we r.na we r.ea ma ~.~a e o.» n m.a n o.e is. sex 3.5 as... see .5: as sense a sense a sense sudden mammmmo mcu>amuoh cardamomoum mea>og moaaomucoouunom odomaeommom HzmnzmmeZH Aoq ensue: oeouum>~mm ZOHHquoufim AmA momma unencummn um mm_weoa< anon oummucouowmun noun: nosam> we charge mam“ ouuoomaoo mes scum memos: .oN «Home For our sample. W and W are the two terminal values which differentiate among age levels best. While .figpily_pgpppipy,remains fairly highly valued over all ages, there is a decided drop between seventh and ninth grades. SogLal pegogpipiop, although relatively low in the seventh grade, continues to drop off in later grades. Of the instrumental values, clea , lo 1 a1, and ingependepp seem to reflect age differences best. nggp_shows a sharp decline as age increases, dropping from second for seventh graders to sixteenth for eleventh graders. Logical shows a sharp rise, from 18th to 9th, from the ninth to eleventh grade. Independent shows a steady increase across the three age groups, going from 15th to 11th to 4th in importance as age increases. Less strongly, age-related increases occur with gppppg_lpyg, pepponpible, pelf-ppptpolleg, brpagggpdpd, and gppppl_; age-related decreases occur with palvatiog, foggiyipg, lpyppg, and pplppg, The strongest curvilinear trend occurs with p_pgppp_pj_ppggnpliphpppp, which rises from 11th to 8th in the ninth grade and drops off to 12th in the eleventh grade. The value rankings of the Ss were then analyzed by means of the Kruskal-Wallis analysis of variance technique according to their stage of moral development. The same criteria for determining reliable value differences between groups were maintained: a strong, reliable differ- ence was to be inferred if ps<310 for Sppp,time l and time 2 value rankings or p‘<.01 for the combined analysis; a somewhat less strong and reliable difference was inferred if p«<.05 for the combined analysis. 81 Table 27 summarizes these findings. Where linear trends for median or rank differences seemed to be the rule across ages (Table 26), the pattern across stages has become somewhat more complex. As our specific value peak predictions in Table l and Table 2 made clear, simple linear increases or decreases in value rankings across the stages of moral reasoning were not anti- cipated. Freedom, for example, is ranked 1st by the first two stages, drops off somewhat at stages 3 and 4, then returns to 1st rank at stage 5. Looking at Table 27 in context of our previous findings, one terminal value, freedom, and one instrumental value, obedient, stand out. Not only do these two values discriminate best among the Ss at different stages of development, they discriminate in a way consistent with the moral development typology. In Table 22, the combined time 1 time 2 value data, freedom was correctly predicted to peak at stages 2 and 5 and obedient was correctly predicted to peak at stages 1 and 4.15 When the value data for each grade level and each time is separately considered, there are six opportunities to predict peaks (one peak prediction for Se at each of the three grade levels for both time 1 and time 2) for each value. Table 23 shows the within-grade peak predictions for freedom and obedient were correct at two of the three grade levels at time 1. Time 2 data (not shown) indicates that the predictions for freedom and obedient were correct at all three levels. Thus, Sppggpp, and obedient not only strongly differentiate among Sp at the various stages of moral development, they differentiate in a consistent, reliable manner over all S9 or for S9 within each grade. These two values will be considered in more detail later in this section. 82' .9352; scam—sou um no.Vm .¢>oz< assume—o mugfimzuaoxmsmx now .3232; momma—moo mom Ho.Vm Hm confines «1063 a 7063 soon no.0. 3.V .— .<>oz< assumes unaszaamxmsux you unmade» mono human 0 «mma<> was "Emma A .uumhdmcm «use: 1 .7083 ocean—Eon 05 Beam mum muoouo 3cm."— ouamomaoo men 92:3 53va m ea o.oa ma m.HN e n.N ca n.e e n.e meeeawuem a N.a m m.m N n.e n m.N as o.oa ssoauaea< we n.ea e a.» ea N.Na ea a.ma ca o.oN azmaamno amucmasuumcH ea n.ma n o.a Na n.na r a.» m n.n asses: No votes e. N o.m ma N.Ha ea r.aa ma a.ma Ha o.aN sosnssn-masm m n.n we n.sa NH n.aa oa a.ce N n.N aseusosm aseoassz we e.Na e n.e Na n.aa ea N.Na n n.N assass>~nm Na n.aa om N.a ca m.oa Na n.na ma m.NH uzazrsm mezzo a o.N N N.n N N.« a o.m a N.m alderman .wmmmmmmm area .5: seem .5: arse .eaz mere .5: use .2.: 0 IF IF 1. M] I G n owmum e ammum m owmum N ommum H swoon .ucuflmo.—0>on any“: we mowmum ucuuowma um am 9392 anon «omen—commons some: 33.3 no muoouo scam someone—o0 men mean suave! KN 0.3mm. 83 Age-related and Stage-related value Differenges Table 26 and Table 27 present an interesting picture when considered together. Only two values, salvation and forgiving, appear on both lists as values which discriminate well among both age-groups and moral stage- groups. The remaining values differentiate among g9 strongly on only one of the variables. This implies that even though there is a strong correlation between age and developmental level of moral reasoning, the values which define age and stage differences do so independently. Obviously, such a statement requires statistical support which, unfor- tunately, is not forthcoming.16 The age-related value differences also give rise to speculation that there is a developmental pattern of value change related to inter- personal behavior which is distinct from the moral judgmental dimension of Kohlberg. Note in Table 26 that five of the seven instrumental values which (in Table l) were thought to be related to the moral dimen- sion (responsible, self-controlled, lo in , broadminded, and forgiving) discriminate well among different age gs. Only one, forgiving, also discriminates well among stage-groups. The other four values are among the least-best stage-group discriminators of the instrumental values. These findings generally suggest that values are more a product of the sub-culture of age group than of the underlying structural varia- tion which the Kohlberg typology is so intimately related. Thus, Kohlberg's own suggestion that the value content for individuals below his post-convention stages is largely culture-bound is given some support. However, as Tables 22 and 23 indicated, predictable value variation occurs across the stages of moral development when the cultural variation is at 84 a minimum. In another culture, predictable value variation should also be found, but the specific values involved might or might not be the same as we found. Additionally, we would predict that fgeedom and obedient would consistently vary with stage of moral development across cultures although their relative overall ranks would vary from culture to culture. A word must be said about salvation and forgiving, which are the two values which are ranked discriminably different by both stage and age groups. Considering that Rokeach (1969) found galvation and forgiving to be the most distinctively Christian values in a representative sample of adult Americans, it is not surprising that children growing up in the culture would show difference on these values as they develop. Neither is it surprising that the moral dimension would also show distinctions on these values. It is here that the interaction between stage and age might be most revealing. It should be noted that the hypothesized relationship between stage of moral reasoning and value ranking was as reliable for salvation (see Table 22 and Table 23) as any value. The predicted peaks for salvation were correct for the combined analysis and for the time-l within-grade analyses. We should also note in Table 27 that inner harmony discriminates strongly among gs on the moral development dimension. Initially, there seemed to be little to suggest that inner harmony was a value which would reflect the structural variation found in Kohlberg's schema. However, following the median ranks across the stages of moral develop- ment, it appears that the pattern for inner harmony is much like that for obedient. Also for both the time 1 and time 2 data, there is a 85 significant (p‘<.05 one-tailed) correlation between inng;_hg;ngny,and obedient. These correlations are not