Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Relational structures among worldview, self-view, moral inclusiveness, and moral orientation : a holistic… Yu, Angela Yan-Yan 2008

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2008_fall _yu_angela.pdf [ 6.2MB ]
Metadata
JSON: 24-1.0054481.json
JSON-LD: 24-1.0054481-ld.json
RDF/XML (Pretty): 24-1.0054481-rdf.xml
RDF/JSON: 24-1.0054481-rdf.json
Turtle: 24-1.0054481-turtle.txt
N-Triples: 24-1.0054481-rdf-ntriples.txt
Original Record: 24-1.0054481-source.json
Full Text
24-1.0054481-fulltext.txt
Citation
24-1.0054481.ris

Full Text

RELATIONAL STRUCTURES AMONG WORLDVIEW, SELF-VIEW, MORAL INCLUSIVENESS, AND MORAL ORIENTATION: A HOLISTIC AND COMPLEMENTARY PERSPECTIVE  by ANGELA YAN-YAN YU B.Sc., University of Toronto, 1980 M.A., University of British Columbia, 2001  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY  in THE FACULTY OF GRADUATE STUDIES  (Human Development, Learning and Culture)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) August 2008 ©Angela Yan-Yan Yu, 2008  ii  ABSTRACT The overall goal of this study was to develop a comprehensive model of moral development to explicate the complexities of everyday morality. Based on a holistic and complementary perspective, the model relates the constructs of ―worldview‖ and ―self-view‖ to represent the influence of cultural individualism-collectivism and personality on moral development. It posits that worldview shapes self-view and moral inclusiveness [what is included in one‘s moral consideration (Carter, 1980)] and then worldview, self-view, and moral inclusiveness jointly influence moral orientation (the main hypothesis). Interacting with situational factors, moral orientation would further influence moral judgment and behavior, thus connecting habitual morality with reflective morality. The specific objectives were: (a) to examine the relational structures among worldview, self-view, moral inclusiveness, and moral orientation; and (b) to construct a moral orientation index that measures various moral orientations, including egocentric, family, care, norm, justice, biocentric, and religious orientations, reflecting the corresponding worldviews, self-views, and levels of moral inclusiveness. Moral inclusiveness was operationalized as relationship closeness to different social groups at expanding levels of moral inclusiveness: (1) self only, (2) family, (3) peers, (4) society, (5) humanity, (6) nature, and (7) God. Survey data were gathered from 640 Grades 8-12 students and 472 adults. Structural equation models (SEM) were developed using the student data while making some comparative references to the adult data. Results of SEMs generally supported the main hypothesis. For example, individuals scoring higher on vertical individualism, social Darwinism, and independent self scored higher on egocentric orientation; individuals scoring higher on vertical collectivism and closeness to family scored higher on family orientation; and individuals scoring  iii  higher on horizontal collectivism and interdependent self scored higher on care orientation, at Levels 1, 2, and 3 respectively. Individuals scoring higher on collectivism, horizontal worldview, and moral self scored higher on justice orientation, but individuals scoring higher on closeness to nation scored lower on justice orientation at Level 5. The primacy of worldview, particularly collectivism, over self-view and moral inclusiveness in moral orientation development has implications for education. Exploring the roots of habitual morality, this study advances theory by integrating different schools of moral psychology with cultural psychology.  iv  TABLE OF CONTENTS Abstract …………………………………………………………………………………………ii Table of Contents ……………………………………………………………………………….iv List of Tables …………………………………………………………………..………………..ix List of Figures …………………………………….……………………………………………xii Glossary ……………..……………………………………………………..………………….xiii Acknowledgements ….……………………………………………………..………………….xv Dedications ………...…………………………….…………………………………………….xvi CHAPTER 1: Introduction ……………………………………………………………………1 Background to the Problem ….………………………………………………………1 The Research Problem ……………………………………………………………….2 A Holistic and Complementarity Perspective ….……………………………….4 Significance of the Study ……………………….……………………………….……7 Summary and Overview ………………..…………………………….……………….8 CHAPTER 2: Review of the Literature ...……………….……………….……………………9 Theory of Moral Development …….……………………………….……………….9 The Kohlbergian Approach …………………..……………….……………….9 The Moral Judgment Interview ……………………….……………….10 The Defining Issues Test …………..………………….……………….12 Gilligan‘s Theory of Two Moralities ……………………….………………..13 Defining Moral Development, Moral Orientation, and Moral Judgment …….16 Defining Moral Judgment and Moral Orientation ….………………….17 Defining Moral Development ………………………………………….18 A Comprehensive Model of Moral Development ….………………………………28 Worldview …….……………….…………………………………………..…32 Worldview as Individualism-Collectivism .…………….………………32 Cultural versus Personal Worldview …...……………….………………36 Worldview and Moral Inclusiveness ………..….…………….………………37 Moral Inclusiveness as Relationship Closeness ……………….………………39 Moral Inclusiveness and Moral Orientation ….……………….………………40  v  Hypothesis ….……………….……………………………………………………...46 Level 1 ….……………………………………………….…….………………49 Rationale …………….……………..….…………………..……………49 Level 2 ….……………………………………………….…….………………51 Rationale ………….....………………..…………………..………….…51 Level 3 ….……………………………………………….…….………………52 Rationale ………..…...………………...…………………..……………52 Level 4 ….……………………………………………….…….………………52 Rationale ……..………………………...……………….………………53 A Caveat ...……………….………………………………...……………53 Level 5 ….……………………………………………….…….………………54 Rationale …..………..……………….…………………..……………..54 Level 6 ….……………………………………………….…….………………55 Rationale .…………...……………….…………..………..……………56 Level 7 ….……………………………………………….…….………………57 Rationale .…………...……………….……..……………..……………57 CHAPTER 3: Methodology ...……..……….…….……………………………..……………58 Research Design ………………………………...….……………….………………58 Sample Selection ..………………………………...….……………….……………58 The Pilot Sample ……………………………..……………….………………59 The High School Student Sample …………….……………….………………59 The Adult Sample (Comparison Sample) …………………….………………61 Data Collection Procedures ………………………..……………….………………62 The High School Student Sample .…..……….……………….………………62 The Adult Sample ……………………………….……………….……………62 Measures ….……………….……………………….……………….………………63 Demographic Data ….…………………………..…………….………………64 Moral Inclusiveness ………………………….……………….………………64 Circles of Relationship Closeness ….……………….………………….64 Reliability and Construct Validity ….……………….…….……………66 Worldview ….……………………………………..………….………………68 Measurement of Horizontal and Vertical Individualism and Collectivism ……………………………………………………………. 68 Social Darwinism, Anthropocentrism, and Spiritual Worldview .….….69 Self-View ….……………….………………...……………….………………72  vi  The Self-Construal Scale ….……………..…………….………………72 Moral Self, Ecological Self, and Religious/Spiritual Self ………………74 Moral Orientation ………………………...….……………….………………76 The development of the Moral Orientation Index .…….………………76 Reliability and Construct Validity ….………………….………………77 Validity Scale ………………………………..……………….………………80 Issues in Model Testing …………………………………………….………………81 Model Identification ………………………….……………….………………81 Model Estimation …………………………….……………….………………82 Goodness-of-Fit Indices ………………….….……………….………………83 Absolute Fit Indices …………………….……………….………………84 Incremental Fit Indices ……………..….……………….………………86 Summary ………….………………………………..……………….………………86 CHAPTER 4: Results ………..……………………………….……………….………………88 Preliminary Analysis ……………..………………..……………….………………88 Sample Size and Missing Data ………..……..……………….………………88 Outliers and Normality ………..……………..……………….………………89 The Modeling Process …..……………..………………..……………….…………90 Step 1: Fitting the Measurement Model ……..……………….………………90 Step 2: Comparing the Measurement Model to Alternative Models ……….…91 Comparing Alternative Models ……………………….………………94 Step 3: Validating a Structural Model …………………………..……………95 Tests of Structural Models .………………………………………...………………96 Predicting Egocentric Orientation at Level 1 …………………………………96 Predicting Family Orientation at Level 2 ……………………………………100 Predicting Care Orientation at Level 3 ……………………...………………103 Predicting Norm-Maintaining Orientation at Level 4 ………………………107 Predicting Justice Orientation at Level 5 ……………………………………111 Predicting Biocentric Orientation at Level 6 …………………..……………115 Predicting Religious Orientation at Level 7 ……………..………………..…119 Testing Individual Path Hypotheses ……..………..……………….………………122 Summary ………………….……..………………..……………….………………124 CHAPTER 5: Discussion and Conclusions …………….…..……………….………………125 Limitations of the Study  .……………..………………..……………….…………125  vii  Implications of the Results …..…………..………………..……………….………127 The Multiplicity of Moral Orientation ….…………..……….………………127 The Distinctiveness of Egocentric, Family, and Care Orientations ……128 The Distinctiveness of Norms and Religious Orientations ....…………131 The Distinctiveness of Justice and Biocentric Orientations ….…..……134 From ―Is‖ to ―Ought‖ ……………………………..………….………………138 Worldview and Self-View ……………………….………….………………142 Implications for Education ………………………………….………………144 Future Research Directions ………………..………………..………….……….…147 Future Research with the Moral Orientation Index ………….………………150 Conclusions …………………..…..………………..……………….………………153 References  …………………………………..………………..……………….…..…………156  Appendices  ….……………………………..………………..……………….………………173  Appendix A  The Measurement of Horizontal and Vertical Individualism and Collectivism ………………………………………………………………… 173  Appendix B  The Circles of Relationship Closeness ……………………………………… 174  Appendix C  The Moral Orientation Index ………………………………………………... 176  Appendix D  Student Recruitment Letter …………………………………………………. 178  Appendix E  Parent Consent Form ………………………………………………………... 180  Appendix F  Signature Page ………………………………………………………………. 182  Appendix G  Adult Introductory Letter ……………………………………………………. 184  Appendix H  Adult Cover Letter …………………………………………………………... 185  Appendix I  Demographic Data Sheet for Students ………………………………………. 186  Appendix J  Demographic Data Sheet for Adults ………………………………………… 187  Appendix K  The Self-Construal Scale ……………………………………………………. 188  Appendix L  Univariate Summary Statistics for Observed Variables for the Student Sample ………………………………………………………………………. 189  viii  Appendix M  Description of Marker Variables ……………………………………………. 192  Appendix N  UBC Research Ethics Board's Certificates of Approval for the Pilot Study… 195  Appendix O  UBC Research Ethics Board's Certificates of Approval for the Student Study ………………………………………………………………………… 197  Appendix P  UBC Research Ethics Board's Certificate of Approval for the Adult Study… 199  ix  LIST OF TABLES Table 1  Definitions of and Items for the Moral Orientation Index (MOI) Subscales …. 41  Table 2  Different Aspects of Worldview, Self-View, Moral Inclusiveness, and Moral Orientation at Different Levels of Inclusiveness ...……………………. 48  Table 3  Age Means in Years for Gender and Grade for the Student Sample …………. 60  Table 4  Age Means in Years for Gender and Ethnic Groups for the Student Sample … 60  Table 5  Age Means in Years for Gender and Religious Groups for the Student Sample ………………………………………………………………………… 61  Table 6  Item Summary and Rotated Factor Loadings for Relationship Closeness for the Student Sample …………………………………………………………… 67  Table 7  Factor Score Correlation Matrix and Internal Consistency Reliability Estimates for Relationship Closeness for the Student Sample ……………….. 68  Table 8  Summary of Items, Rotated Factor Loadings, and Internal Consistency Reliability Estimates for Worldview Subscales for the Student Sample ……... 71  Table 9  Summary of Items, Rotated Factor Loadings, and Internal Consistency Reliability Estimates for Self-View Subscales for the Student Sample ……… 75  Table 10  Rotated Factor Loadings and Internal Consistency Reliability Estimates for Moral Orientation for the Student Sample ……………………………………. 78  Table 11  Moral Orientation Index Factor Score Correlation Matrix for the Student Sample and the Adult Sample ………………………………………………... 80  Table 12  Indices of Model Fit and Comparison ………………………………………... 84  Table 13  Invariance Tests of Factor Structures for Egocentric Orientation Between Students and Adults …………………………………………………………... 97  Table 14  Correlation Matrix for Observed Variables at Level 1 for the Student Sample and the Adult Sample ………………………………………………… 99  Table 15  Goodness-of-Fit of Alternative Models Predicting Family Orientation at Level 2 ………………………………………………………………………. .101  Table 16  Invariance Tests of Factor Structures for Family Orientation Between Students and Adults ………….……………………………………………… .101  x  Table 17  Correlation Matrix for Observed Variables at Level 2 for the Student Sample and the Adult Sample ………………………………………………. 103  Table 18  Goodness-of-Fit of Alternative Models Predicting Care Orientation at Level 3 ………………………………………………………………………. 104  Table 19  Invariance Tests of Factor Structures for Care Orientation Between Students and Adults …………………………………………………………. 105  Table 20  Correlation Matrix for Observed Variables at Level 3 for the Student Sample and the Adult Sample ……………………………………………….. 107  Table 21  Goodness-of-Fit of Alternative Models Predicting Norm Orientation at Level 4 ………………………………………………………………………. 108  Table 22  Invariance Tests of Factor Structures for Norm Orientation Between Students and Adults …………………………………………………………. 109  Table 23  Correlation Matrix for Observed Variables at Level 4 for the Student Sample and the Adult Sample ………………..……………………………… 110  Table 24  Goodness-of-Fit of Alternative Models Predicting Justice Orientation at Level 5 ………………………………………………………………………. 112  Table 25  Invariance Tests of Factor Structures for Justice Orientation Between Students and Adults ……………………………………………….………… 113  Table 26  Correlation Matrix for Observed Variables at Level 5 for the Student Sample and the Adult Sample ………………..……………………………… 114  Table 27  Goodness-of-Fit of Alternative Models Predicting Biocentric Orientation at Level 6 ……………………………………………………………………. 116  Table 28  Invariance Tests of Factor Structures for Biocentric Orientation Between Students and Adults …………………………………………………………. 117  Table 29  Correlation Matrix for Observed Variables at Level 6 for the Student Sample and the Adult Sample ………………..……………………………… 118  Table 30  Goodness-of-Fit of Alternative Models Predicting Religious Orientation at Level 7 .....………………………………………………………………… 119  Table 31  Invariance Tests of Factor Structures for Religious Orientation Between Students and Adults …………………………………………………………. 120  xi  Table 32  Correlation Matrix for Observed Variables at Level 7 for the Student Sample and the Adult Sample …….…………………………………………. 121  xii  LIST OF FIGURES Figure 1  Seven levels of moral inclusiveness ……………………..…………………… 23  Figure 2  A proposed comprehensive model of moral development .…………………... 29  Figure 3  Hypothesized structural model ……………………………………………….. 46  Figure 4  Hypothesized and alternative models ………………………………………… 92  Figure 5  Theorized model for egocentric orientation at Level 1 ……………………….. 98  Figure 6  Hypothesized model for family orientation at Level 2 ……………………… .102  Figure 7  Modified model for care orientation at Level 3.……………………………... .106  Figure 8  Modified model for norm orientation at Level 4 ……………………………. .110  Figure 9  Hypothesized model for justice orientation at Level 5………………………. .113  Figure 10  Modified model for biocentric orientation at Level 6 ………………………. .118  Figure 11  Modified model for religious orientation at Level 7 ………………………... .121  xiii  GLOSSARY Anthropocentrism: the view that humans are the center of the universe (Campbell, 1983). Biocentric orientation: moral consideration is extended to include all life forms and the earth. Biocentrism: the view that the universe is the creator of life (Campbell, 1983). Care orientation: based on an assumption of connectedness, good behavior is defined in terms of getting social approval and maintaining good relationships Collectivism: the view that the basic unit of survival is the group and not the individual, thus stressing conformity, security, and closeness to ingroup (Hui & Triandis, 1986). Ecological self: the self is defined in terms of connection with other people, animals, and plants and as part of nature. Egocentric orientation: decisions and reciprocity are based on self-serving purposes. Ethnocentrism: the view that one‘s ethnic group is the center of life. Familism: the view that family is the center of life (i.e., life is passed on through family linkages). Family orientation: family responsibility is based on duty rather than good relationships; family integrity is highly valued. Horizontal collectivism: denoted by lateral relation to authority and collectivist relation to group (Triandis & Gelfand, 1998). Horizontal individualism: denoted by lateral relation to authority and individualist relation to group (Triandis & Gelfand, 1998). Horizontal worldview: denoted by lateral relation to authority, equality, benevolence, and universalism (Triandis & Gelfand, 1998). Individualism: the view that the basic unit of survival is the individual rather than the group, thus stressing self-direction, hedonism, and distance from ingroup (Triandis, 1989). Independent self: the self is defined as an entity separate and independent from groups (Triandis, 1989). Interdependent self: the self is defined in terms of relationships and social connections (Markus & Kitayama, 1991). Justice orientation: morality is guided by a justice moral principle that treats all human beings as equals and that goes beyond race and social boundary.  