large (r1-.2051; r2-.2226) but support the suggestion that obedient and inne; hagggny tend to be related to the moral judgmental dimension in a similar manner. Laggg, harmony, while showing a general pattern across the stages similar to obedient, does not show the extreme variation of obedient. Inner harmony also tends to be more highly ranked by the conventional stages, 3 and 4, but without the severe drop at stage 5. The interpretation of igggg_ harmogy as a value reflecting some underlying structural variation is confounded by the lack of any consistency in the within-grade analyses. For grades seven, nine, and eleven, inner hgggggy peaks at stages 1, 4, and 2 respectively. This inconsistent pattern suggests that there may be some strong interaction between moral stage and age for these gs. EEEEDGH and OBEDIENT as Defining moral values It was noted in the last section that freedom and obedient were not only the strongest, most reliable values for discriminating among §s at the various stages of moral development, but these values also were the most predictable and consistent values related to the moral development typology. This suggests that, within the gs' culture, the structural differences encountered in the development of moral reasoning are, to significant extent, reflecting an individual's value of freedom as a desirable end-state of existence and his valuing of obedient as a desirable mode of behavior. It might be suspected that freedom and obedient are value-opposites and are tapping the two directions of a value continuum. Here this the 86 case, we would expect a significant negative correlation between Ss' ranking of freedom and obedient. While the correlations between fgeedom and obedient for both time 1 and time 2 are negative, neither is suffi- ciently large to reject the hypothesis that the population correlation is zero.17 The median ranks for freedom and obedient across the stages of moral development (for the combined time l-time 2 analysis) are given in Figure 5. For freedom, the grand median is 4.0 and for obedient, 12.0. Recalling the stage-typical descriptions of Kohlberg's moral stages, it seems that §s at stage 1 would be the most likely to be responding to the values on the value Survey as merely cultural labels and rules rather than as personal preferences. This is a result of the relative undifferentiated, unintegrated structure which, in moral matters, defers to the superior power. It is most difficult at this stage to say that an individual "has a value" in that value preferences are inextricably tied to the authority present in the situation. Since the value rankings took place within the school before an adult observer and the §.identified himself on the value Survey, we may assume that effects of acquiesence will be maximum at stage 1. The approval motive may be strong at stage 3 as well, thus further confounding the responses. However, it is assumed that these §s will be ranking the values in a socially acceptable way to some degree. To the degree that socially desirable responses are given, the cultural homogeneity of the S9 will reduce stage-related response differences making the value Survey's task that much more demanding. We have already seen valid stage- related value differences occurring, assuring us that these children 87 Freedom lLllllJ n-u—or-or-I wNHOOmNOUJ-‘th—I Lil ll 1 median Rank Order 16 - Obedient l 2 3 4 5 Stage Figure 5. Median Ranks for Freedom and Obedient by Stage of Moral Development (combined time-l - time-2 analysis). 88 are not responding to the value Survey with socially desirable responses only. The high value of freedom at stage 1 is partly a response to a label highly valued by the culture and may be considered as a good estimate of the socially desirable response. Thus, deviations from this point may be thought to represent stage-typical value differences. Obedient, however, was predicted to peak at stage 1 (and at stage 4) not because the culture values obedient but because the stage-l §fs morality is one of obedience to the power and authority. Moving to stage 2, we find that one of freedom's peaks was pre- dicted for stage 2 due to the hedonistic orientation of the gs at this stage. LThe increase in median over stage 1 is very slight but it is nonetheless higher. Obviously, since freedom is already high at stage 1, there is little room to move upward and the median differences could not become too large in any event. However, while the median differ- ence is slight, the percent of stage-2 gs ranking freedom first is twice that of stage-l §s who rank freedom first (232 for stage 2; 12% for stage 1). \ The higher ranking of freedom at stage 2 is coupled by a sharp drop in the median rank for obedient. Not only do the stage-2 hedonists value freedom more highly, their instrumental relativistic orientation reflects a relaxation over concerns with the consequences of rigid deference to authority found at stage 1. ’_ At stage 3, this valuing of obedient is little changed from stage 2. However, freedom has taken a (relatively) sharp drop, falling over two full ranks. Given the strong cultural emphasis on freedom and the emphasis at this stage for being "nice" and maintaining approval by 89 showing loyalty and conformity to the established order, this drop in the median rank of freedom is an even more compelling indication that the label "freedom" is not being responded to in simply a socially- desirable manner, but is being responded to, at least in part, as a personal standard reflecting the morality of this stage which, in stressing conformity and approval, de-emphasizes personal freedom. The failure of obedient to rise much from the stage-2 level reflects this stage's concern with maintaining a friendly, flexible interpersonal concordance in moral matters rather than a rigid, strict rule orientation such as found at stage 4. This stage-4 rigid rule orientation emphasizing doing one's duty and respect for and obedience to authority was predicted to reflect a high value for obedient. The data in Figure 5 confirms this. Of all the stages, obedient is ranked highest at stage 4. Freedom drops slightly as might be expected. The slightness of the drop may be due to the strong cultural influence which, in this sample, keeps freedom relatively high for all §s. The development of the child's morality from the conventional stage-4 morality to the autonomous principled level of stage-5 morality finally has polarization effect on these two values: freedom rises to its highest and obedient falls to its lowest. In fact, freedom now has the highest median rank (supplanting a world at peace) and obedient's median is eighteenth in order (down from sixth in order at stage 4). These values are reflected in the stage-5 moral concern with individual rights and responsibility and an awareness that personal obligation, outside the legal sphere, is a result of freely entered agreement. The 90 emphasis is on democratic agreement in the legal sphere and on the individual's personal standard evaluated within a social-contract orientation. Thus, freedom for self and others is a primary pre- requisite while the notion of unquestioned duty and respect for authority within the value obedient has become outmoded and useless. Once again, with five of the six stage-5 §s in the eleventh grade, a strict stage-interpretation unconfounded with age is difficult. Sufficient numbers of stage-S moralists at all grade levels would aid in our interpretation. However, the fact that our value peak pre- dictions held up fairly well in the eleventh grade even without con- sidering the stage-5 moralists suggests that stage of moral develop- ment is more strongly related to these value differences than age, although some interaction between stage and age would not be unlikely. Rank-difference Score As Figure 5 indicates and the discussion above endeavored to point out, neither freedom nor obedient alone are as indicative of stage of moral development as when they are considered together. A simple ‘measure which utilizes both value ranks was later devised to further simplify the analysis. The Rank-difference Score is a single number indicative of the distance between the ranks of freedom and obedient.18 Table 28 summarizes the mean Rank-difference Scores for the five stages and Table 29 summarizes an ANOVA on the Rank-difference Scores by stage and grade level. Only the main effect of stage of moral development is significant in this analysis. The absence of any strong age-related effect on the Rank-difference Scores and the absence of any strong interaction effects gives further 91 . Table 28. Haans and Standard Deviations of Rank-Difference‘ Scores by Stage of Moral Development. Stage 1 2 3 4 5 n 8 3O 20 14 6 Mean 5.38 8.53 5.20 2.36 8.83 3.0. 