xiv  Moral development: the tendency (as a personality trait rather than an ability) to take an increasingly wider perspective or more standpoints from different groups in defining social facts and moral values. Moral inclusiveness: what is included in one‘s moral consideration (Carter, 1980). Moral judgment: the use of any moral orientation(s) for interpreting or resolving or making decisions for certain moral dilemmas, thus reflecting the interplay between (dispositional) moral orientation and contextual factors, such as the content of the dilemma. Moral orientation: a personal preference for a moral perspective or decisional strategy that one tends to use in interpreting or resolving moral dilemmas or both (Yu & Kishor, 2001a). Moral self: the self is defined in terms of moral traits, such as caring, compassionate, fair, generous, hardworking, helpful, honest, friendly, and kind. Norm or norm-maintaining orientation: social responsibility implies maintaining the given social order and upholding the existing laws and regulations. Religious self: the self is defined in terms of religious faith or spiritual life. Religious/Spiritual orientation: morality is guided by one‘s religious or spiritual beliefs. Religious worldview: the view that God is the creator of life, spirit is fundamental, and matter is derivative (Smith, 2001). Spiritual worldview: the view that the world is governed by some spiritual power, force, being(s), or laws (King, Speck, & Thomas, 2001). Social Darwinism: the view that humans, like animals, must compete in a struggle for survival. Vertical individualism: denoted by linear relation to authority and individualist relation to group (Triandis & Gelfand, 1998). Vertical collectivism: denoted by linear relation to authority and collectivist relation to group (Triandis & Gelfand, 1998). Vertical worldview: denoted by linear relation to authority, hierarchy, power, and achievement (Triandis & Gelfand, 1998). Worldview: a set of assumptions about physical and social reality that may strongly influence cognition and behavior (Koltko-Rivera, 2004).  xv  ACKNOWLEDGEMENTS I would like to express my gratitude to all the participants in the pilot study, the student study, and the adult study for sharing their worldviews, self-views, and inner thoughts with me. I am particularly indebted to the anonymous adult participants at UBC and University of Victoria, who did not only participate in the study but also helped recruit other participants at UBC and University of Victoria and in their local communities. I would also like to extend my gratefulness to the participating teachers and principals in British Columbia and instructors in the Faculty of Education at UBC for their time and effort in making this research possible. My appreciation is equally extended to the Faculty of Education, the Sauder School of Business, Departments of Biology, Chemistry, Electric Engineering, and Zoology at UBC, the Caribbean African Association UBC, the UBC Alma Mater Society, and the Tung Lin Kok Yuen, Canada Society in Vancouver for allowing me a platform to distribute survey forms to potential adult participants. Finally, I would like to thank my Program Advisory Committee: Professors Nand Kishor (the Advisor), Marion Porath, and Mary Bryson for their support and guidance for all these years, and my Research Supervisory Committee: Professors Nand Kishor (the Supervisor), Marion Porath, and Seong-Soo Lee for their patience, encouragement, and valuable advice on my thesis.  xvi  DEDICATION To all my research participants  1  CHAPTER ONE: INTRODUCTION The overall goal of this dissertation was to develop a comprehensive model of moral development (Figure 1) that relates the constructs of ―worldview‖ and ―self-view‖ to represent socio-cultural and personal influence on moral development. The specific purposes were: (a) to examine the relational structures among worldview, self-view, moral inclusiveness, and moral orientation in adolescence and adulthood; and (b) to construct a multidimensional moral orientation scale that includes various moral orientations reflecting different worldviews and self-views. It is hoped that this research might contribute to the integration of the diverse research literature on moral development. Background to the Problem In the second half of the past century, the cognitive-developmental approach, particularly Kohlberg‘s (1981, 1984) stage theory of moral judgment development, was the dominant view as well as the subject of debate in the field of moral psychology. Kohlberg proposed that moral judgment develops in universal and invariant stages and that justice as a culturally universal or nonrelative moral concept is the primary virtue. Kohlberg claimed that his theory accounting for the development of justice structures through an invariant sequence of stages could lead us out of the problems of both individual and cultural relativity. The Kohlbergian theory, however, was criticized for neglecting the impact of culture and social context in morality (e.g., Bandura, 1991, 2002; Simpson, 1974; Snarey, 1985; Tappan, 1997). Empirical research indicated that moral judgment might be less of a developmental variable than a personality factor (Carr, 2002; Gibson, 1990). Theorists also argued for the need to embrace moral reasoning strategies or moral orientations outside Kohlberg‘s scheme (e.g., Snarey & Keljo, 1991; L. J. Walker, Pitts, Hennig, & Matsuba, 1995).  2  So far, a more comprehensive model of moral development that takes into account the various moral orientations as well as the influence of culture and personality factors is still lacking. It would thus seem timely and essential to develop a model of moral development to integrate different cultural and personal aspects with various moral orientations, by utilizing the constructs of worldview and self-view. This model might help to explain the relativity problem and the relation between ―is‖ and ―ought‖ in morality. The Research Problem Worldview is a notion common in everyday language as well as in academia. Scientists (e.g., Bohm & Hiley, 1993), educators (e.g., Ambrose, 2000), and researchers (e.g., Guba & Lincoln, 1994) have used worldview to refer to a theoretical perspective or a set of basic beliefs or philosophical assumptions that can affect the nature and the course of their practice. Only recently, however, has Koltko-Rivera (2004) provided substantive evidence to justify ―worldview‖ as a valid psychological construct. Koltko-Rivera defined worldview as a set of assumptions about physical and social reality that may strongly influence cognition and behavior and proposed a comprehensive, multidimensional model of worldview. In this research, I focused on the existential aspects of worldview that are most pertinent to moral development. The influence of worldview on morality has been suggested by many theorists (e.g., Rest, Narvaez, Bebeau, & Thoma, 1999; Vandenberg, 1991). For example, Jankowski (1991) described a link between a social Darwinist worldview (the view that humans, like animals, must compete in a struggle for survival) and predatory behavior in young gangsters in urban America. These youngsters grew up in a neighborhood that was filled with competitiveness, illegality, and aggressive behavior. Though not deliberately taught, they developed a social Darwinist worldview that guided their moral reasoning and behavior.  3  I (Yu, 1998) also reported a case study regarding a Chinese woman who converted herself to Buddhism. The dominant theme in the participant's life was that her faith changed her worldview from a self-centered, materialistic worldview to a spiritual one that stressed the interconnectedness of all beings, the law of karma, and principles of cause and effect. Prior to converting to Buddhism at age 32, she was egocentric, haughty, extravagant, stubborn, and quick-tempered. Because of a terrifying bloody gang-fight that she witnessed when she was 12, she had extreme aversion to the sight of blood, which would remind her of the horrifying bloody scene and make her feel anxious and distressed. She would walk away from any accidents, trying to avoid seeing blood. After her conversion, she was able to focus her attention on others in pain rather than her own personal distress. She became capable of regulating her own fear and distress, empathizing with others including accident victims in blood, and behaving altruistically. This research provided tentative evidence for a link between worldview and moral development. Compared to worldview, the self is a popular and much researched construct among psychologists. The psychological self or self-concept is generally referred to as a class of beliefs, conceptions or cognitions that one has about oneself, including one‘s physical image and psychological traits, abilities and behaviors, attitudes and values, social roles and ethnic identity, and so on (Rokeach, 1973). The influence of self-view on moral orientation has also been well documented by researchers (e.g., Gilligan, 1982; Lyons, 1983). Because self-view, self-concept, self-conception, self-construal, and self-perception are synonymous, they will be used interchangeably henceforth. Theorists have argued that conditioned by the social (class), (sub-) cultural, and historical context in which it is situated (e.g., Baumeister, 1997; Cushman, 1990; Danziger, 1997; Holland, 1997), the specific content of self in a culture reflects the culture‘s worldview  4  and philosophical beliefs (Markus & Kitayama, 1991; L. J. Myers et al., 1991). Thus, worldview and self-view are closely related. It is not clear, however, how worldview and self-view are related to moral development, because empirical investigation of the interrelations among these constructs simultaneously is almost non-existent. I attempted to do just that in this study, from a holistic and complementary perspective. A Holistic and Complementary Perspective In most Eastern cultures, it is commonly believed that everything contains pair(s) of opposites that are interdependent and complementary to each other, and that wholeness or harmony could only be achieved by combining these opposites. Thus, the balancing of opposite tendencies and elements is greatly emphasized; imbalance or extremism is seen as the cause of physical and mental pathologies (Sinha & Tripathi, 1994). The concept of interdependence of opposites can be understood in the facts that one can see beauty as ―beauty‖ only as contrasted to ugliness and good as ―good‖ as contrasted to evil. That is, beauty and ugliness, good and evil are all relative, neither one can exist without the other. Therefore, the explicit duality of polarities embodies an implicit unity. Balance between the opposites is called the Middle Path or Golden Mean, representing a state of ideal unity and harmony of the polarities (Lao Tzu, 1972). The principles of interdependence of opposites can be seen everywhere, even in the field of quantum mechanics. According to classical physics, waves (carriers of energy) and particles (basic units of matter) are distinct. Nevertheless, subatomic particles can manifest themselves in either waves or particles as a response to the scientists‘ experimentation. That is, mind and matter are not distinct. In other words, objectivity is impossible and there is no external material reality independent of our consciousness. Bohr (1934) used the term complementarity to express the paradox of quantum interconnectedness of wave and particle. The concept of  5  complementarity as a theoretical framework that acknowledges the simultaneous coexistence of mutually exclusive descriptions of psychological phenomenon has also been advocated by a number of psychologists (e.g., Hyland, 1985; Snyder, 1983). Morality is characterized by the self-other(s) polarities that characterize human life and result in conflicts in social and interpersonal relationships that require well-balanced solutions (Yu, 1999). A complementary view is thus compatible with the current study. Stressing the interdependence and unity of opposites, the complementary view is also holistic. A holistic view of the individual functioning and developing as an integrated (among the individual‘s mental, behavioral, and biological components), indivisible organism and as part of an integrated, complex, and dynamic person-environment system has been advocated by many psychologists (e.g., Allport, 1937; Bergman, Cairns, Nilsson, & Nystedt, 2000). The individual and the environment in which individual functioning and development occurs are thus inseparable. To understand individual functioning, we must consider the environment in which it occurs. Magnusson and Stattin (2006) distinguished three environmental positions along the dimension of proximity to the individual‘s experiences: the immediate, the proximal, and the distal environments. The immediate environment refers to the temporary situations in which an individual‘s current functioning takes place. An individual‘s current functioning is contingent on the functional interaction with its immediate environment, mainly social interaction with other individuals (e.g., parents, siblings, peers, and teachers). The proximal environment, which is to some extent individually unique, refers to the social relationships the individual has with family members, peers, and other individuals in various settings, such as at home, school, and neighborhoods. Patterns of individual functioning established across situations over time characterize the individual‘s personality and social relationships or proximal environment. The  6  distal environment refers to the social, political, cultural, and physical settings that envelop and shape proximal environments, at a more general level. Thus, the three levels of environment are interconnected. Each level of environment builds on the previous level, but has its own valid existence and an implicit, stable relationship with the level above and below. Moreover, according to Magnusson and Stattin (2006), a holistic approach considers individuals as active, intentional agents in the person-environment interaction process at all levels of the person-environment system. As active and purposeful agents, individuals select and interpret information from the environment and transform the information into internal and external activities (e.g., feelings, thoughts, actions) to deal with the changing environment or to change the environment. Such processes are characterized by continuously ongoing interactions among mental, biological, and behavioral components of the individual and social, cultural, and physical components of the environment. Such a holistic view ―implies that a guiding principle in the individual‘s inner life and in his or her dealings with the external world resides in the functioning of the integrated mental system, including self-perception, perception of others, and worldviews‖ (p. 409). A holistic perspective is thus most suitable for the current study in search of the guiding principles in the individuals‘ inner life and their dealings with morality. By emphasizing the complementary characteristic of the holistic perspective described above, the holistic and complementary approach herein highlights the dualistic aspects of life (e.g., person-environment, self-other) and regards achieving a balance between opposites as the ideal state of human and moral development. Consequently, my holistic and complementary perspective is not value-neutral and differs from most holistic models (e.g., Allport, 1937; Bergman et al., 2000; Magnusson & Stattin, 2006) that simply offer a way to understand or explain individual functioning and development.  7  Adopting a holistic and complementary approach to the study of worldview, self-view, moral inclusiveness, and moral orientation, I hoped to integrate cultural psychology with moral psychology as well as different schools in moral psychology. By examining the relational structures among these constructs, I tried to reveal the dynamic relations of cultural and personal factors with moral inclusiveness and moral orientation. To investigate moral orientation more inclusively, I initiated the construction of a multidimensional moral orientation scale to include various moral orientations that reflect diverse worldviews and self-views. Significance of the Study Dichotomized theorizing undermines the interdependent and complementary nature of the opposites (e.g., Guisinger & Blatt, 1994). Decontextualized theorizing failing to take into account social, cultural, and contextual factors influencing human behavior (Cushman, 1991; Hansford & Hattie, 1982) can hamper our understanding of human motivation and development (Cushman, 1991). A comprehensive theory of moral development that takes into account the influence of culture and individual differences on morality is needed to guide psychological research and education reform. Integrating morality with worldview and self-view that represent cultural and personal influence on moral development simultaneously, my proposed model of moral development is contextualized and holistic. It will therefore be valuable in theory and research and affect the thinking behind moral education. Research in moral psychology has been guided by different approaches in the last century. Diversity in the field of moral psychology is beneficial because the complexity of morality requires investigation from various angles. Knowledge gained from diverse perspectives must be integrated, if moral psychology is to be advanced. Integration of diverse research literature on morality, however, has been overdue. In this study, I have tried to fill this  8  gap by developing a multidimensional moral orientation scale that includes various moral orientations. These diverse orientations reflect not only different worldviews and self-views, but also the various reasoning strategies identified by researchers as commonly used by people in making their real-life moral decisions. The multidimensional moral orientation scale to be developed can thus be used to understand the diverse moral orientations employed by an individual or a group of individuals. That is, such a scale could be a useful tool for research into moral orientation and the complexity of inner moral life at the individual and the group levels. Further, it could also help school counselors and teachers in understanding, counseling, and educating students in their personal, social, and moral development. The results of this study may also provide new insights into the theory and research of moral development. Summary and Overview To investigate the influence of culture and personality on moral development, I explored the relational structures among worldview, self-view, moral inclusiveness, and moral orientation simultaneously in adolescence and adulthood. To this end, I surveyed a sample of high school students and a sample of adults. The next chapter starts with a critical review of the literature, followed by a proposed model of moral development and research hypotheses. Then I describe the methodologies in Chapter 3 and present the findings in Chapter 4. Finally, in Chapter 5, I discuss the limitations of the study, the implications of the results, and future research directions.  9  CHAPTER TWO: REVIEW OF THE LITERATURE In this chapter, I first review the literature on moral development. Based on the review, a comprehensive model of moral development is proposed. The model tries to account for the influence of cultural and personal factors on moral development by explicating the interrelations among worldview, self-view, and moral development. I then explain how the constructs were operationalized and present my research hypothesis. Theory of Moral Development The Kohlbergian Theory Piaget (1932/1965) proposed that children‘s moral judgment develops in two stages, from concrete and external to abstract and internal. Adopting Piaget‘s cognitive developmental approach to moral judgment, Kohlberg (1971) expanded Piaget‘s two-stage model into a sixstage model of moral judgment and claimed that moral judgment developed in universal and invariant stages. The six stages are grouped into three levels of development: the preconventional (Stages 1 and 2), the conventional (Stages 3 and 4), and the postconventional (Stages 5 and 6). Stage 1 is defined by an Obedience and Punishment Orientation that values deference to power and avoidance of punishment to satisfy one‘s needs, whereas Stage 2 is defined by an Egoistic and Instrumental Orientation where reciprocity and fairness are based mainly on self-serving purposes. At Stage 3, good behavior is defined in terms of getting social approval, maintaining good relationships, and living up to social expectations, reflecting a Social Approval Orientation. At Stage 4, respect for authority, maintaining the given social order, and fulfilling social duties take precedence over interpersonal responsibility, reflecting a Maintaining Norms Orientation. Stage 5 is marked by a Social Contract Orientation. Awareness of the relativism of personal values and opinion leads to an emphasis on procedural rules for reaching  10  agreement. Duty is defined in terms of contract and community welfare. Stage 6 is marked by an Ethical Principle Orientation that morality goes beyond obeying the legal standards and is guided by the conscience and self-chosen moral principles of justice that are abstract, universal, and consistent. The Moral Judgment Interview. Kohlberg collected data by interviewing a sample of White adolescent boys about hypothetical moral dilemmas (e.g., Heinz and the drug) on what action ought to be done (e.g., should Heinz steal the drug to save his dying wife?) and why. Interview data were then scored by trained scorers based on a scoring system. Kohlberg referred to this procedure as the Moral Judgment Interview (MJI; see Colby & Kohlberg, 1987). Kohlberg (1981, 1984; Colby & Kohlberg, 1987) redefined his theory of moral development and the scoring system for interview data, on the basis of empirical findings from his 20-year longitudinal study of the White male sample. Due to the fact that none of the longitudinal participants had evidenced a focus of Stage 6 reasoning, Stage 6 became a theoretical stage. Kohlberg, however, retained his claim that the stages are invariant and universal, with hierarchical integrations of structured wholes, albeit his all White male sample. Research evidence for Kohlberg‘s cognitive-developmental model based on the current MJI and its scoring system has been inconsistent. On one hand, cross-cultural longitudinal studies (Colby, Kohlberg, Gibbs, & Lieberman, 1983; Snarey, Reimer, & Kohlberg, 1984) reported few instances of stage skipping or regression, supporting the notion of invariant and universal stage development. On the other hand, research indicated that respondents scored higher on Kohlberg‘s hypothetical moral dilemmas, but lower on real-life or real-life-like moral dilemmas, suggesting that moral stages can be a function of dilemma content (e.g., Carpendale &  11  Krebs, 1995; Krebs, Denton, Vermeulen, Carpendale, & Bush, 1991; L. J. Walker, de Vries, & Trevethan, 1987; Wark & Krebs, 1996, 1997). Besides, reviews of cross-cultural studies using the MJI indicated that Kohlberg‘s test is biased toward urban societies and middle-class populations and under-represents some culturally specific moral judgments (Snarey, 1985; Snarey & Keljo, 1991). For example, a communitarian orientation is common in folk-cultural groups and working-class communities (Snarey & Keljo, 1991). The Confucian concepts of filial piety (i.e., children should obey and take care of their parents) and collectivistic orientation are documented in Chinese populations (Dien, 1982; L. J. Walker & Moran, 1991). Compared to the Americans, Hindu Indians tend to value interpersonal interdependence over individual autonomy in their moral reasoning (J. G. Miller & Bersoff, 1995). The concepts of karma and dharma, and the role of compassion are fundamental to a Buddhist worldview and morality (Heubner & Garrod, 1993). All this research suggests that cultural worldviews may predispose individuals toward the use of or preference for certain moral reasoning strategies. Furthermore, studies on real-life morality conducted in the United States (Colby & Damon, 1995) and Canada (L. J. Walker et al., 1995) showed that spiritual or religious beliefs (e.g., universal love) are involved in everyday morality. Colby and Damon (1995) reported that moral judgment scores were related to participants‘ educational levels rather than to participants‘ moral commitments, because only half of the moral exemplars (people who have shown extraordinary moral commitment) attained Kohlberg‘s postconventional level, but the other half scored at Kohlberg‘s conventional level. Colby and Damon concluded, ―One need not score at Kohlberg‘s highest stages in order to exhibit high degrees of moral commitment and exemplary behavior‖ (p. 367). Moreover, individual differences in verbal fluency and scorer subjectivity  12  tend to complicate the assessment of moral judgment in the MJI (Rest, 1979). Colby et al. (1983) and Snarey et al. (1984) proposed that confound of the MJI with education and verbal ability may be partly responsible for the developmental findings reported by Kohlberg. The Defining Issues Test. There were disagreements among Kohlbergian psychologists, as well. Gibbs (1979) argued that postconventional stages could be an existential phase of adult development in metaethical positions. Retaining the postconventional stages, Rest‘s (1976, 1979) additive, complex stage model suggested that the acquisition of higher stages results in no loss of availability of the lower stages, and that individuals may use many types of reasoning simultaneously. Rooted in the complex stage model, the Defining Issues Test (DIT; Rest, 1993; Rest, Cooper, Coder, Masanz, & Anderson, 1974) was developed to test moral judgment development in the Kohlbergian tradition on a continuum. As a production task, the MJI asks for the participant‘s spontaneous production of moral judgments. In contrast, as a recognition task, the DIT tests participants‘ levels of preference for different moral thoughts and may thus avoid confound of verbal ability (Rest, 1979). Derived from Kohlberg‘s (1969) theory and interview data, the DIT is a rating and ranking test assessing how people define and judge moral issues in dilemmas. Unlike the MJI, the DIT is able to assess Stage 6 thinking in participants, because recognition and preference tasks are easier ways of manifesting discrimination of higher levels of thinking than are production tasks. Stage 1, however, is not assessed in the DIT (Rest, 1979). On the basis of 25 years of research using the DIT, Rest et al. (1999) postulated three moral schemas: Personal Interest, Maintaining Norms, and Postconventional schemas. Supposedly originating in childhood, the Personal Interest schema includes concerns for oneself as well as for one‘s kin, friends, and personal acquaintances, thus combining Kohlberg‘s Stages 2 and 3 because the scores of these two stages were correlated in DIT data and fused into a single  13  factor in factor analysis. It is considered primitive, for it fails to attain a sociocentric perspective (i.e., to consider how cooperation is organized in society). Derived from Kohlberg‘s Stage 4, the more advanced Maintaining Norms schema begins in adolescence and it signifies the attainment of a sociocentric perspective. Derived from Kohlberg‘s Stages 5 and 6, the Postconventional schema is the most advanced and based on sharable ideals and logical coherence. Nevertheless, the DIT is plagued by the same problems that have plagued the MJI. Similar to the findings with the MJI, research with the DIT indicated that participants scoring high or low on principled scores did not differ in their responses in real-life moral dilemmas, confirming that Kohlberg‘s model of moral judgment fails to address everyday morality (Wygant, 1997). Confound of the DIT scores with education and verbal ability (Chovan & Freeman, 1993; Sanders, Lubinski, & Benbow, 1995) may be partly responsible for the reported developmental findings in the DIT, as they are in the MJI. The DIT also fails to represent other forms of moral reasoning, such as religious reasoning, so moral judgment of people who base their moral decision in terms of religious beliefs is likely to be underestimated or misconstrued as Stage 4 reasoning (cf. Rest et al., 1999). Gilligan’s Theory of Two Moralities There is another area of confusion in Kohlberg‘s theory. As mentioned earlier, Kohlberg (1971) first defined moral stages in terms of moral orientations. In typologizing moral philosophic theories, Kohlberg (1976) later identified four primary moral categories and called them ―moral orientations‖: normative order, utilitarian, ideal-self or perfectionism, and justice orientations. Kohlberg argued that these moral orientations could be found at each of the moral stages and ―define four kinds of decisional strategies, each focusing on one of four universal  14  elements in any social situation‖ (p. 40). Kohlberg, however, considered justice orientation as the basic moral principle that signifies the highest development of moral judgment. Kohlberg‘s preference for individual rights and formal principles of justice has incited a lot of controversy in moral psychology (Puka, 1991). The most notable was Carol Gilligan‘s (1982) assertion of two moralities, ―justice‖ and ―care,‖ that are gender specific. Gilligan contended that women have a moral ―voice‖ or caring orientation that is different from Kohlberg‘s justice orientation. Gilligan argued that a morality of justice, focusing on rules, rights, reciprocity, autonomy, rationality, and impartiality, fails to account for the morality of females. According to Gilligan, the morality of females, anchored in empathic affects and compassion, focuses on care and responsibility, and avoids inflicting harm on others. Gilligan (1982; Gilligan & Attanucci, 1988) asserted that self-conceptions are related to conceptions of morality in women‘s thinking. According to Gilligan, the identity of a woman is defined in terms of relationships and social connections. Thus, women tend to view detachment or abandonment as problems. Responsibility and caring about self and others becomes the moral imperative and requires the construction of a moral conflict in its contextual particularity. On the contrary, Gilligan argued, a justice perspective focuses on problems of inequality and oppression to be resolved by the notions of reciprocal rights and equal respect. Respecting the rights of others requires a separation identity and noninterference, that is, not to interfere with others‘ autonomy. Thus, the two orientations are incompatible and gender-related. Gilligan (1982) first proposed a three-level model of care ethic to explain females‘ social experience and moral development, based on interviews with women in consideration of abortion. Later on, Gilligan (1990; Gilligan, Brown, & Rogers, 1990) retreated from her developmental model of care ethic in favor of a literary framework for describing moral  15  orientations. Gilligan and Wiggins (1987) proposed that the development of morality is rooted in and proceeds through relationships that start in early childhood. Originating in early childhood experience and self-conception, moral orientation becomes a personality rather than a developmental variable (cf. L. J. Walker, 1997). Research with adolescents and adults indicated that more females than males possess a feminine or relational self-conception that mediates gender differences in moral orientation (Lyons, 1983; Pratt, Golding, Hunter, & Sampson, 1988; Skoe & Marcia, 1991). That is, both men and women may subscribe to the ethic of care or justice, depending on their self-concept. Because self-concept may have a direct influence on moral orientation, regardless of gender, this study investigated the relations between self-concept and moral orientation in a gender-neutral way. Gilligan (1995) further made a distinction between a ―feminine‖ ethic of care and a ―feminist‖ ethic of care. The former is selfless or self-sacrificial care, based on faulty interpersonal relationships, thus signifying an opposition between relationships and selfdevelopment as well as a disconnection from emotions within a patriarchal world, whereas the latter begins with connection, and the relational world includes not only close interpersonal relationships, but also international relationships and relationships with the environment. Gilligan‘s notion of feminine care is somewhat similar to Kohlberg‘s (1976) Stage 3 good boynice girl orientation. A feminist care orientation, however, is akin to Campbell‘s (1983) notion of biocentrism that regards all life, including humans, animals, and plants as having intrinsic value (Nash, 1989). Nevertheless, the distinction between the two kinds of care ethics has largely been ignored by researchers. Treating (feminine) care and justice moral orientations as unidimensional constructs and trying to find intrapersonal consistency in moral orientation across dilemmas,  16  most researchers (e.g., Garrod & Beal, 1993; D. K. Johnson, 1988; L. J. Walker, 1989, Wark & Krebs, 1996; Yacker & Weinberg, 1990) investigated