5.96 5.52 6.25 6.52 7.84 a Rank-Difference score - (rank of obedient)-(rank of freedom). 92 Table 29. Analysis of variance on Rank-Differencea Scores by Stageb of Moral Development and Grade Level Source df HS F p Grade (A) 2 25.3 0.65 n.s. 'Stage (3) 2 142.2 3.64 (.05 A X B 4 25.5 0.65 n.s. Within cell 55 39.1 - - a Rank-Difference score - (rank of obedient) - (rank of freedom). b §s at stages 2, 3, 5 4 only included in this analysis. 93 support to the notion that freedom and obedient are values related to the moral development dimension. Even when all S; are used in an analysis, the mean Rank-difference score for the ninth grade gs (5.5) is well below that for both the seventh grade (6.5) and eleventh grade (6.9) §s. Thus, there is not even a hint of any linear relation- ship between age and Rank-difference Score as might be anticipated. From Table 28, we see that, in terms of Rank-difference Scores, stage 1 and stage 3 are similar and stage 2 and stage 5 are similar while stage 4 is quite unique. However, the similarity in Rank-difference Scores for stages 1 and 3 does not reveal the tendency for both freegom and obedient to be ranked lower at stage 3. Likewise, the Rank-difference similarity between stages 2 and 5 does not reveal the more extreme median positions of the two values at stage 5. There are, then, both quantitative and qualitative differences among the stages of moral development for these two values. It un- doubtedly would be possible to derive more elaborate and sophisticated methods for scoring value patterns which reflect and identify value aspects of such phenomena as structure of moral development. This, in fact, will be encouraged in the final section of this paper. However, it is sufficient for our purposes here to have explicated reliable and valid value components related to the underlying structure of moral reasoning. These more sophisticated and refined techniques will require samples more nearly suited to the purpose and considerably more sub- jects with which to work. Summary of Hypotheses Before moving to a final discussion, let us review the conclusions 94 related to the specific hypotheses. HI. H4.l H402 value system stability, contrary to this hypothesis, did not appear to show a strong linear increase with higher moral development. However, there was a strong indication that value system stability increases with age. A’53 .8 Instrumental value system stability, contrary to this hypo- thesis, related less to stage of moral development than did terminal value system stability. The moral values, identified by Rokeach as a subset of the instrumental values, did not, contrary to this hypothesis, differentiate better among gs at different stages of moral i development than did the non-moral instrumental values.{i,. w Contrary to this hypothesis, a global measure of value system similarity did not reflect the pattern of moral development.¥ . etr‘ This hypothesis that the content of values may be predicted from the structural component within a given cultural con- text was strongly supported. value peaks were predictable from stage of moral development and a number of values dis- criminated reliably among §s at different stages. Freedom and obedient, especially, were related to moral development. I DISCUSSION Kohlberg (1968) makes the point that moral content or value for individuals at the preconventional or conventional levels (stages 1 through 4) is largely accidental or culture-bound. However, in the higher stages, "Socrates, Lincoln, Thoreau and martin Luther King tend to speak without confusion of tongues, as it were. This is because the ideal principles of any social structure are basically alike, if only because there simply aren't that many principles which are articu- late, comprehensive and integrated enough to be satisfying to the human intellect (p. 30)." Rokeach (1968), in his conceptualization of adult values, also suggests that there are only a limited number of means and ends which the individual holds and applies universally and consistently. In his definition of terminal and instrumental values, Rokeach makes clear that a "value" has application across situations, across individuals, across social systems, across objects. Thus, the values of Rokeach and the universal, comprehensive, consistent abstract principles which guide moral judgments at the post-conventional level (stages 5 and 6) of Kohlberg's typology function in remarkably similar ways. But, as Kohlberg implies in the quote at the beginning of this section, only a handful of men achieve the highest stages. Rokeach's value definition does not deal with those individuals who are unable to say that they believe a particular mode of conduct is 95 96 uniyersally preferable. With the value Survey, the individual who, in one sense, has no values as defined by Rokeach is operationally forced to generate a value system of value terms which may be meaningless for him. The Kohlberg typology, with its structural dimension, readily identifies those individuals for whom such a concept as justice is not universally conceived nor applied. Thus, in populations of individuals who are at the Kohlberg's stage 5 or 6, the value Survey would seem to be most valid for indicating differences in values which have universal application. For individuals at lower stages of development, the value Survey might be seen as forcing them to order some value terms which, in fact, may not reflect consistent, universal, and comprehensive prefer- ences within the individual's belief systems. These lower-stage indi- viduals might be seen as responding to the value terms merely as cliches' and perceived social norms which have little personal refer- ence or utility. Given the conceptualization and operationalization of the value concept by Rokeach, the preceding argument against the validity of the value Survey for those whose developmental stage of moral reason- ing has not reached stages 5 or 6 is not untenable. The purpose of the research reported in this paper was to ascertain whether or not that argument was supported in practice. I have concluded that the arguments above are not supported in practice. The Value Survey does tap developmental structural differ- ences in moral reasoning across pl;,the stages of the Kohlberg typology. The nature of the value-ranking procedure, in that it generates a value system for an individual whether or not one actually exists, 97 obviously makes it impossible to determine where one group would syste- \ matically rank all or most of the values higher than another group. Such an occurrence is suggested in comparing value responses of stage-6 individuals to those of stage-2 individuals. The absolute degree of internalization of all the values theoretically could not be tapped by the ranking procedure (unless stability of rankings was related to degree of internalization as suggested in H1.). The results of this work indicate, however, that even where we have apparently maximized the opportunities for minimal value differences, reliable value differences consistent with the complex underlying structural variation gp_occur. This indication of the validity (concurrent) of the Value Survey may not be interpreted to mean that the value differ- ences found "explain" all or even much of the moral judgmental differ- ences. The interpretation which is consistent with the findings is that certain values on the Value Survey validly reflect the structural dimension in the development of moral reasoning where other contaminating variables are reasonably controlled. Had our sample of 78 gs been drawn from a large, urban, socially and ethnically mixed school system, the value variations related to other socio-cultural variables might be expected to "wash-out" the relatively small stage-related value variation. Even though valid and reliable value differences have been found, the failure of some of our predictions raises further questions which this data does confront. For example, the failure of equality to peak for stage-5 §s in the eleventh grade and its failure to discriminate well among the various stage §p cannot be dismdssed without concern for .. __ 98 the implications these failures may carry. Eguality clearly is related to the moral development concept in theory and the discrepancies noted in this data should be further pursued in other studies in order to determine the age-stage interactions and the relationship of the value equality to the moral dimension both in children and adults in this and other cultures. The generalization we make from the data is that the structural differences in moral reasoning existing in adults and children are identifiable in value terms. That is, judgments made by individuals in moral situations do refer to both terminal and instrumental values which the individual holds. In moral dilemmas, an individual makes a judgment about the good and the right of action based not only on his preferences about those specific modes of behavior but also on his preferences for desirable end-states of existence which may be affected by the action in question. There is some factual as well as theoretical support for the suggestion the value concept as currently operationalized may not be maximally powerful for the younger adolescents or for groups with large numbers of preconventional moralists. While the Value Survey was tapping reliable and valid differences at these levels, they were not as overwhelming as the interview data would suggest they might be.. A modification of the value terms specifically for use with these younger and less developed individuals might well result in more compelling and powerful results which cannot be obtained with the current set of "adult" values. The value ranking format, however, seems entirely adequate if large amounts of data are being sought. Complication of 99 the current format might not give additional results worth the additional time and cost involved in administering and scoring. As an example, Penner, Homant, and Rokeach (1968) found that the rank- order procedure compares favorably with the more laborious paired- comparison method. Although the moral judgmental dimension and the value dimension as currently operationalized cannot be viewed as isomorphic, there is reason to believe that the more complex, tedious, and expensive inter- view techniques required for assessing level of moral development might be supplanted, in some cases, by a value-ranking procedure. In adults, the Value Survey would appear to be already taking the structural varia- tion into account. In children and adolescents, the particular value terms now in use may not be ideal for this purpose, but, as we have seen, are not insensitive to it. Specific value differences found in our data may or may not be generalizable to the other sex, to other communities, to other races, or to other time periods.19 The age-related value patterns (as opposed to the stage-related patterns) are particularly interesting and they may have implications for our educational and socialization practices. Further research to consider these aspects is encouraged. 1. FOOTNOTES The "behavior" in Havighurst and Tabs (1949) is based on adult and peer ratings of §fs character reputation rather than any observed behavioral criterion. "A belief is any simple proposition, conscious or unconscious, inferred from‘what a person says or does, capable of being pre- ceded by the phrase 'I believe that...'" (Rokeach, 1968, p. 113). A belief system, for Rokeach "represents the total universe of a person's beliefs about the physical world, the social world, and the self. (123)" An attitude, for Rokeach, is defined as "a relatively enduring organization of beliefs around an object or situation predisposing one to respond in some preferential manner (1968, p. 112)." Cf. Rokeach, 1968: "A value...is a standard or yardstick to guide actions, attitudes, comparisons, evaluations and justifications of self and others (p. 160)." Maccoby (1968), in discussing Kohlberg's theory: "If values are being internalized...should it not be true that they guide behavior? Should we not require that measures of moral judgment predict actual overt conformity, as a test of the validity of the moral-judgments analysis? Kohlberg recognizes the impor- tance of the issues involved in the mesh between moral values and 'moral behavior. He does insist that moral judgments are of im- portance in their own right--that society cares not only about what an individual does but cares also about the nature of the moral judgments an individual is able to make concerning his own behavior and that of others. He points out that law requires that the individual shall be able to distinguish 'right from wrong' before he may be punished for a deviant act and that, furthermore, the law judges behavior in terms of the intent as well as the consequences of an individual's actions. Therefore, it is impor- tant to understand the development both of concepts of right and wrong and of the ability to judge on the basis of intentionality, even if these aspects of moral development are unrelated to overt deviant or conforming behavior in specific situations. For Xohlberg, morality itself is defined more by the cognitive contents of moral judgments than by individual items of behavior. "Nevertheless, Kohlberg does concern himself with the corres- pondence between the two classes of phenomena...He cites evidence 100 101 to show that there are low, but positive, correlations between the maturity of moral judgments and resistance to cheating, teachers' ratings of conscience, and teachers' ratings of fairness with peers. He notes, further, that maturity of moral judgments is associated with a child's ability to resist ex- ternal pressure to commit an "immoral" action. Kohlbefg's _ interest in conceptual development stems in part from his assump- tion that thought and action are not really separable, and hence he is impelled to search for instances in which one can be predicted ‘from the other and to suspect that in cases of lack of corres- pondence the wrong thing has been measured (pp. 239-240)." The "moral values" specified by Rokeach (1969) are: clean, foggiving, h.__ef_el ul. hone t. levies. ob die t. MW. and self: gontrolled. The remaining instrumental values are competence values except popyageous and gheegful, which are indeterminate. "One wants to say that what makes a moral situation moral must be some element in the situation, whatever it might be, which itself is uniquely moral. Like a human who is said by some to be human because he possesses a unique human soul, so a situation is supposed to be moral because it has some unique moral component (e.g., some quality, relation, rule) in it. But when we look for the alleged moral component and fail to find its "moral-ness" we are puzzled. We think perhaps we ought to look harder. And perhaps we should. We also perhaps ought to look at the model we are using to answer the question "What makes a moral situation moral?" A moral situation may be less like a human with a soul and more like, say, a lion. A lion is not normally thought of as a lion because it has lion parts, but the other way around; the parts are labelled lion parts because they are parts of a lion. Similarly it may be that a moral situation is not moral because of any one of its parts (e.g., rules) but because of the way the parts are put together as a whole. Thus the rule (or value judgment) cited in a moral situation may not be what makes the situation a moral one. Rather, the rule (or value judgment) cited may be treated as a moral rule (or value judgment) because it is cited in a moral situation. If the situation makes the rules and other parts of a situation moral and not the other way around, the task of the philosopher is not to look for the one moral "quality" or whatever, but rather to characterize those combinations of circumstances or aspects which make a situation a moral one. And this is just what this studyiis about. Just as the lion is composed of a series of parts no one of which is labelled "lion," so moral situations may best be thought of as having such "parts" as agents, actions, rules, patients, conditions, and value judgments--no one of which can be labelled intrinsically moral (Fotion, 1968, pp. 47-48)." In a cross-sectional developmental study, the passage of individuals through the stages of development over time is represented by the responses of different individuals at different stages at one time. The resulting responses are thought to represent the pattern of 10. 