gender differences in moral orientation with dilemma-based measures, either with the MJI or DIT procedures. A meta-analysis on gender differences in moral orientation (Jaffee & Hyde, 2000) reported that small gender differences for both moral orientations in the hypothesized directions (i.e., males scoring higher on justice and females higher on care) were found with dilemma-based measures and that moderately large gender differences were found among measures that did not include a dilemma. Jaffee and Hyde suggested that research on gender differences in moral orientation must avoid confound with dilemma content. Defining Moral Development, Moral Orientation, and Moral Judgment With regard to the issue of confounding moral orientation with dilemma content, Yu and Kishor (2001a) proposed to differentiate between dispositional and situational moral orientation. They defined dispositional moral orientation as a personal preference for a moral perspective or decisional strategy that one tends to use in interpreting or resolving moral dilemmas or both, whereas, situational moral orientation is one‘s use of a moral perspective or decisional strategy for interpreting or resolving a specific moral dilemma or both. For illustration, a person who has a general preference for Stage 2 egoistic orientation as his or her dispositional moral orientation is likely to choose Stage 2 reasoning as his or her situational moral orientation when facing a dilemma that involves conflicts between self-interests and relational responsibility, such as in the Heinz dilemma in the DIT. In this story, the conflicts between self-interest (e.g., the risk of getting shot as a burglar or going to jail for stealing the drug) and relational responsibilities (stealing the drug to save one‘s spouse) is spelled out in one of the statements. Conversely, he or she may switch to Stage 3 or 4 reasoning as his or her situational moral orientation when dealing  17  with a dilemma that involves little or no self-interest. For instance, in the DIT story that asks whether an escaped prisoner who had been a good and productive citizen for the past eight years since his escape from prison should be reported and sent back to prison, no stage 2 item is presented. When self-interest is down played, he or she may focus on getting social approval or maintaining law-and-order. This may help explain why participants scored higher on hypothetical dilemmas, but lower on real-life dilemmas (e.g., L. J. Walker et al., 1987; Wark & Krebs, 1996, 1997), because hypothetical situations do not truly involve the participants‘ selfinterests, whereas real-life situations often do. Situational moral orientation thus reflects the interplay between dispositional moral orientation and contextual factors, such as the content of the dilemma. In such a sense, the MJI and the DIT, assessing moral judgment with hypothetical or real-life dilemmas could be considered as measures of situational moral orientation. In other words, the construct of moral judgment is essentially the same as the construct of situational moral orientation, that is, one‘s use of any moral orientation(s), moral perspective(s), decisional strategy(s) or moral stages in the Kohlbergian language for interpreting or resolving or making decisions for certain moral dilemmas. Defining moral judgment and moral orientation. It has been assumed that moral orientation is conceptually independent of level of moral reasoning and thus may be related more closely to the content, rather than the structure, of reasoning. However, an empirical confound between moral orientations and moral stages in the Kohlbergian typology has been reported, with the normative and utilitarian orientations more typical at lower stages and the justice and perfectionism orientations more typical at higher stages (L. J. Walker, 1989). These findings suggest that it is difficult to separate content from structure, especially when structure or stage  18  can be confined by the content of the dilemma (e.g., Carpendale & Krebs, 1995; Krebs et al., 1991). These findings also suggest overlap between moral judgment and situational moral orientation, when they are both assessed with moral dilemmas. Thus, moral judgment is defined herein as the use of a moral perspective or decisional strategy in interpreting or resolving moral dilemmas that is assessed with a dilemma-based measure. To retain its dispositional connotation, moral orientation is defined as a personal preference for a moral perspective or decisional strategy that one tends to use in interpreting or resolving moral dilemmas or both that is assessed with a dilemma-free measure. To define moral judgment as situational moral orientation may serve to connect the construct of moral judgment with the construct of moral orientation. While research has identified more and more moral perspectives or orientations (e.g., filial piety) beyond Kohlberg‘s (1981, 1984) scheme, the usefulness of Kohlberg‘s developmental theory seems to have diminished. To be able to accommodate the increasingly diversified morality in the 21st Century, psychologists in the field of moral psychology need to find a bridge to integrate the diverse knowledge gained in the last century. Defining moral judgment as situational moral orientation does not only provide such a bridge, but also allows for the expansion and revision of the Kohlbergian developmental model to reflect the multiplicity of morality in this global age. The construct of moral orientation enables us to assess moral development without confound of dilemma content and to examine interactions between dispositional and situational factors in moral reasoning. Defining moral development. According to Dawson (2003), developmental stages can be considered as orders of hierarchical complexity. Focusing on the levels of abstraction in moral reasoning, the Kohlbergian approach describes moral development as the development of moral  19  concepts evolving from simpler ideas to ideas that are more complex, such as the shift from conventional to postconventional thinking (Rest et al., 1999). Focusing on the cognitive structure of moral judgment, the Kohlbergian approach confounds moral development with education and verbal ability (Colby et al., 1983; Sanders et al., 1995; Snarey et al., 1984). Defining moral development from a cognitive approach also fails to explain why the attainment of postconventional thinking is unrelated to exemplary moral behavior and commitment (Colby & Damon, 1995). One solution to these problems is to define moral development in terms of perspective taking. Perspective taking is a developed capacity and the ability to take another‘s perspective is usually developed adequately by mid-adolescence (Selman, 1976). Kohlberg (1981) stressed that, like cognitive-developmental levels, perspective-taking ability setting the upper-bound limits on moral reasoning is essential but not a sufficient condition for moral development. Kohlberg further postulated that there exist three major levels of social perspective (i.e., the standpoint one takes in defining social facts and moral values) corresponding to the three major levels of moral judgment development. The preconventional level is marked by a self-interested egoistic social perspective, and the conventional level is denoted by a member-of-society perspective. The postconventional level is characterized by a prior-to-society perspective where individuals rationally define their values and principles in a generalizable way. So far, research has not been able to provide evidence for the priority of logical to moral reasoning in development (Damon, 1975; Dawson, 2003). In a sample of juvenile delinquent boys, positive correlations among cognitive development (measured by the pendulum task and balance tasks), perspective-taking stage, and moral judgment were found. When perspectivetaking ability was partialed out, the statistically significant correlation of cognitive development  20  with moral reasoning disappeared, but the correlation of perspective taking with moral judgment remained statistically significant when cognitive development was partialed out (Lee & Prentice, 1988), suggesting that perspective taking plays a more important role than cognitive development in moral reasoning. Research that perspective taking may lead to a merging of the self and others (Davis, Conklin, Smith, & Luce, 1996) and that perception of oneness or merging of the self and others may result in helping behaviors (Maner et al., 2002) further elucidates the role and importance of perspective taking in morality. Nevertheless, perspective-taking ability is essential but not a sufficient condition for moral development, as pointed out by Kohlberg (1981), because possessing perspective-taking ability does not guarantee the use of it, particularly during moral reasoning. On the contrary, as a personality trait, perspective-taking tendency has low correlations with cognitive or intellectual ability in adolescents and adults (Bernstein & Davis, 1982; Hogan, 1969). These findings suggest that perspective-taking tendency is not an ability but a social or personal style, that is, a dispositional likelihood to use one‘s role-taking ability to receive the others‘ viewpoints (Davis, Luce, & Kraus, 1994). Thus, one way to define hierarchical complexity in the moral domain is to define it as the tendency (as a personality trait rather than an ability) to take an increasingly wider perspective or more standpoints from different groups in defining social facts and moral values, as suggested by many theorists (e.g., Carpendale, 2000; Carter, 1980). Carter (1980) argued that Kohlberg (1971) should focus on delineating ―a sequential typology of development in moral thinking from egoism to universalism … rather than something unnecessarily contentious,‖ such as linking principled morality to any normative ethical position (p. 88). Carter further proposed that Kohlberg‘s moral stages could be reconceptualized as stages of ―what is to be included in one‘s ethical deliberations‖ or moral  21  inclusiveness (p. 94). The moral journey or moral development could thus be described as the stage-by-stage expansion of one‘s moral inclusiveness, represented by seven concentric circles with Stage 1 as the smallest, innermost circle and Stage 7 the largest, outermost circle. The contents of moral inclusiveness starts from the self (Stage 1) to including at least one other (Stage 2), one‘s peer group (Stage 3), one‘s nation (Stage 4), the broader group described by a bill of rights (Stage 5), humans as a whole (Stage 6), and finally encompasses the cosmos (Stage 7). When one‘s ethical consideration includes more and more people or groups, one‘s moral reasoning matures from preconventional to conventional, then postconventional, and finally Stage 7 cosmic reasoning. The notion of a Stage 7 is not new. Kohlberg (Kohlberg & Ryncarz, 1990) theorized about a ―metaphorical‖ Stage 7 since 1973 and suggested that it may provide answers to the ultimate question of ―why be moral?‖ in an immoral world. According to Kohlberg and Ryncarz, a metaphorical Stage 7 involves experience or concepts of a non-egoistic or nondualistic variety, characterized by a sense of unity with the cosmos, nature, or God (which includes and represents all the higher beings in any religious or spiritual beliefs in this thesis) that is underlined by a natural law theory or orientation. As stated by Kohlberg and Ryncarz, a natural law theory assumes that human moral law is a part of the larger natural law or order embodied in the cosmos. Natural law theory differs from divine command theory that asserts that morality is defined by divine command as revealed by the Bible or other documents of revelation. Kohlberg and Ryncarz further conceded that well developed moral intuitions, informing a general natural law, parallel intuitions about nature or ultimate reality. That is, Stage 7 reasoning can be reached by intuitions. This may help to explain why moral exemplars that relied on intuitions or spiritual or religious beliefs to guide their moral reasoning in Colby and Damon‘s (1995) study failed to  22  express an advanced form of moral judgment, because intuitive knowledge could be difficult to express, as contended by Rest (1973). Therefore, Stage 7 is a meaningful stage that should not be overlooked. Defining moral development in terms of an increasingly wider social perspective in moral reasoning seems to be a better alternative than defining moral development in terms of an increasingly complex cognitive structure. From a holistic perspective, however, the expanding concentric circles can be viewed or interpreted not only as increasingly wider social perspective or expanding moral inclusiveness, but also as a nested arrangement of structures, representing multiple levels of expanding environment, similar to Bronfenbrenner‘s (1979) multiple layers of context. Situated at the center of the concentric circles, the self can be conceived as embedded in multiple layers of expanding environment, as illustrated in Figure 1. That is, the multiple concentric circles represent both the different levels of environment that the individual may experience in life and the different levels of moral inclusiveness resulted from multi-level person-environment interaction. Conceiving the concentric circles as multi-levels of expanding moral inclusiveness and environment thus places moral inclusiveness development in the context of social relationships interacting with the individual‘s psychological and biological processes. In other words, the social environment provides opportunity for the development of moral inclusiveness. Whatever the individual brings into the environment, such as empathic skills and perspective-taking tendency interacts with the environment and results in moral inclusiveness. Because the concentric circles expand horizontally, it is better to use the term ―levels‖ rather than ―stages‖ to describe and designate the development of moral inclusiveness and the expansion of environment. Contingent on person-environment interaction, the development of moral inclusiveness need not follow a universal and invariant sequence. The simultaneous  23  existence of different levels of environment in our life allows for the simultaneous existence and development of different levels of moral inclusiveness in our mental systems. The degree of inclusiveness at a level reflects the person-environment interaction outcome at the same level, which may differ from that at another level. As a multifaceted construct, moral inclusiveness may help revise and expand Kohlberg‘s stage theory to include other theorists‘ ideas, such as Campbell‘s (1983) notion of biocentrism, which is of particular relevance today when environmental concerns are rising. To address environmental concerns and to provide a more comprehensive account of moral and human development, seven levels of moral inclusiveness are delineated, as shown in Figure 1. 1 Self  2 Family  3 Peers  4 Society  5 Humanity  6 Nature  7 Universe/God Figure 1. Seven levels of moral inclusiveness.  