11. 12. 102 movement that the construct being measured shows if the same indi- viduals were measured at these various intervals. Thus, when "movement" of values in the hierarchy across ages or stages is alluded to, it should be recalled that this is an inference about a longitudinal pattern from the cross-sectional data. With this in mind, value "peaks" refer to the developmental group in which a value reaches its highest average rank when com- pared to other groups. Thus, when relative to gs at other stages of development, gs at a given stage will be predicted to rank a y’ value highest on the average, the value is said to "peak" at this given stage. Each male student in the three grades was given a copy of the letter with parental request form to take home and was asked to return it the following day whether or not approval was given. One week later a second copy of the form'with an additional cover letter was mailed to the parents of each child who had not re- turned the form asking them to sign it and have their son return it. A few gs were absent on one or the other test days. They were ad- ministered the value Survey on the first day they returned to school. The longest test-retest interval for any §_was three weeks, four days. Moral maturity (MM) scores are the sum of weighted situation score weights. As an example, the MM score for subject #30 - 26. Situation: III Iv I VII Weighted Global Score: 2(4) 1(3) 2 2(3) ; Z 1 2 2 2 2 2 3 2 7 14 8 3 1 1 2 6 3 4 1 1 4 was . _ 6 _ .. ifo26-HH score Socio-economic level was determined from.§s' reports of parental occupations and educational achievement. High socio-economic level (roughly "middle-class" and up) was inferred from such occupations as: professional or semi-professional; teacher; engineer; shop foreman; small business, sales, or skilled trades with some college education. Low socio-economic level (roughly "lower middle-class" and below) was inferred from such occupations as: small business,ssles, or skilled trades with no college educa- tion; electrician; semi-skilled factor work; unskilled labor. 13. 14. 15. 16. 17. 18. 19. 103 Rokeach reports the following test-retest median stability co- efficients for samples of college students: Hill7; at 3 weeks: terminal - .78; instrumental - .72 N-100; at 7 weeks: terminal - .78; instrumental - .71. These value peak predictions may be classed as a binomial experi- ment as there are p_independent trials (predictions) with only two possible outcomes, ‘correct' and 'incorrect' and we are interested in the number 'correct' in the p_trials. However, p(correct) is not identical for each of the p'trials because in some cases the £22, highest medians have been predicted and in the others, only the single highest median. Where (in Table 18) only one peak has been predicted, the chance p(correct)-l/5-.2. Where two peaks are pre- dicted, p(both correct)-(2/5) (1/4)-.l. Thus, if for simplicity, we count each two-peak prediction as correct only if both peaks are correct, and consider this equivalent to a one-peak prediction with p(correct)-.2, we would actually overestimate the chance probability, making our statistical decision more conservative. The chance probability of guessing correctly both peaks for either value is .10. The probability of getting all four correct by chance is .01. Even if the two stages at which one value peaks can be eliminated from consideration in predicting the peaks of the other value, the chances of guessing all four peaks correctly is only one out of thirty. A non-parametric two-way ANOVA technique such as that outlined by Wilson (1956) would be useful for analyzing stage by age interaction. However, the cell n's in such an analysis would be of such a small order that any interpretation of the results would be dubious. The correlations (r) between rank of freedom and rank of obedient are, for time 1: -.l749; for time 2: -.0252. The hypothesis p20 may be tested with the £_ratio: Vn-Z c- 1' with N-2 df (Maya, 1963). 31'1“! For N578, ¢* -.05 (l-tailed), the critical value of £f1.67. The value of p,associated with r- -.l749 is 1.55. Thus, there is in- sufficient evidence to reject the null hypothesis. The Rank-difference Score is derived by subtracting the §fs rank for freedom on the terminal scale from his rank for obedient on the instrumental scale. Thus, Rank-difference Scores may range from +17 (freedom ranked 1 and obedient ranked 18) to -17 (freedom ranked 18 and obedient ranked 1). An eyeball comparison of the value medians of my §s with the value means of 564 New York City Public School children at the same grade 104 levels (i.e., 7th, 9th, and 11th grades) reveals some interesting pattern similarities and differences. For example, salvation, in the Michigan sample (boys only) stays above rank 10 for 7th and 9th grades before dropping 17th, where it stayed consistently in the New York sample. The most striking comparison, however, seema to be general similarity in the developmental patterns of the New York City and Michigan samples. (The unpublished New York value data is the work of R. P. Beech and Aileen Schoeppe of New York University). REFERENCES Bakan, D. The test of significance in psychological research. Psychological Bulletin, 1966, pg, 423-437. Baldwin, A. L. Ipeoriep of child deyelopmpnt. New York: Wiley, 1967. Barton, A. Measuring the values of individuals. Religious Egpgation, 1962, 51, Research Supplement, 62-97. Clausen, J. A. (Ed.). SOQialization and society. Boston: Little, Brown, 1968. Dukes, W. F. Psychological studies of values. Psyphological Bulletin, 1955,122, 24-50. . Dunn, L. M. Expanded mppppl f0; she Eeappgy Pipgpxg ypgapplgpy Igpt. Minneapolis: American Guidance Service, 1965. Flavell, J. H. The developgental psychology of Jean Piaget. Princeton, N. J.: D. VanNostrand, 1963. Potion, N. Moral Situations. Yellow Springs, Ohio: Antioch Press, 1968. Frankena, W. K. Ethics. Englewood Cliffs, N. J.: Prentice-Hall, 1963. Furth, H. G. Piaget and knowledge. Englewood Cliffs, N. J.: Prentice-Hall, 1969. Haan, N., Smith, M. B., Block, J. Moral reasoning of young adults: political- social behavior, family background, and personality correlates. Jppgpgl of Personality and Social Psychology, 1968, 19, 183-201. Hartshorne, H., and May, M. A. Studies in the natpye of phapappepz 22]., I: Studies in Deceit; vol. II: Studies in Service apd Self-gonprpl; Vol. III: Studies in the Organization of Character. New York: Macmillan, 1928-1930. Havighurst, R. J., and Tabs, H. Adolescent character and perponality. New York: Wiley, 1949. Hays, W. L. Statistics for psychologists. New York: Holt, Rinehart a Winston, 1963. Hollen, C. C. The stability of values and value systems. Unpublished master's thesis, Michigan State University, 1967. 105 106 Homant, R. J. The meaning and ranking of values. Unpublished master's thesis, Michigan State University, 1967. Homant, R. J. Semantic differential ratings and the rank-ordering of values. Edpgational and Psychological Measurement, 1969, 22, 885-889. Inhelder, B., and Piaget, J. The ggowth of logical thinking from child- hood to adolepcence. New York: Basic Books, 1958. .Inkeles, A. Society, social structure, and child socialization. In J. A. Clausen (Ed.), Socialization and society. Boston: Little, Brown, 1968. Pp. 75-129. Kessen, W. Research design in the study of developmental problems. In P. H. Mussen (Ed.), Handbook of research methods in child develop- ment. New York: Wiley, 1960, 36-70. Kluckhohn, C. Values and value-orientations in the theory of action. An exploration in definition and classification. In T. Parsons and E. Shilo (Eds.), Toward a general theory of action. Cambridge: Harvard U. Press, 1959. Kohlberg, L. The development of modes of moral thinking and choice in the years 10 to 16. Unpublished doctoral dissertation, University of Chicago, 1958. Kohlberg, L. The development of children's orientations toward a moral order. Paper presented at Human Development Symposium, University of Chicago, April 1960. . Kohlberg, L. Moral development and identification. In H. Stevenson (Ed.), Child Psychology. 