Due to an initial lack of differentiation between self and other (e.g., caregiver) in infancy (Hoffman, 1976), the infant and its environment is one. Thus at the first level of inclusiveness, there is only the self, including its own body and all its needs. The second level of inclusiveness begins when self-other distinction is made, and the self is expanded to include at least one other. This ‗other‘ refers to one‘s caregiver(s) or family, from a developmental point of view. Family constitutes the primary immediate and proximal environments for most young children, until  24  they have a chance to meet and play with their peers. By interacting with their cohorts, children develop social relationships outside their families and expand their inclusiveness to include their friends and peer groups at Level 3. As children go to school or the local community, their social environment expands. So does their moral inclusiveness from friends and peer groups to their schools, the local community, and eventually their nation at Level 4. For Level 5 inclusiveness, it should be anthropocentrism (i.e., the human race as a whole) rather than sociocentrism (i.e., the society) or the broader group as described by a bill of rights or constitution, as Kohlberg (1971) assumed. Whether the broader group is described by a bill of rights or constitution, its group members still share the same social structures and conventions, and most of all, collective interests. However broad the group is, it is just a group of people. Thus, it should be included in Level 4 inclusiveness instead. To reach Level 5, individuals must embrace humanity as a whole. Thanks to the advanced technology in this information age, we can now meet people from all over the world and listen to their stories and points of view on TV and the Internet. In a sense, advanced information gadgets can help create an anthropocentric environment for the new generations. I can see that nationalism and ethnocentrism are becoming outdated, as indicated by the borderless movements (e.g., Doctors Without Borders) and the notion of global citizenship. The notion of global citizenship can also be applied to Level 6, when the globe refers to the Earth as a whole. At Level 6, moral inclusiveness includes not only humanity but also other species and nature. Human beings are viewed as part of nature and no more important than other species. By treating humans and nature equally, Level 6 inclusiveness is essentially a broadening of Level 5 inclusiveness. Reading books or watching films that use human identities to construct anthropomorphic identities for animals and plants may help children and adolescents feel  25  empathy with nature and endow it with moral standing, so that we can create an environment for Level 6 inclusiveness without going out into the wilderness to experience nature directly. The willingness to humanize other animals and nature at Level 6 can be a declaration of common humanity that can bring out the best in people and make it difficult for them to remove humanity from their conduct (cf. Bandura, 2002). Level 6 inclusiveness thus reflects a decisive feature of human morality, namely, the humanization of all living and non-living beings, and should not be taken as another domain unrelated to human morality. At Level 7, moral inclusiveness further expands from nature to the entire universe or God, involving spiritual or religious experience that is characterized by a sense of oneness or connectedness (cf. Kohlberg & Ryncarz, 1990). When the self is merged with the universe or God (Yu, 1999), other-concerned feelings are often one‘s spontaneous reactions to other‘s conditions, and one acts accordingly, without the awareness of the actions being moral (Blum, 1987). These other-concerned feelings may lead to dissolution of the self-other distinction, however fleetingly and resolve the moral conflicts between self and others. Therefore, genuine goodness or morality comes from the unity of life or spiritual awareness of oneness. Newberg and Newberg (2006) argue that humans are born with essential brain functions underlying religious and spiritual experiences. Spiritual development, however, depends on support from the environment (Hay, Reich, & Utsch, 2006), such as family (Boyatzis, Dollahite, & Marks, 2006) and culture (C. N. Johnson & Boyatzis, 2006). Contingent on personenvironment interaction, spirituality can thus be viewed as a continuum and a universal phenomenon applying to all people, from the mystics and the devout believers on one end to the materialists and the earnest nonbelievers on the other (McSherry & Cash, 2004). Research has also confirmed the involvement of spiritual or religious beliefs in everyday morality (Colby &  26  Damon, 1995; L. J. Walker et al., 1995); even children have been found to have powerful insights in relation to spirituality (McSherry & Smith, 2007). Thus, spiritual or religious experience should not be excluded from a theory of moral development. The above description of moral inclusiveness delineating the types of environment giving rise to the expanding levels of inclusiveness is developmental and hierarchical. While Levels 2 and 3 inclusiveness involves concrete interpersonal relationships, Levels 4-7 inclusiveness involves social relationships with a group that require some form of operational thought and the ability to abstract, to form social concepts, and to extend empathy from one individual to the whole group. Similar to perspective-taking ability, the highest level of empathy is usually developed by late childhood, according to Hoffman (2000). Then empathy becomes a response partly to one‘s mental image of the other(s) and partly to the other‘s immediate distress, so that the child is able to empathize with a whole group of people. I also assumed that most individuals will have experienced most types of environment and developed certain degrees of inclusiveness at most levels of inclusiveness by early adolescence. Adolescence therefore deserved the most attention in this research. As we move from childhood to adolescence, adulthood, and old age, these types of environment are not lost, but undergo changes. In the same way, the focus of our moral inclusiveness at each level may change, transform, or transcend as our proximal and distal environments vary. For example, workplace may replace school for working adults at Level 4. When the child becomes a parent, the ‗other(s)‘ in the individual‘s Level 2 inclusiveness may change from his or her parents to his or her own children. For that reason, we do not ―reach‖ a lower level and then ―move on‖ to a higher one. The degree of inclusiveness at any level may  27  increase or decrease, fluctuate or stabilize, dependent on the continuous multi-level personenvironment interaction processes. Theoretically, Level 7 inclusiveness signifies the highest level of moral development, but it is not the whole picture. For example, if an individual fervently includes his or her God at Level 7, while excluding people of other religions at Level 5 or other animals at Level 6, he or she can hardly be considered morally developed. The complexity of human morality thus requires a holistic and complementary perspective that takes into account not only the multi-level person-environment interaction but also the multifaceted and often oppositional nature of human morality. To develop a comprehensive model of moral development based on such a holistic and complementary perspective was the primary objective of this dissertation to offer a more comprehensive picture of human morality. In sum, by transforming stages of moral judgment into levels of moral inclusiveness, the development of moral inclusiveness can be understood as the continuous ongoing multi-level person-environment interaction. Moral inclusiveness may orient one‘s attention toward the predicaments of those within one‘s moral inclusiveness, thus activating moral orientation compatible with one‘s moral concern. One‘s tendency to focus on a specific level of inclusiveness in defining social facts and moral values can be considered as one‘s (dispositional) moral orientation and assessed with dilemma-free measures. When it is assessed with dilemmabased measures (e.g., the DIT), it is called moral judgment, which reflects the interaction between moral orientation and situational factors. Based on these definitions, a comprehensive model of moral development employing the constructs of worldview and self-view to elucidate environment and personality influence on moral development is proposed in the next section.  28  A Comprehensive Model of Moral Development If moral development is the expansion of the circle of inclusiveness in moral deliberation (Carter, 1980), it is crucial that we understand what may predispose individuals to focus on a specific level of inclusiveness in moral reasoning, given developed empathetic and role-taking abilities. As argued by Wainryb (2004), prescriptive moral concepts are not applied in a vacuum, but against a background of specific beliefs about the relevant facts of reality, although these beliefs of reality may not be the true reality. I refer to these beliefs of reality or theories of the world as worldviews. As the background of morality, worldview affects not only how we interpret the moral meaning of social facts and values (e.g., Dien, 1997; Wainryb, 2004), but also our moral inclusiveness, which prompts us to adopt compatible moral orientation(s). Moral orientation then interacts with situational factors and affects our moral judgment and behavior. In other words, our worldview informs us what is moral, who should be included in our moral deliberations, and what should be done. What we perceive or believe the world ―is‖ thus results in us doing the ―ought‖ that is consistent with our worldview, despite its illusiveness. Moral behavior may in some way serve to maintain our worldviews, as Jensen (1997) put forward. Our worldview also affects our self-view (e.g., Markus & Kitayama, 1991; L. J. Myers et al., 1991), which predisposes us to embrace moral orientation(s) in accord with our self-view (Gilligan, 1982; Lyons, 1983). Because the world is a stage, where individuals play out their selves, one‘s perception or theory of the world (i.e., worldview) becomes the background of one‘s self-view. It thus seems plausible that worldview affects the development of self-view rather than the reverse.  29  On the basis of the above arguments, I proposed a comprehensive model of moral development that worldview influences self-view and moral inclusiveness. Then worldview, selfview, and moral inclusiveness jointly influence moral orientation, which in turn interacts with dilemma content and situational factors and influences moral judgment and behavior. When one‘s moral judgment is in harmony with one‘s moral orientation, moral judgment is likely to lead to compatible moral behavior. When one‘s moral judgment is inconsistent with one‘s moral orientation, moral behavior is likely to be inconsistent, too. When the situation requires quick decision and action, moral orientation may lead directly to behavior, bypassing moral judgment. The comprehensive model of moral development does not only present a way of organizing and describing moral behavior, as do most cognitive-developmental and social-cognitive theories (Selman, 1976), but also explains the ‗causes‘ of behavior. The proposed model of moral development that explicates the interrelations among worldview, self-view, moral inclusiveness, moral orientation, moral judgment, and moral behavior is presented in Figure 2.  Self-View  Moral Orientation  Worldview  Moral Inclusiveness  Moral Judgment  Behavior  Figure 2. A proposed comprehensive model of moral development. Human morality is conceptualized as a complex, integrated, adaptive open system in the proposed model, which can be conceived as consisting of two parts. The first part explains the genesis of moral orientation and the relations of worldview to self-view, moral inclusiveness, and moral orientation in the proximal and distal environment. The influence of culture and  30  individual characteristics (e.g., psychological needs, empathic capability) on moral inclusiveness and moral orientation is subsumed under the constructs of worldview and self-view. Integration of person-environment interaction processes at all levels of the person-environment system is assimilated into the construct of moral inclusiveness. Because there are seven levels of moral inclusiveness reflecting seven corresponding worldviews, self-views, and moral orientations, there are also seven sets of relational structures among worldview, self-view, moral inclusiveness, and moral orientation, one at each level of moral inclusiveness. Defined as a personal preference for a moral perspective or decisional strategy that one tends to use in interpreting or resolving moral dilemmas or both, moral orientation is the result of the interplay among worldview, self-view, and moral inclusiveness, as shown in Figure 2. The influence of distal environment (e.g., culture), proximal environment (social relationships), and personal factors on morality is thus interwoven into the construct of moral orientation. Moral orientation at a level thus sums up the person-environment interaction process at the same level of the person-environment system. Constituting one‘s values and attitudes, one‘s worldview, self-view, moral inclusiveness, and moral orientation are all part of one‘s mental system, thus somewhat trait-like and resistant to change once established. Among the seven moral orientations, one is likely to develop a preference for one or more moral orientations that one tends to use as situational moral orientation in resolving moral conflicts. Supported by one‘s dominant worldview, self-view, and moral inclusiveness, one‘s most preferred moral orientation(s) is likely to be regarded as the guiding principle(s) in the realm of morality. For example, when resolving the Heinz dilemma, a participant insisted that it was Heinz‘s moral principle to steal for or to get rid of his wife, depending on whether he liked her or not (Rest, 1979). It was obvious that this participant preferred an egocentric orientation  31  and considered it the guiding principle in not only his but also everybody‘s dealings with the external world. The second part of the model describes the relations among moral orientation, moral judgment, and behavior, given a specific moral situation (the immediate environment). One‘s preferred moral orientation(s) serves to sensitize oneself to certain issues in the particular situation. Moral judgment and the ensuing action are the result of the interplay among the competing moral orientations (as highlighted by the content of the dilemma), the specific context in which the dilemma takes place (e.g., hypothetical or real-life situation, the presence of bystanders or observers), and the time allowed for ethical deliberation. Conflict arises when there are two or more moral orientations competing for situational moral orientation. Competing moral orientations may stem from the individual having developed preferences for two or more dispositional orientations. The individual‘s most preferred dispositional orientation may also compete with a less preferred orientation, when the latter is underscored by the situational factors (e.g., the context and the content of the dilemma) as the most appropriate situational moral orientation. Either way, moral judgment is the resolution of the competing orientations. Indecision in moral judgment arises from having two or more equally appealing situational moral orientations. Non-action may be the result of indecision in moral judgment. However, even with a well-deliberated final judgment, action may not be forth coming when the final judgment is inconsistent with one‘s most preferred dispositional orientation or when the situation is seen as lacking urgency or being overly demanding or restricted. Behavior in a specific situation is thus a function of the individual‘s current internal processes reacting to and interacting with situational features. Because moral judgment and behavior are context specific, they are more state-like and less predictable.  32  Being context specific, moral judgment and behavior are difficult to assess. Research on moral judgment and behavior will require authentic or experimental settings that can be very expensive and time consuming. Due to the constraints of time and resources, it was impossible to examine the whole model in this dissertation. I thus focused on the interrelations among the first four factors of the model in this exploratory study, leaving out moral judgment and behavior for future investigation. Because there are no existing measures assessing the seven moral orientations (although the case can be made for more) reflecting the seven levels of moral inclusiveness, I endeavored to construct a multidimensional moral orientation scale for this research. Constructing such a multidimensional moral orientation scale and investigating the relational structures among worldview, self-view, moral inclusiveness, and moral orientation are essential for they pave the way for research into the whole or the second half of the model. Before I elaborate on the interrelations among worldview, self-view, moral inclusiveness, and moral orientation at each level of moral inclusiveness, I must first explain how worldview and moral inclusiveness were operationalized. Worldview As the background and basis of self-view and morality, worldview is an important but under researched construct. It is vital to identify worldview elements that have influence over the development of self-view, moral inclusiveness, and moral orientation at the same time. I chose individualism-collectivism, one of the most researched cultural dimensions to represent worldview elements because of its moral significance and its impact on self-view. Worldview as individualism-collectivism. Emerging from the biological sciences, Western psychology has been influenced by Darwin‘s theory of evolution by natural selection.  