92nd Yearbook of the National Society for the Study of Education. Part 1. Chicago: University of Chicago Press, 1963 (a). Kohlberg, L. The development of children's orientations toward a moral order: I. Sequence in the development of moral thought. Vita Humans, 1963, 9, 11-33 (b). ' b,’Kohlberg, L. Development of moral character and moral ideology. In H. Hoffman & L. Hoffman (Eds.), Review of phild deve10ppgnt research. Vol. 1. New York: Russell Sage Foundation, 1964. Kohlberg, L. The child as a moral philosopher. Ps cholo Toda , 1968, ‘2-4, 24-30. Kohlberg, L. Moral judgment interview coding manual. Mimeo., Laboratory zk of human development, Harvard University, Cambridge, 1969. 107 Kramer, R. B. Changes in moral judgment response pattern during late adolescence and young adulthood: Retrogression in a development sequence. Unpublished doctoral dissertation, University of Chicago, 1968. Levitin, T. values. In J. P. Robinson and P. R. Shaver (Eds.), Measures of social psychological attitudes. Ann Arbor: Institute for Social Research, 1969. Pp. 405-502. Lovejoy, A. 0. Terminal and adjectival values. Journal of Philosophy, 1950, 41, 593-608. V/*Maccoby, E. E. The development of moral values and behavior in child- hood. In J. A. Clausen (Ed.), Socialization and society. Boston: Little, Brown, 1968. Pp. 227-269. MacCurdy, J. T. Psychopathology and social psychology. Part III: Hierarchies of interest. British JOurnal of Psychology, 1950, 41, 1-13. Murray, H. A. Explorations in personality. New York: Oxford Univ. Press, 1938. Mussen, P. H. (Ed.) flpndbook of researph methods in phild developppn . New York: Wiley, 1960. :ngenner, L., Homant, R., and Rokeach, M. Comparison of rank-order and paired-comparison methods for measuring value systems. Perceptual and Mpto; Skills, 1968, 21, 417-418. Piaget, J. The moral d at o t e hild. New York: The Free Press, 1965. Translation of book originally published in 1932. . Pittel, S. M. and Mendelsohn, G. A. Measurement of moral values: a review and critique. Ppychologica; Bulletin, 1966, pg, 22-35. Rokeach, M. Beliefs, attitudes, and values. San Francisco: Jossey-Bass, 1968. Rokeach, M. I. Value systems and religion. 11. Religious values and social compassion. Review of Religious Research, 1969, 1;, 2-38. Rokeach, Mg, and Parker, S. values as social indicators of poverty and race relations in America. The Annals of the Amegigsn Agadepy of Political and So ial Scien e, 1970, 388, 97-111. ;%< Rokeach, M; The measurement of values and value systems. In G. Abcarian (Ed.), Political participation and social psychological processes. Columbus, Ohio: Charles E. Merrill. In Press. Siegel, S. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill, 1956. 108 Turiel, E. A ex e W ptages ip the phild'p gpza; jpdgmgnpp. (Doctoral dissertation, Yale University) Ann Arbor, Mich.: University Microfilms, 1966. No. 66-1, 115. Turiel, E. An experimental test of the sequentiality of developmental stages in the child's moral judgments. Journal of Personality and Social ngghology, 1966, 3, 611-618. Winer, B. J. Statiptical prinpiples in expepimental design. New York: MbGraw-Hill, 1962. Yarrow, M. R. The measurement of children's attitudes and values. In P. H. Mossen (Ed.), Handbook of repearch methods in child develop- ment. New York: Wiley, 1960. Pp. 645-689. APPENDIX A Definition of Kohlberg's Moral Stages 1. Preconventional Level Atthis level the child is responsive to cultural roles and labels of good and bad, right or wrong, but interprets these labels in terms of either the physical or the hedonistic consequences of action (punishment, reward, exchange of favors) or in terms of the physical power of those who enunciate the rules and labels. The level is divided into the following two stages: Stage 1: The punishment and obedience orientation. Orientation toward punishment and unquestioning deference to superior power. The physical consequences of action regardless of their human meaning or value determine its goodness or badness. Stage 2: The instrumental relatiyipt orientation. Right action consists of that which instrumentally satisfies one's needs and occasionally the needs of others. Human relations are viewed in terms like those of the marketplace. Elements of fairness, of reciprocity and equal sharing are present, but they are always interpreted in a physical, pragmatic way. II. Conventional Level At this level, maintaining the expectations of the individual's family, group, or nation is perceived as valuable in its own right, re- gardless of immediate and obvious consequences. The attitude is not only one of conformity to personal expectations and social order, but of loyalty to it, of actively maintaining, supporting, and justifying the order and of identifying with the persons or group involved in it. At this level, there are the following two stages: Stage 3: The interpersonal concordance or "good boy - nice girl" orientation. Good behavior is that which pleases or helps other and is approved by them. There is much conformity to stereotypical images of what is majority or "natural" behavior. Behavior is frequently judged by intention--"he means well" becomes important for the first time. One seeks approval by being "nice." 109 110 APPENDIX A - continued Stage 4: The "lay and ordep" opientatiop. Orientation toward authority, fixed rules, and the maintenance of the social order. Right behavior consists of doing one's duty, showing respect for authority and maintaining the given social order for its own sake. III. Postconventional, Autonomous, or Principled Level At this level, there is a clear effort to define moral values and principles which have validity and application apart from the authority of the groups or persons holding these principles and apart from the individual's own identification with these groups. This level has two stages: Stage 5: The social-contract legalistic orientation. Generally with utilitarian overtones. Right action tends to be defined in terms of general individual rights and in terms of standards which have been critically examined and agreed upon by the whole society. There is a clear awareness of the relativism of personal values and opinions and a corresponding emphasis upon procedural rules for reaching consensus. Aside from what is constitutionally and democratically agreed upon, the right is a matter of personal "values" and "opinion." The result is an emphasis upon the "legal point of view," but with an emphasis upon the possibility of changing law in terms of rational considerations of social utility (rather than freezing it in terms of Stage 4 "law and order"). Outside the legal realm, free agreement, and contract is the binding element of obligation. This is the "official" morality of the American government and Constitution. Stage 6: The universal ethical principle orientation. Right is defined by the decision of conscience in accord with self-chosen ephical principles appealing to logical comprehensiveness, universality, and consistency. These principles are abstract and ethical, (The Golden Rule, the categorical imperative) they are not concrete moral rules like the Ten Commandments. At heart, these are universal principles of justipe, of the reciprocity and eguality of the human ri hts, and of respect for the dignity of human beings as individual persons. Note: from Kohlberg. 