33  Selection can operate at many levels, such as the individual level and the group or collective level. Darwin used the principle of group selection to explain the evolution of human morality, suggesting that a group that possesses many members who are willing to sacrifice themselves for the common good would be victorious over most other groups. In the 1960s, however, group selection was attacked by many biologists and the idea of genuinely altruistic behaviors being part of our evolutionary legacy was abandoned. With the belief that, to survive, individuals must act in their own interest to secure the scarce resources, the concepts of individual selection and human selfishness became established in evolutionary biology (Sober & Wilson, 1998). When individuals are regarded as the basic units of selection and survival, egoism is seen as necessary and selfishness as virtuous (e.g., Hardin, 1977; Perloff, 1987; Rand, 1964). Morality thus takes on different meanings for individuals who embrace individualism. In congruence with individual-level functionalism, the development of self-concept in humans was considered a means of competing with one another for limited resources (Gallup, 1997). When individuals rather than collectives are regarded as the basic units of selection and competition, individual goals logically precede group goals. Self-reliance means one must be able to compete with others for scarce resources, and one is free to do so in one‘s own way. Interpersonal and social relationships are conducted in utilitarian or social exchange terms and become instrumental to achieving personal goals. Consequently, the self is defined as an entity separate and independent from groups; concern for and emotional attachment to groups is low. Social behavior is determined more by personal attitudes than by social norms; hedonism, competition, and justice are highlighted to secure personal interest. Such an independent view of the self is more prevalent in cultures that embrace individualism, such as North American and Western European cultures (Triandis, 1989).  34  On the contrary, an interdependent view that defines the self in terms of relationships and social connections is more common in collectivist cultures, such as most Aboriginal, Asian, African, and Latin American cultures. Collectivist cultures tend to embrace a more holistic worldview that sees individuals as part of natural, ancestral, and spiritual world. Self is thus defined as part of group(s), and emotional attachment to groups is high. Interpersonal and social relationships can be described as communal. Group goals are given priority over personal goals. Duty, obligation, cooperation, social norms, and sociability are emphasized to ensure the integrity and wellbeing of the groups, especially family integrity. In contrast to the independent view, the interdependent construal of self reflects a holistic, collectivistic worldview that values cooperation and sharing of resources rather than competition among individuals, thus attesting to group-level functionalism. These differences in self-view can have differential effect on individual experience, including cognition, emotion, and motivation (Markus & Kitayama, 1991). Triandis (1996) further distinguished between vertical (emphasizing hierarchy, power, and achievement) and horizontal (emphasizing equality, benevolence, and universalism) individualism (stressing self-direction, hedonism, and distance from ingroup) and collectivism (stressing conformity, security, and closeness to ingroup). Triandis and Gelfand (1998) reported that horizontal individualism (HI; denoted by lateral relation to authority and individualist relation to group, as found in Australia and Sweden) was predicted by self-reliance only, and vertical individualism (VI; denoted by linear relation to authority and individualist relation to group, as found in France and the United States) by both competition and hedonism. Whereas, both associated with sociability, vertical collectivism (VC; denoted by linear relation to authority and collectivist relation to group, as found in China and Japan) and horizontal collectivism (HC;  35  lateral relation to authority and collectivist relation to group, as found in the Israeli kibbutz) were not very distinct, except that VC was predicted by family integrity, and HC by interdependence. Anyway, these four dimensions (VI, HI, VC, and HC) can be distinguished by different sets of values and measured by different sets of items (see Appendix A for items of the Measurement of Horizontal and Vertical Individualism and Collectivism; Triandis & Gelfand, 1998). Despite a contrast between independent and interdependent self-construals (Markus & Kitayama, 1991; Triandis, 1989), these two self-views are not necessarily polar opposites or mutually exclusive (Singelis, 1994). Many theorists have advocated a balanced model of development to denote mature personality and development and envisioned a balanced self as scoring high on both independence and interdependence (e.g., Guisinger & Blatt, 1994; Imamoglou, 1998; Yamada & Singelis, 1999), echoing the holistic and complementary view. Similarly, although research indicated that individualism and collectivism appear to be polar opposites of a single dimension at the cultural level, they are not mutually exclusive at the individual level (Singelis, Triandis, Bhawuk, & Gelfand, 1995). Many theorists have also argued that these concepts are mutually dependent and coexist in cultures and individuals (e.g., Sinha & Tripathi, 1994; Shimizu, 2000) and that dichotomous theorizing can undermine the interdependent and complementary nature of the opposites (e.g., Guisinger & Blatt, 1994; Ho & Chiu, 1994). To avoid polarized theorizing, I adopted a holistic and complementary approach herein that argues for the multiplicities of individualism and collectivism beyond dichotomization and their coexistence within the same individual. Turiel (2004) also contended that the notion of culture may serve to stereotype people and groups and that the prevailing individualism-collectivism distinction fails to reflect the diversity of worldview and self-view, especially in present pluralistic societies. I thus proposed  36  to use the term ―worldview‖ in lieu of the terms of cultural individualism-collectivism to represent the various dimensions or combinations of individualism and collectivism. Cultural versus personal worldview. People‘s worldviews may be influenced by numerous sources, such as parental worldviews, religious beliefs, education, and popular cultures. These sources may convey conflicting messages. Moreover, many of our cultural truisms are contradictory to reflect the dualistic realities of life (Darley, 1995). When the observed facts fail to follow logically from their conception of reality, some individuals may choose to discount them, but some may change or reorganize their worldviews accordingly. Either way, individuals must decide for themselves what kind(s) of worldview to adopt or develop their own philosophies of life, based on their own experiences and personal characteristics (e.g., affiliation, achievement needs). It is here where the term ―personal worldview‖ comes in to keep the cultural and individual levels of analysis distinct. Personal worldview, which is a product of the reciprocal interplay of personal and social influences, provides a framework through which people perceive their world and conduct their life, moral or otherwise (cf. Dien, 1997). It is possible that some individuals may have difficulty developing or holding on to a coherent and consistent worldview, because of the inconsistency they have experienced (e.g., what is preached is different from what is done by the adults, conflicts between perceived reality and one‘s worldview). Without a coherent and consistent worldview, one may feel lost, not knowing what decision to make and not knowing what to do. Coherent or not, our worldview will influence our self-concept and our relationships with others (M. Harris, Fontana, & Dowds, 1977; Markus & Kitayama, 1991; L. J. Myers et al., 1991). After all, kindness and cruelty are expressions of and are affected by one‘s image of the world, image of the self, and the conception of the self-other relationships, as summed up by Staub (1984).  37  Worldview and Moral Inclusiveness As the root of moral development, worldview was operationalized in this study as the various types of individualism and collectivism because they have moral meanings that match well with the seven levels of moral inclusiveness discussed previously. Individualism and collectivism are defined as the pursuit of self-interest and the pursuit of collective interest, respectively (Hui & Triandis, 1986), notwithstanding their multiple forms. Hogan (1975) stated that there are different types of individualism, representing different philosophical positions with regard to basic human nature. For example, ideological individualism sees social convention as artificial and fake, and alienated individualism sees social institutions as evanescent and people who identify with their social roles as inauthentic. Romantic individualism sees people as naturally good and society as corrupted, but egoistic individualism sees people as fundamentally selfish, self-centered, and aggressive. Although romantic individualism paints a more positive view of human nature, individualism in general seems to denote a negative worldview that portraits the social world as corrupted and untrustworthy. Such a pessimistic worldview reinforces the belief of an immoral world as imparted by Kohlberg and Ryncarz (1990), rendering morality impracticable and selfishness viable. With a negative worldview, excluding others from one‘s moral concern becomes proper. Sampson (1977, 1988) also distinguished two forms of individualism, representing opposing views of self in relation to others: ensembled individualism (views self as a free, independent agent, but self cannot be defined apart from others) and self-contained individualism (views self as distinct from others, and self is defined by an exclusion of others). Ryckman, Hammer, Kaczor, and Gold (1996) associated the former with personal development competition (the focus of competition is not on winning, but to facilitate personal growth) and the latter with  38  hypercompetition (an indiscriminate need to compete and win at all cost in order to feel good about the self). Self-contained and egoistic individualism may be more akin to vertical individualism, all emphasizing competition. Combining hypercompetition with vertical individualism paints a picture of Social Darwinism, which is likely to promote independence and egoistic orientation, aggressiveness and illegality, as described by Jankowski (1991). In contrast, ensembled and romantic individualism may be more similar to horizontal individualism, which is more likely than other forms of individualism to foster the development of a healthy independent self-view that may facilitate the development of interdependent self. Either way, individualism is likely to lead individuals to focus on the first level of moral inclusiveness that includes no other but the self, unless it is balanced with collectivism. There can be several kinds of collectivism as well (Uleman, Rhee, Bardoliwalla, Semin, & Toyama, 2000). Hui and Triandis (1986) deduced that collectivism could be summed up by the word ―concern,‖ which refers to bonds and a sense of oneness with others, due to the recognition that the basic unit of survival is the group and not the individual. Different kinds of collectivism thus represent different philosophical positions with regard to the nature of life or the basic unit of survival. For example, familism views family as the center of life (i.e., life is passed on through family linkages); sociocentrism or ethnocentrism views one‘s society or ethnic group as the center of life. Anthropocentrism sees humans as the center of the universe, biocentrism views the universe as the creator of life (Campbell, 1983), and religious or spiritual worldview views God as the creator of life, spirit as fundamental, and matter as derivative (Smith, 2001). Different levels of collectivism can thus coincide with different levels (Levels 2 to 7) of moral inclusiveness.  39  Moral Inclusiveness as Relationship Closeness Moral inclusiveness is not a new construct. Cooper (2003) argues that defining moral development as progress toward greater moral inclusiveness is central to William James‘s (1897) moral theory. Research has supported that individuals with a highly self-important moral identity are more likely to expand their circles of moral inclusiveness to include out-group members and have more favorable attitudes toward relief efforts to aid out-group members (Reed & Aquino, 2003). However, according to the self-expansion model of Aron and Aron (1986), the primary motive for inclusion of others in the self (IOS) is self-expansion, so that one could gain access to others‘ resources, identity, and perspectives when the others are included in the self. Research has indicated that, even though the initial motivation to include others in the self could be selfserving, the process of IOS eventually leads to a motivation to benefit the other along with the self, due to cognitive confusion of self and others in close or new relationships (Aron, Aron, Tudor, & Nelson, 1991; Cialdini, Brown, Lewis, Luc, & Neuberg, 1997; Maner et al., 2002). To investigate the IOS process, Aron, Aron, and Smollan (1992) constructed the Inclusion of Other in the Self Scale (IOS Scale) to tap people‘s sense of being interconnected with another. The IOS Scale is a single-item pictorial measure of relationship closeness. In the IOS Scale, respondents are asked to select the picture that best describes their relationship from a set of Venn-like diagrams each representing different degrees of overlap of two circles. The theoretical approach of Aron et al. (1992) for understanding and assessing relationship closeness has gathered much support (Agnew, Loving, Le, & Goodfriend, 2004). For example, Mashek, Cannaday, and Tangney (2007) developed a single-item pictorial measure of community connectedness based on the IOS Scale. In two studies, one with college students and one with incarcerated male offenders, Mashek et al. found the single-item pictorial measure  40  correlated positively with lengthy multi-item measures of psychological sense of community and self-report community helping activities, but negatively with self-report anti-social behavior. Tropp and Wright (2001) adapted the IOS Scale to produce the Inclusion of Ingroup in the Self Scale to study ingroup identification. They found cognitive confusion of self and ingroup among people high in ingroup identification (i.e., the ingroup is adequately included in the self-concept), but not in those low in ingroup identification. Moreover, Mashek, Aron, and Boncimino (2003) reported that the greater confusions between self and close other are due specifically to interpersonal closeness rather than greater familiarity or similarity with close other. As a holistic process of cognitively including another person within one‘s self-concept (Aron, Mashek, & Aron, 2004), relationship closeness seems to capture well the concept of moral inclusiveness for relationship closeness can lead to cognitive confusion between self and other that spontaneously activates an altruistic orientation (Aron et al., 1991; Cialdini et al., 1997; Maner et al., 2002). Thus, moral inclusiveness was operationalized as relationship closeness. The overlapping circles representing self and other can best portray the degree of moral inclusiveness at each level (see Appendix B). Thus, relationship closeness is also viewed as a continuum from completely detached to completely overlapped, reflecting the continuous degree of inclusiveness from zero to 100%. The degree of moral inclusiveness at each level, in conjunction with worldview and self-view, affects the development of moral orientation. Moral Inclusiveness and Moral Orientation To investigate the multiple moral orientations reflecting the seven levels of moral inclusiveness, a measure of multidimensional moral orientation is required. It was the second objective of this study to develop such a measure, called Moral Orientation Index (MOI), by revising and expanding the Objective Measure of Dispositional Moral Orientation (Yu, 2001) to  41  include various moral orientations beyond the Kohlbergian schemes. The MOI assesses how people generally solve problems or make decisions on a continuum. Table 1 lists the definitions of and items for the moral orientation subscales of the MOI at each level of moral inclusiveness. Table 1 Definitions of and Items for the Moral Orientation Index (MOI) Subscales  Level 1. M32. M20. M33. M1. M14. M26. M36. M30. M25. Level 2.  Egocentric Orientation (Decisions and reciprocity are based on self-serving purposes) In solving problems, I try to protect my own benefits first. I make decisions based on my own needs and desires. I don‘t waste my time and effort on something that‘s not going to benefit me. I do what I feel is best for me, regardless of how that might affect others. When I do people a favor, I expect them to return a favor in the future. If people are mean to me, I‘ll be mean to them. My benefits and goals are more important than my family‘s benefits and goals. I prefer to forgive than to revenge. (Reversed) In making decisions, I weigh out the cost and benefit. Family Orientation (Family responsibility is based on duty rather than good relationships; family integrity is highly valued.)  M8. Family relationships are more important than friendships. M10. Responsibilities to my family are more important than my own rights. M4. I try not to let my family down in whatever I do. M23. When I make decisions, I take my family‘s opinions and benefits into consideration. M29. Responsibilities to one‘s family are more important than duties to the society. Level 3. Care Orientation (Based on an assumption of connectedness, good behavior is defined in terms of getting social approval and maintaining good relationships.) M6. I try to make everyone happy when making decisions. M7. When my desires conflict with my friends‘ benefits, I rather give up my desires than to hurt my friend. M37. I am willing to agree in order to have good relationships with my friends and family. M24. I try to avoid making decisions that may damage my relationships. M3. I am willing to forgive others who have done me wrong. M5. I worry about what others may think of me when I make decisions.  42  Table 1 (Continued) Definitions of and Items for the Moral Orientation Index (MOI) Subscales  Level 4.  Norm (Norm-Maintaining) Orientation (Social responsibility implies maintaining the given social order and upholding the existing laws and regulations.)  M39. I try to follow the existing rules and laws of the society. M13. Students should obey their teachers. M31. We must comply with all the laws, although some of them may seem unnecessary or unreasonable at times. M28. We must maintain social order to protect personal and public safety. M21. Workers and soldiers must show respect to their superiors. Level 5. Justice Orientation (Morality is guided by a justice moral principle that treats all human beings as equals and that goes beyond race and social boundary.) M19. All individuals are equal, regardless of sexual orientation. M38. All individuals are equal, regardless of religious or spiritual beliefs. M12. All individuals are equal, regardless of race. M18. In resolving conflicts, each person‘s rights are equally important. M35. For the sake of national security, sometimes it is impossible to treat everybody equally and fairly. (Reversed) Level 6.  Biocentric Orientation (Moral consideration is extended to include all life forms and the earth.)  M27. Taking care of our natural environment is more important than following the law. M17. Protecting the natural environment is our social responsibility. M2. Preserving nature is more important than expanding business and consumption. M9. The law of nature is more fair and just than the law of any human society. M11. All life forms, including humans, animals, and plants are equally valuable. M15. Animals have a right to live just as much as humans do. Level 7.  Religious/Spiritual Orientation (Morality is guided by one‘s religious or spiritual beliefs.)  M22. I make decisions based on my religious or spiritual beliefs. M34. In resolving conflicts, I rely on the teachings of my religion or spiritual beliefs. M16. Obeying God or other higher being(s) is more important than obeying one‘s parents or manmade laws. M40. I pray to God or other higher being(s) for guidance when solving problems.  The items of all subscales were randomly ordered as a single scale. Participants are asked how they generally solve problems or make decisions, to rate each sentence on a scale from 1 (Strongly Disagree) to 7 (Strongly Agree) to indicate how much they agree or disagree with the  43  sentence, and to circle the number that best matches their answer. Visual representations of the 7-point rating scale were provided to reduce boredom (see Appendix C). As shown in Table 1, for Level 1 inclusiveness that contains only the self, an Egocentric Orientation is most compatible because decisions and reciprocity are based on self-serving purposes. This first level of inclusiveness and moral orientation coincide with Kohlberg‘s (1971) preconventional morality. Because self-interest is now an existing norm in the Western cultures (D. T. Miller, 1999), it is no longer preconventional, but conventional. Rather, I call it egocentric orientation. It may have become a leading orientation worldwide, thanks to the widespread influence of the individualist Western cultures all over the world through trades and the media. To reflect Level 2 inclusiveness that centers on family relationships, a Family Orientation or Filial Piety (Dien, 1982) is ideal. Filial piety, which dictates that it is the duty of children to care for their (grand) parents physically and psychologically, is accepted as consistent with the law of nature in most Eastern cultures because our life came from our parental lines. Thus, filial piety or duty-based familism is different from social approval reasoning. Because it is consistent with the law of nature, family orientation is not a property of the East, but a universal concept that can be found in all cultures. My notion of family orientation thus differs from the Personal Interest schema of Rest et al. (1999) or the social approval or good-boy/nice-girl orientation of Kohlberg (1981). Specifically, a family orientation values family integrity and family responsibility, based on duty rather than good relationships. For Level 3 inclusiveness that focuses on friendships and peer relationships, I chose Gilligan‘s (1982) notion of care rather than Kohlberg‘s (1981) social approval or good-boy/nicegirl orientation as the Level 3 orientation. It is because the former is based on an assumption of connectedness and sympathy (Gilligan, 1982), whereas the latter is rooted in self-interest (Rest et  44  al., 1999). That is, the current Care Orientation defines good behavior in terms of getting social approval and maintaining good relationships for the sake of relationships rather than self-interest. At Level 4, one‘s moral inclusiveness expands from one‘s peer groups to one‘s school or workplace, the local community, and one‘s nation. To preserve these social institutions, individual members must uphold the respected norms, authorities, and social orders of these social institutions. Therefore, a Norm Orientation that views social responsibility as maintaining the given social order and upholding the existing laws and regulations seems indispensable for Level 4 inclusiveness. Although norm orientation may be more advanced than family and care orientations because it denotes the attainment of a sociocentric perspective (Rest et al., 1999), national conflicts and ethnic cleansing can be justified as patriotic acts to preserve group norms and social order. Similarly, cruelty to fellow human beings can also be committed at lower levels of inclusiveness for the sake of protecting individual (Level 1) or ingroup interests (Levels 2-3). For Level 5 inclusiveness that includes the human race, a Justice Orientation that treats all human beings as equals beyond race and social boundary is most appropriate. Justice orientation is more adequate and morally advanced than lower level orientations for it can prevent cruelty to fellow human beings, by treating all human beings equally. However, individuals who embrace humanity as a whole may have not yet grasped the interconnectedness of all beings and fail to consider the earth and other animals when making moral decisions. When human beings are seen as the center of life, individuals at Level 5 can justify their exploitation of the environment and other animals on anthropocentric terms despite their justice orientation. Even Kohlberg (1984) once argued for ―justified killing‖ when his 4-year-old son refused to eat meat for six months because it is bad to kill animals (pp. 14-15).  45  To reflect Level 6 inclusiveness that includes all life forms and nature, a Biocentric Orientation that extends moral consideration to all life forms and the earth is most suitable. As argued above, justified killing and abuse of animals may occur at Level 5 inclusiveness for the benefit of humanity, despite justice orientation. The occurrence of cruelty and violence against animals, however, can lead to a slippery road of spreading justified killing of and cruelty toward not only animals but also human beings by degrading the victims to subhuman objects or low animal forms. Only by honoring Level 6 inclusiveness and biocentric orientation that advocate the humanization of all beings, can cruelty and violence be minimized if not eliminated from humanity. Therefore, biocentric orientation is more satisfactory and morally advanced than lower level orientations, including justice orientation. For Level 7 inclusiveness, I chose a Religious/Spiritual Orientation (i.e., morality is guided by one‘s religious or spiritual beliefs), because it is commonly practiced by lay people as well as moral exemplars (Colby & Damon, 1995; L. J. Walker et al., 1995). Referring to institutions and systems of practices and beliefs within which a social group engage, religion can provide a platform for the expression of spirituality, but can also inhibit individual spirituality (Oser, Scarlett, & Bucher, 2006). A religious orientation may risk being biased toward religiosity and fails to represent the kinds of orientation that are not based on religion but are important for Level 7 inclusiveness, such as universal love or compassion. Without embracing biocentric and justice orientations, individuals may base their moral judgment on divine command theory or religious orientation only, which is just a form of conventional morality (i.e., upholding the norms of a religious institution). Some may take the Bible literally and assume that humans should have dominion over the earth and other animals; some may even kill people or wage war in the name of their religion. Thus, religious orientation by itself cannot be considered as more  46  advanced than other orientations. Religious or spiritual development cannot be understood by spiritual worldview, religious self-view, closeness to God, and religious orientation alone, but in relations to other moral orientations, particularly egocentric, justice, and biocentric orientations. To sum up, moral orientation is informed by moral inclusiveness, as well as worldview and self-view, as suggested by the proposed model of moral development presented in Figure 2. The causal relations among worldview, self-view, moral inclusiveness, and moral orientation were the focus of this study and the core of the main hypothesis. Hypothesis There is one main hypothesis in the current study. It is best stated by the hypothesized structural model depicted in Figure 3, which represents the first part of the comprehensive model of moral development. Based on the above literature review, the structural model posits that worldview shapes self-view and moral inclusiveness, and then worldview, self-view, and moral inclusiveness jointly influence moral orientation. Thus, the hypothesis can be ramified in terms of five individual path hypotheses (see Figure 3):  Figure 3. Hypothesized structural model Path 1: Path 1 illustrates the hypothesis that worldview shapes self-view. This hypothesis was based on literature supporting the influence of worldview, such as individualism  47  and collectivism, on the formation of independent and interdependent self-concept (e.g., Markus & Kitayama, 1991; L. J. Myers et al., 1991; Triandis, 1989). Path 2: Path 2 shows the hypothesis that worldview influences moral inclusiveness. This hypothesis was based on research on how our theory of the world affects our interpretation of the moral meaning of social facts and values (e.g., Dien, 1997; Jankowski, 1991; Wainryb, 2004), which in turn defines for us what is moral (Wainryb, 2004) and therefore who should be included in our moral inclusiveness. Path 3: Path 3 presents the hypothesis that worldview shapes moral orientation. This hypothesis was based on research that supported the influence of worldview on moral judgment and action (e.g., Jankowski, 1991; Wainryb, 2004) and the current definition of moral judgment as situational moral orientation in this study. Path 4: Path 4 illustrates the hypothesis that self-view induces moral orientation. This hypothesis was based on research supporting correlations between self-view and moral orientation, such as the positive correlations between interdependent self and care orientation (Gilligan, 1982; Lyons, 1983). Path 5: Path 5 represents the hypothesis that moral inclusiveness influences moral orientation. This hypothesis was based on Carter‘s (1980) theory that moral stages or orientations are stages of moral inclusiveness. As defined in this study, there are seven levels of inclusiveness, reflecting seven sets of worldview, self-view, moral inclusiveness, and moral orientation, one at each level of inclusiveness. The hypothesized structural model needed to be tested seven times, each with a different set of observed variables for worldview, self-view, moral inclusiveness, and moral orientation at each level of inclusiveness. As a result, the main hypothesis (the hypothesized  48  structural model in Figure 3) is comprised of seven subsidiary models, one at each of the seven levels of inclusiveness. That is, the constructs of worldview, self-view, moral inclusiveness, and moral orientation were all multidimensional and had different aspects at different levels of inclusiveness. Each aspect was measured with a unique set of observed variables. The different aspects of worldview, self-view, moral inclusiveness, and moral orientation at each level are presented in Table 2. Table 2 Different Aspects of Worldview, Self-View, Moral Inclusiveness, and Moral Orientation at Different Levels of Inclusiveness. Levels of Inclusiveness  Worldview  Self-View  Moral Inclusiveness  Moral Orientation  Level 1  Vertical Individualism & Social Darwinism  Independent Self  Self (i.e., no relationship closeness to any groups)  Egocentric Orientation  Level 2  Vertical Collectivism  Interdependent Self  Closeness to family  Family Orientation  Level 3  Horizontal Collectivism  Interdependent Self  Closeness to friends (and peers)  Care Orientation  Level 4  Horizontal Collectivism & Vertical Worldview  Interdependent Self  Closeness to society (school, local community)  Norm Orientation  Level 5  Collectivism & Horizontal Worldview  Moral Self  Closeness to humanity, indicated and assessed by anti-nationalism  Justice Orientation  Level 6  Collectivism, anti-Social Darwinism & antiAnthropocentrism  Ecological Self  Closeness to nature  Biocentric Orientation  Level 7  Spiritual Worldview  Religious Self  Closeness to God  Religious Orientation  As illustrated in Table 2, the various forms of individualism and collectivism can be put together to form worldviews and self-views that coincide with the seven levels of moral  49  inclusiveness and orientation. As the content of inclusiveness expands from the individual (the self) to collectives (groups or nature), one‘s moral orientation and judgment are likely to mature (Carter, 1980). Thus, worldview, self-view, moral inclusiveness, and moral orientation at the same level are interrelated as described by the five individual path hypotheses. In the following, I present specific hypotheses on each subsidiary model, followed by the supporting rationale. To signify which individual path hypothesis the text is referred to, the path hypothesis number is inserted in a pair of brackets [e.g., Path 1] at the end of the relevant text. Note that all the paths in Figure 3 are unidirectional reflecting the hypothesized pseudo causal or predictive relations between the latent constructs. Hence, hypotheses at each level were stated in predictive terms that match the direction of the arrows of the paths in Figure 3. Level 1 It is hypothesized that at Level 1, both vertical individualism and social Darwinism positively predict independent self [Path 1] and egocentric orientation [Path 3]. Independent self also positively predicts egocentric orientation [Path 4]. However, no relationship closeness to any groups have positive correlations with vertical individualism, social Darwinism [Path 2], and egocentric orientation [Path 5]. Rationale. Individualism has been seen as the breeding ground of independent self-view (Markus & Kitayama, 1991), which can be problematic for it fails to meet the basic human needs of connectedness (e.g., Maslow, 1954). In the absence of traditional values (e.g., family, community), the self becomes the value base (Baumeister, 1997), which leads to moral decay and the rise of crime and violence (Dalrymple, 2001). According to the holistic and complementary viewpoint, however, neither individualism nor the independent self is to blame for delay in moral development, because they can be  50  balanced by collectivism and interdependence, but the exclusion of others in one‘s moral inclusiveness, which is a sign of extremism. As argued previously, horizontal individualism may foster the development of a healthy independent self-view. Although independence may give rise to egocentric orientation, it can also serve as a secure base, enabling the individual to bond with others and develop interdependence. Thus, independence is also valued in collectivist cultures. To be a constructive member of rather than a burden on the