1969. APPENDIX B ASPECT LIST Grouping I. The Categories A. Prima-Facie Obligations Aspect 1: Extga-Legal or Mopal Nogpp. Ways of invoking and conceiving Aspect In: of rules, norms, and role-stereotypes. Metaethical issues about relativity of moral norms and rules. Aspect 2: Legal Norms. Ways of invoking and defining legal norms. Aspect 2M: (Either explicitly or if norms about stealing, etc., are talked about in ways implying crime, police, etc.) Metaethical issues about relation of morality to law. Issues of legitimacy of civil disobedience. (Aspect 2211: General reasons for making and keeping laws is listed under Grouping III as a value.) B. Conceptions of Prima-Facie Rights (These are all defined under Grouping III where each value aspect has a rights subaspect and a value subaspect.) C. Conceptions of Dutiful Choice Aspect 3: Concept of "should" op "opght" for an actor in a choice Aspect 3R: Aspect 3M: Aspect 4 situation involving a conflict between rules or between rules and the interest of self or of others. Obligation when have right not to fulfill obligations- the relation between rights and obligations. Metaethical issues about relativity of obligations. D. Taking Responsibility Limiting gonseguences and pepsons actor is responsible for. Aspect 5: Limiting autonomous choice by reliance on advice or compromise with others. Aspect 6: Acpountability. Limiting accountability for(not) performing an act because of ignorance, lack of self control, etc. E. Praising and Blaming-the Worth of Persons and Personal Actions 111 112 APPENDIX B - continued Aspect 7: Culpability or Blame. Judgments of whether to blame someone as a person when he has violated a norm or obligation. (Aspect 8: Praise apgfiAdggration.) F. Meting Out punishment and Reward Aspect 9: Rules for Punishing. When, how, how much to punish. (Aspect 231 : the general purpose of punishment, its basic functions as expiative and preventive--is listed under Grouping III as a value.) Aspect 10: Rules for Rewarding. When, how to reward. (Aspect 2411: purpose or function of reward--is listed under Grouping III as a value.) Grouping II. The Principles G. Considerations of Prudenpe Aspect 11: Fear of Punishment and anticipation of guilt (or shame) as reasons for following norms. (Aspect 12:Anticipation of Reward or of pride or self-esteem as reasons for following norms.) Aspect l3: Anticipation of pain to the self, of injury or failure as reasons for following norms. (Differs from Aspect 11 in that these bad consequences are not punishment--they may be interpersonal, however. The harm to the self coming from disruption of de- sired relations is Aspect 13, altruistic relations are Aspect 15.) (Aspect l4: Antipipation of pleasure to self (outside a defined reward system) as a reason for following norms.) H. Consideration of Welfare of Others (Note where the welfare is a matter of definite values of Aspect 2211, Maintenance of Law; of Aspect 26, Life; of Aspect 27, Property; of Aspect 28, Liberty; of Aspect 29, Love and Fraternity--it is scored under Grouping III values.) Aspect 15: Welfare of other individuals (Love and friendship as altruistic motives or reasons for helping others or conforming come here, the reasons for entering into and maintaining love or friendship relations are scored Aspect 29, Love.) Aspect 16: Welfare of group, institutions, and societies, as a reason. Aspect l7: Aspect 18: Aspect l9: Aspect Aspect 20: Aspect 21: Grouping III. Aspect 22: Aspect Aspect Aspect 23: Aspect Aspect (Aspect 24: Aspect Aspect Aspect 25: Aspect Aspect Aspect Aspect Aspect 26: Aspect Aspect Aspect 27: Aspect Aspect Aspect 28: Aspect 113 APPENDIX B - continued I. Considerations of Respect Respect for persons and personal authority as a reason. Respectyfor the group, for group consensus, and for social order as a reason. J. Considerations of Justice Maintaining positive reciprocity and trust. 19RX: Defining or justifying obligations by stating actor should exchange places with the victim- Golden Rule. Maintaining negative reciprocity by vengeance or by refusal to honor non-reciprocal demands. Distributive Equality -- Maintaining equality or equity (equality relative to need) as a reason. The Basic values and Rights Security oELaw and Legal Order as a value. 221: (not used as comes under Aspect 2) 2211: Reasons why laws and their enforcement are necessary or desirable. ' Punishment as a value 231: (not used as comes under Aspect 9) 2311: Reasons, purposes of punishment, its basic functions as expiative and preventive. Reward as a value.) 241: (not used as comes under Aspect 10) 2411: Reasons, purposes of reward. Contract, Promise and Non-Deception as values. 251: Definition and Use of Contract and Promise- Xeeping Concepts. 25 I: Reasons for Maintaining Contract and Promise 25 : The Social Contract--the contract of the individual with abstract institutions or with society. 25T: Truth values. Life as a value. 261: Definition of the nature of Life's value, of the Right to Life, e.g. of what lives are valuable under what conditions. 2611: The reasons why life is morally valuable. Property as a Value. 271: Definition of Property Rights and values. 2711: Reasons for maintaining property rights. Liberty or Autonpgy as a Value. Definitions of rights and values of freedom from coercion. 281: 114 APPENDIX B - concluded Aspect 2811: Reasons for valuing freedom, for having rights of liberty. Aspect 29: Love and Fraternity as Values. Aspect 291: Definition of the obligations and nature of a good relationship or of a good love motive. Aspect 2911: Reasons for love and friendship being valued. Aspect 30: Sexual Values Aspect 301: Definition of appropriate sexual relations. Aspect 3011: Reasons for valuing appropriate sexual relations. Note: from Kohlberg, 1969. meowumumowmnoo omumwamwoa nomwumuaaau: mo mmoaommm non .haoosm menus as: 053 one a season so use moaonw nuaoo now umouxo .m owmum mom on omwsuonuo umoe n owmum mm on“: uuosuo .oamao no omammm mamas moamuomaum Homes moans ou mouuommq no moauom :36 mo mooooovom moqmwooo mo moamwomaum Hmuoa unmumomusmumsm ou zoom on no woman you auqammoa umoo Ham mom oaoammommou Hmmuo>uoo ho vomwauouoo .o moo~m> mumoauuqaaoo mo asoummoan Houseman monoumummoo gamma mo uxou muowooo mo mosses a mu amouomuuooo wouaamwasu m magmas oowono Homegumu nmoo ma unoasmaosm mo 3ou> one moumuooAm Hmwuumm one muozuo mo «unwau an no madam floouomuu amowwo~osozmmummaumuwHaas use as common oamHo wmauoommom ou mouuaaa umouuamwoa an momaauouoo .m Hm>oummmuao weamamumamansma A~m>ommmmv Homomuom moauafiaoummommou moowwm one Amman m.o:o you man uo>o auumoaum moxmu 1mm wmafiaaufiow one moaom moons mom usmwu momma umss moo .o.av o>uumumxo mmoemnoua Hmoauowoumo onu wmusoaaom ou mouwaafi umouoouoasm .Hmoamowoumo .o no>wuoa smegma mo unsooom moxmuuuawuoa> mo>auos wmwamomooo mo ammo oxmu one mouuuuo om on» no soauouauoom m was aoHoOoomm ooow wnw>ms unuom mason umo~o>omoo mo mamas mamoom moow .oamooa ~m>ommanmam mo ooammoumxo uom ou moumoaHOuosmHo oomoomaxo onu an vouaawa umoa owns .HmoammuOuououm .m aauou> onu ou mongoose on» Assam...wav amouo o>au=uaumom no moduuuomom .m> smaaoom no oumoum nosuoamn .smua ou Houses mo o>wumo>oum .umoumoa osu cu moumoau0108mao maxm onu mo caveman nsmummfi .owumw>«umamm .N you may no some ua>oo on» no :mmoeooow: o>aumufimuou one owmamo .moooosoom mouom no one: o>aouooo .omxwm .momoaumooomo unoo snowmanm mo momma nmmoauosom so mouuawfi .amomouxo .owumausaomom .~ Amm .Oe .m auuoaso amouu ad hoaamsom .uuouu we huuamGOuuom muss: mouuouauo sumaauas sou ma mummaos .maamao momma «no mo mo oumwaos one mm as same on» ma Mano umoawmsn Hmuoa ouna umoumo Amman oommoousqma ago no auuooamuoou mom uoommou anamo>ams o>m£ uom moov mowumon muons uHoa m.m~om onuv ooeoosum .9 common mo moooonvonmoo .uoommou Humans sou mo mwuouuuo oumaau .uonwau ~a>uousmasa gnomes mom muoomoo .huaammommon nfis on ma huouooo mo sumo uoommou umomou mo one auumsouommo we was Boom consumes» one name no moose: uaoumouw «.muuosaaoo madmamuou ou muaamsoo ma anemones .uuauu mosamwooom auauoousm mo onu wo moow umoomouw mom summaos_oso mo wmmsuau umoo he momuuom zuwooumuoou nmoauoosw mom magma» on» .oummaos moaeoasoo uxma ammowusm ou moumowuo .m .u.moo uuosuo .moHou onu one duos no snowmoo ou ouuaooahm mom amoumxo o>nn osoa muons omonu one has» awesome mouoomoou enemas: sumaauas nouuaamsoonu odomnoauoon mouuauoeusm .huuannos one can macaw Humoausuwum mmosuo madman: one smash uoo .muoov mo unnamed on» no open ummumomaa no and mo summed onu mo Hs>a> sou wouumdoa> uo>o odds» mo .xuos one mouse mom monsoon husmaooom m auauonuso name was mono: onu .uovuo mwmqaoow ovao>s was mono: mo choose one we muwooumaoou mom woodman Hammouma amwoom mo oommmoumwms any m.omo moamumama usage sou .a .mooum noon madame numooo mu muwfimovm .mmoauma no» Hmeomumo>moo onaammoaumu moaumaou amsoau anonuo mo ~m>oummm ou mousnmuo>mm mean we can» noomum so goes voomm woa>o~ one enqufioums wagon one amouumaou ooow use Homoumuw an muuooumwoou zuumoeusm you uoomnou .nuomuo wounds: am woo» woo mammamuoama mu ooow .m omo~w>mum om moguosuom uouam nos-m ago new masons omomuo>o ou suuoa aqueous . nmoo me uom moo: .mmosuo nmuafiasuo ousfioaom no o>au ousoauuum woo «woo mom mama may nonuo may no moss summonoo gamma» nooauuauv mom: .ommmsoxo uou .huwuonunm mom uooma moon oeu ou Hmumoasmuomw named mo>uon some: use» now ago an auwooumaoou moms non Hmnowomo>ooo among we use: omam ma woo» woo ma uewwu no ooow one .N noummavuoosm mom moonwau-uosom nonmoaoomoou museums amouumfiou fiasco meson homeomon nommm~03,mmass mo mauou mownoo mom unmanmamom no: .amooumuoouumoc uumooom ou sommwoooo aw uoomuou mu moon on» summon u.maooo mo meow cu modumumowuo .a STE 36633 3:33 a: .2 3833 GA .2 36633 ST: 3833 o>auoowmumao one huaooum uuoom .o>«unusEEOOuooauasn moomuom muosoo mo saunas: oomoooum mw~omaoeu mom can muswmu humonomm yon .umouu amwoou mom aummmoooo «a ammo new uoomaou one goes Iflufimdho HQHOE G M0 UHOO Ofiu a.“ moomo gamma mo monsooumama omomoo ow some: mmoqumaom amuoa mo maamwoom umo .muoaoom mu amouumaom Hausa mo mouumomsom on» ma unnuu Humane mo and mo oouou onu o>mn momma» humanoum owed muss: ago you uoomuou genus oommmoumama osu .uomuumoo omohom .o .maounam owumadmuoOu .oauqumuwmmo mu mnemuwmumoo wmuzmo> mo mauou ow anaemia ups one accuuwemuoo ooaasuoo .uo>oso= o=~o> Ammoaumm mo moouo Hmwoon no one name .uoomueoo finesse on» we once .uusmuu uoomoo canon omm uowwu aqueous nuaasoo .moqumwaaoo wo mouumoasou mass: awesome so moo munwum homomoum .dmuuo>usa oaumo sou ma owafi onu ma umoaoomwm mo moowumoo comm .m .uoouo .omua mo meomumaou fimwoou ma muaauomaaou finance ago no meoaumoesou sou one u« now moduoououm ago we moouo msoamqaom you owe: oeu onm u.ooo ooh ma mono: ommsou onu omm ago: one: umeu uomm any nomaoom on» «0 once on» mammooo sumo no sense a coma once use» wow so oommo .oo:~0uom ohm munwwu homomomm ummumomau adoumswuao sou ma owed gamma ou mOaumumowuo Hmuuuowoomo m .e .oo oases a“ oumsou ooumsuu up Had: no» on momsouaumomoum ooom m one: on .mauoo muonuo mo heummao omm moauooumo omm muosuo mmwuowommmmwo ouo>m ou oaou ou ooummuouooom muawwu muuomoun onu so oommo osam> m.oma~ omen: omox mum momuaomm ooo mumoaoomwm .m Amo>gfi ooov monomomm momma sow: mom: zone one: so one omo unmmommom no“ mo momma owned .ommmnoxo nauo>onnuwmanu mo omo mass «0 munuuoeso umooon ago on muaflmueoasuumma onm mommamwumo mo assumes a ma no oommo magma» homomoum :ouofiomom: mud ou ooozoou osHm> n.0maa moon mumoeoouwm omm momwaoum .N muomso mo mo>mHm .nusoumm mo homomoum mnu sou mommoumom mud mo owqumoum noaano .w.o .muosuo mo auuoaoum onu on Scum no nuoofioo amowmhnm wmuomao uom mmo ofimoom umnu common uumooo< .oomsm mo osam> scum oosmuowowumuo mum omm uuo~ as» ou usmuosom nmco gunman mason one magma» monomomm hammofio mos o=Hs> m.ou«~ amass once any up some can momuaomm .H Asa sooasu:m oa~m> one on ouuau movaoneou avauw some cu aw 0N ecu we soon a «mm. ow.~ n a.a a N.m o ~.m saunas one. am.c a o.» a n.k m a.“ aaeaeeuaum 6:49 «cc. an.oa we m.na ma m.ea «a «.ma coauaawoooe auaoom «so. a~.n ca n.oa ea a.~a ea a.~a uuoauou-uaom sac. an.m he m.na ea n.oa a n.oa coasusaum ANN. ae.~ ea m.ma he e.na ma «.ma «newsman awn. co.“ m o.m no m.~H a ~.oa kuauaoom aueoauaz nmo. NA.G a o.a ca a.oa ea n.~a 056; «Essa: Aka. mn.n no c.aa Na n.aa Na o.- access: coca“ mam. e~.a A m.» e o.» a A.» sausages: Ame. no.3 N m.n N «.3 a A.m anemone coo. ma.aa e o.a n m.h m o.e augusuom aaaaum ans. no.a n «.6 n n.» e m.n kuaaaswu as». m~.c «a k.aa a ~.ca ca o.aa assume no case: < one. em.o a n.m a ~.~ ~ n.m «666m u. eases 4 sec. ma.o «a o.aa m n.» ea a.aa necessaaasouo< mo «scum < can. ee.~ as m.oa ma a.aa “a n.~a omaa weauaoxm ca has. kc.o ma w.ua ea m.aa no o.~a «wag «asuuuomaoo < a a sane .5: same .52 as; .5: asoz< .aaaaz-aux.auu an um a as «case a sense A sense .Avocaoaoo sumo ho>unm oauo> Nissan a Auoawuv doauq scene zoom on mm you euomuo scam ouueoaaoo woo escape! oaaa> Hmeaauoa .m.m manna 133 APPENDIX E (con't) .ho>uam oaan> ecu Ou ouwau oovconoou oosuw sumo aw aw.o~ may no comm G ooo. Nn.o n N.N o N.o oo o.oo Noo. oo.o N N.o N n.o o N.o ooo. NN.o no o.NN o n.o o N.N ooo. so.“ NN o.oo No N.NN No N.NN NNo. No.N NN m.oN N N.N o n.o ooo. oo.mo o n.o oN o.mN NN N.oN moo. oN.n oN o.oN «N n.No oo N.NN moo. .oNuoo a N.N oN o.oN no n.NN ooo. no.N oN N.¢N NN o.NN oN «.mN ooo. oo.o N m.m o o.« a o.N «No. oN.n NN o.oN oN N.o m N.o mNo. No.N NN N.oN o m.o N N.N one. oo.m N o.o a N.N m N.o ooo. No.oN oo o.NN no o.oN N N.o oeN. so.N 3N o.NN N o.o No N.NN moo. om.o o N.o oa o.No no N.No «No. No.N o N.o nN n.oN oN o.NN moo. NN.n m n.o n N.N o m.o a o some .soo some .ooz some .ooz ooozo oooNo:-ooxoouN No No No oo ooouo o ooouo N coupe voaaouucoosuaom afloaocomaom ouaaom ucowoooo mca>OA deadwoa HoauuoaaoucH uaooconovcH o>wuacqwaaH amuse: damage: mcw>uquh oaoowouaoo cacao Hamuoono oNooooo voocaavaowm eaoquwoa< .Avocuaaoo mono hoauam oaao> «Imago a Anoaauv Ho>og omega seam ua aw.uou uuovuo scam ouuuomaoo one seafloor oa~o> aaucuaauuoan .m.a amass 134 APPENDIX B (con't) .ho>uam oa~a> sou ca 00439 ooocomeou_m.nouo aoaaoon wouachN one e.c och o 135 noN. mo.o o N.N o o.o n N.o o o.N oN o.oN soooNz ooo. oN.N N o.N m N.o o N.N n o.N o o.N oNoooooNum cone ooo. Nm.o oN o.oN oN o.NN oN o.NN oN o.oN NN N.NN ooNoNoooouN NoNuoo oNo. Noo.NN o o.o mN N.NN oN o.NN nN N.NN NN o.NN ooooooNINNoo oNo. oN.NN oN o.NN a n.o NN o.NN oN N.NN n N.N ooNoosNuo ooo. No.n NN N.NN NN o.oN oN o.oN «N N.NN nN o.oN «ooooon oNo. «N.NN m m.n oN o.oN NN o.NN oN o.oN N n.o NoNuooom NoooNooz ooo. oo.N NN o.NN NN o.o N m.o NN N.NN oN o.NN o>oN «noon: Noo. No.NN NN o.NN oN N.o oN o.oN NN o.NN NN o.NN Nooauuo noooN Non. N.N NN o.oN o N.N o o.o N o.N o o.o o.ooNoooo ooo. oN.NN N o.N m N.N N N.N N o.N N N.N aooooum Noo. No.o a N.n N N.N a o.o o o.o m n.o NoNuoooo NNNsoN oNo. oo.o n o.n N o.n n o.n n n.o o o.o NoNNaooo NNo. No.NN oN N.NN o o.o NN N.NN o o.o o n.o Nooooo «6 oNuoz a ooo. No.o N o.N N o.N N o.N N N.N N n.n ooooo on oNuoz < oNN. oo.N oN o.oN NN o.oN o o.oN o o.oN NN o.NN .oaoooo No ooooo a non. «N.N o N.N «N N.NN NN o.NN NN o.NN oN o.NN «NNN ooNoNuxo no ooo. mo.N nN o.oN mN o.NN nN o.NN mN o.NN oN o.NN ooNN oNooouoNaoo < a No None .5! None .5: None :5. None .8: so} .5: <>°z< OHflHflalfiQu—Oflg NH ON 0% 00 OH I “G n owoum a mason n owaum N amoum H omaum .Aoocaoaou sumo Ao>uam os~a> NIaaNu d Hiaawuv ucoamoNo>oa Nunez mo owaum doom uo am you auovuo seam ouwoomaoo vac neeaoux.oaam> Nacuauea ..:.m oNoaH APPENDIX E (concluded) .No>aam oaHo> oau Ou OOHsu Newcomoou N some eoaaooo wouaHch one e.a any 136 NHN. NHN.¢ N N.N N N.N N o.N N N.N N N.NH voHHouueOeImHON NNN. NNNNN N N.N N N.N N N.N N N.N N N.N oHoNoeomeeN NNo. HNN.N NH N.HH NH N.HH HH N.NH N N.N N o.NH ouNHom Noo. NNN.NH NH N.NH N N.N NH N.NH NH H.NH oH o.oH ucOHoooo NNN. NHN.¢ NH N.NH N o.N N N.N N H.N N N.N Nchoa NNo. NNN.N oH o.N NH N.NH NH N.NH NH N.NH NH N.NH HaONNOA NNH. NNN.N N o.N NH N.NH NH N.HH NH N.NH NH N.NH HoaOOOHHounH NNo. NNo.N m N.N NH N.NH NH N.NH NH n.oH N N.N unaccomovaH NNN. NNN.¢ NH N.NH NH N.NH NH N.NH NH N.NH NH N.NH o>HuscHNuaH NNN. NNH.¢ H N.N H N.N H N.N H H.N H N.N unseen NNN. NNN.N NH N.HH N N.NH N N.N N N.N N N.N HamnHuN NNo. Noo.oN NN o.oN mN o.NN o o.N oN N.N o n.o ooNsNouoo NHN. NNm.c N o.N N N.N n N.N a H.N N N.N oaooNouaou NNH. NNo.N NH N.NH NH N.NH N N.N N N.N NH N.NH eNOHu Nam. NHH.o NH N.NH oH N.NH NH N.HH HH N.NH NH N.NH Hamuoooo Noe. NNo.¢ N o.N HH N.NH NH N.N NH N.HH NH N.NH oHooaou NNN. NNN.o N N.N N n.oH NH N.HH NH N.HH NH o.NH voocHavaouN NNo. NNc.HH N o.N N N.N N N.N N N.N HH N.NH caONuHoa< a N Noam do: scum :5: 3.52 do: Jena .5: Noam do: NNoz< oNHHosuHoxesuu NH Nm NM NM NH la: n ONauN o ONuuN N ONauN N ONauN H ONauN .AoocNoaoo sumo No>u=N oaHa> Nausea N HIaaHuv ucoaN—OHONNON Hone! mo ONauN zoom um ow new 88qu scam ouHaoaaoo New 0:3?! 03an HeucoflauuocH N.N OHoaH MAY '6 z. 1553 W "if 145 H I” N” U, Elf. 93 03 12 wrung/171711111;