group, one must be self-sufficient or independent. This kind of self-reliance and independence is essential for both individual and group survival and need not involve competition. A moderate independent self and egocentrism need not interfere with moral development, if they are balanced by interdependence and inclusion of others in one‘s moral inclusiveness. Only when independence mixes with vertical individualism or social Darwinism, the outcome becomes drastic. With the belief that the individual (the self) is the basic unit of survival, individuals may become highly competitive and they tend to see others as competitors for resources and survival, reflecting social Darwinism or extreme vertical individualism. With an extreme individualist worldview that stresses separation and competition between individuals, individuals are likely to adopt a strong independent self-view to the exclusion of others, which may predispose them to focus on competition and to see others, including siblings and spouse as competitors and to act competitively and aggressively rather than cooperatively. Thus, family conflicts and sibling rivalry are more common in individualist than in collectivist cultures (Triandis, 1989).These individuals are less likely to develop relationship closeness with other people or groups, because interpersonal and social relationships are based on utilitarian or social exchange motives (Markus & Kitayama, 1991). Therefore, individuals embracing vertical  51  individualism and social Darwinism are likely to espouse independent self and egocentric orientation, but have little relationship closeness with other people or groups. Level 2 At Level 2, vertical collectivism is hypothesized to positively predict interdependent self [Path 1], relationship closeness to family [Path 2], and family orientation [Path 3]. Interdependent self and relationship closeness to family also positively predict family orientation [Path 4 and Path 5]. Rationale. Most children were born or brought into a family without their say. Most family structures are characterized by vertical parent-child relationships, reflecting a world of vertical collectivism, in which children are raised and expected to learn and conform to all the rules and facts of life, including a family orientation. Thus, vertical collectivism and parent-child relationships are the kinds of social world most children first experience in life, with little choice. Growing up in a collectivist environment, children are likely to see the self and others as interdependent beings, especially when they are still dependent on the others to satisfy their needs. Children‘s relationship closeness with their family and their interdependent self-concept could help the internalization of family orientation. Family membership is usually determined by biological relationships bonding individual family members. However binding, family relationships are usually full of conflicts due to close proximity, so exercising family orientation (to sacrifice self-interest for the family when making decisions) to maintain or mend such unbreakable family relationships becomes wise and even necessary at times. In turn, the practice of family orientation can also help enhance family relationships and foster vertical collectivism. Thus, individuals who embrace vertical collectivism are likely to embrace interdependent self and family orientation and have close  52  relationships with their family. Their closeness to family and interdependence further enhance their family orientation. Level 3 At Level 3, horizontal collectivism is hypothesized to predict positively interdependent self [Path 1], relationship closeness to friends and peers [Path 2], and care orientation [Path 3]. Interdependent self and relationship closeness to friends also predict positively care orientation [Paths 4 and 5]. Rationale. One‘s relationships with peers are likely to reflect horizontal collectivism that stresses connectedness and communality, because all members are similar in age and social status. The elements of connectedness and communality of peer relationships in turn promote a sense of interdependence among its members. Unlike family memberships that are bound by biological links, individuals can enter a relationship with their peers voluntarily. The formation and maintenance of friendships and peer-relationships, however, require give and take, sharing and caring, and therefore a care orientation. Consequently, individuals who embrace horizontal collectivism are likely to develop a strong interdependent self, have close relationships with friends and peers, and adopt care orientation. Their closeness to friends and interdependent selfview further fortify their care orientation. Level 4 Specific hypotheses at Level 4 state that horizontal collectivism and vertical worldview positively predict interdependent self [Path 1], relationship closeness to society [Path 2], and norm-maintaining orientation [Path 3]. Interdependent self and relationship closeness to society also positively predict norm-maintaining orientation [Path 4 and Path 5].  53  Rationale. Social institutions, such as schools, the local community, and so on are places where individuals come together for a particular purpose and where social relationships are often formed. Because most social institutions are vertical in structure but somewhat egalitarian among their members, individuals‘ experiences with social institutions can create in them a vertical worldview combined with horizontal collectivism. Combining vertical worldview with horizontal collectivism, worldview at Level 4 reflects essentially a combination of worldview elements at the first three levels, namely, vertical individualism at Level 1, vertical collectivism at Level 2, and horizontal collectivism at Level 3. To learn, to work, and to play as teams in school promotes interdependence and closeness to the school. Relationship closeness to the school will eventually spread out to other social institutions and the whole society, due to their similarities in vertical structures and communal spirit. Interdependence and closeness to school encourage the students to respect authorities (e.g., teachers) and uphold school regulations (norm-maintaining orientation). Therefore, at Level 4 inclusiveness, individuals are likely to espouse a vertical worldview that emphasizes hierarchy and power, in addition to horizontal collectivism, develop a strong sense of interdependence, feel close to their school and the society, and adopt a norm-maintaining orientation that emphasizes respect for authority and social order. Their closeness to society and interdependent self-view strengthen their norm orientation further. A caveat. If the school (or any social institutions or the society as a whole) creates vertical individualism instead of vertical collectivism, by turning a blind eye to bullying and discrimination or overemphasizing competition and personal achievement, egocentric rather than norm orientation will flourish. Likewise, pathological socialization experiences, in the context of the family or peer experiences can also create a social Darwinist worldview in the individuals  54  (Crick & Dodge, 1994; Peterson & Flanders, 2005). For those who perceive the world around them as a cruel and callous place, independent self becomes the default self-view and egocentric orientation the only survival option. Thus, the hypotheses stated in this section are all under the assumption of a normal and positive environment at all levels, except Level 1. A highly competitive, individualist culture, as those in the West, where self-interest is the norm (Miller, 1999) may be a favorable environment for the development of egocentric orientation, but hardly a positive environment for moral development. Although individuals in such cultures are more likely to accept individualism in general and vertical individualism in particular as the reality and to adopt the independent self-view and egocentric orientation, their degrees of endorsing egocentric orientation may be moderated by the proximal environments they are in (e.g., family, peers, school). Thus, individual development and functioning at any level is contingent on the multi-level person-environment interaction, though I focus my discussion on a level-by-level basis. Level 5 It is hypothesized that at Level 5, higher collectivism and horizontal worldview predict higher moral self [Path 1] and justice orientation [Path 3], but not so much on closeness to nation [Path 2], which represents nationalism. Moral self in turn positively predicts justice orientation [Path 4], but closeness to nation negatively predicts justice orientation [Path 5]. Rationale. I assumed that justice orientation, like care orientation, evolves from horizontal peer relationships. While care orientation focuses on maintaining interpersonal relationships, justice orientation tries to step outside of interpersonal relationships to treat all human beings as equals, representing a horizontal worldview (horizontal collectivism and horizontal individualism). It will also be difficult for those who are kind to all human beings  55  (representing horizontal collectivism) but be mean to their own families (representing vertical collectivism). A balanced worldview that is horizontal (horizontal individualism and horizontal collectivism) and collectivist (horizontal collectivism and vertical collectivism) that considers both individual rights and the collective good is therefore regarded as Level 5 worldview, which is conducive to the development of justice orientation. Embracing the whole human race rather than a small group (e.g., one‘s nation), individuals at Level 5 must be able to detach themselves from their family and friends when making decisions. It requires an intrinsic moral self rather than independent or interdependent selves to enable them to keep a safe distance from close relationships (to be objective) and adhere to justice orientation. They rely on the strength of their moral self to carry out this insistence on doing the right thing, without regard to self-interest and opposition from friends and family (for a review on moral identity see Hardy & Carlo, 2005). Although collectivism is conducive to closeness to nation, nationalism is not a favorable condition for justice orientation. Therefore, at Level 5, individuals are likely to espouse collectivism and horizontal worldview, develop a strong sense of moral self, adopt a justice orientation, but feel little closeness to their nation, reflecting a sentiment of anti-nationalism or some doubts about nationalism. A strong sense of moral self in turn motivates a strong stance on justice orientation, whereas feeling too close to one‘s nation may deter the development of justice orientation. Level 6 Specific hypotheses at Level 6 state that collectivism, anti-anthropocentrism, and antisocial Darwinism positively predict ecological self [Path 1], relationship closeness to nature  56  [Path 2], and biocentric orientation [Path 3]. In turn, ecological self and relationship closeness to nature positively predict biocentric orientation [Path 4 and Path 5]. Rationale. A biocentric view can be derived philosophically (e.g., Taoism; Jenkins, 2002) or scientifically. Taoism regards effort to challenge, exploit, or control people or things as against the law of nature or the Tao (Lao Tzu, 1972). Similarly, many Americans attempt to make sense of the environmental issues by the beliefs that nature is a limited resource and that human domination over nature can upset the balance of nature (Kempton, Boster, & Hartley, 1995). A Taoist or a scientific biocentric worldview, however, may seem oppositional to some religious beliefs regarding the origin of life and can be difficult to describe or comprehend. Built on Level 5 inclusiveness, Level 6 inclusiveness includes other species and mother nature into moral consideration. Thus, Level 6 worldview should incorporate some of the basic elements of Level 5 worldview. I chose collectivism rather than horizontal worldview, because humans‘ relationship with nature is more of a vertical mother-child relationship rather than horizontal peer relations. We call nature ‗mother nature,‘ because nature nurtures our life by providing us land and food. The thought of human supremacy over the earth or other species is antithetical to a Level 6 worldview. I thus proposed that, when Level 5 inclusiveness expands to Level 6, individuals who endorse collectivism are likely to see that humans are not the center of the universe (i.e., antianthropocentrism) and realize that humans must live peacefully and cooperatively with one another and with other species (i.e., anti-vertical individualism or anti-social Darwinism). Taking such a biocentric Level 5 worldview (i.e., collectivism, anti-anthropocentrism, and anti-social Darwinism), individuals are likely to feel close to nature and even see their selves as embedded in nature (e.g., ecological self; Ingalsbee, 1996; Opotow, 2003). Together, Level 5 worldview,  57  concept of ecological self, and closeness to nature promote a biocentric orientation that makes protecting the natural environment a moral responsibility. Level 7 It is hypothesized that at Level 7, spiritual worldview positively predicts religious self [Path 1], relationship closeness to God [Path 2], and religious orientation [Path 3]. Then religious self and closeness to God positively predict religious orientation [Path 4 and Path 5]. Rationale. With regard to religious orientation, I theorized that it begins with a spiritual worldview that introduces individuals to the existence of God, so that they can develop a relationship with God and convert themselves into God‘s followers, thereby cultivating a conception of the self as a religious or spiritual being (i.e., religious or spiritual self, simply called religious self). Together, spiritual worldview, religious self-view, and closeness to God encourage a religious orientation that seeks and follows religious guidance when making moral decisions. It was also expected that spiritual worldview is positively correlated with collectivism, because the ideal inclusiveness of ―all is one‖ at Level 7 refers to everyone, not just the individual and his or her God. In brief, each subsidiary structural model of moral development at Level 1 to Level 7 had different latent variables (different aspects of worldview, self-view, moral inclusiveness, and moral orientation) measured with different indicators, but essentially the same relational structures. Each model was then tested separately. The five individual path hypotheses in the structural model were examined based on the test results from all seven levels. In the next chapter, the methodology used to test the hypothesis is laid out.  58  CHAPTER THREE: METHODOLOGY In this chapter, I describe the research design, the sampling procedures, and data collection procedures. The instruments used for the study are specified and issues in model testing are then discussed. Research Design A survey design was used to discover the relational structures among worldview, selfview, moral inclusiveness, and moral orientation. Self-report questionnaires were used to collect data. The questionnaires were pilot tested and revised accordingly before the final administration to collect the data. There were two samples in this study: students and adults. The relational models were developed using the student sample. The adult sample was used as a comparison sample. The data were mainly quantitative. An alpha level of .05 was used for all statistical tests. Structural Equation Modeling (SEM) was the main data analysis technique, because it allows simultaneous tests of predictive or pseudo causal relationships that are adjusted for measurement errors. Combining confirmatory factor analysis with multiple regressions, SEM can be used to test a theory or a model, in which the constructs are latent variables that cannot be directly measured (Tabachnick & Fidell, 1996), like the variables in this study. Thus, SEM was the most suitable analysis technique for the study. Maximum likelihood estimations were employed to test all models. Post hoc model modifications were performed to develop better fitting models. Acceptable models based on the student data were developmentally referenced to the adult data. Sample Selection Given resources available to the researcher, the study relied on snowball and convenience sampling and voluntary participants. For the pilot sample and student sample, public secondary  59  schools in Western Canada volunteered to participate in the study. The pilot sample and student sample were different groups of students, but both were drawn from Grades 8-12 classes. Students who themselves consented to participate and who had parental permission completed the survey measures. For the adult sample, university students and adults in Western Canada voluntarily completed the survey measures at home and mailed it back to the researchers. The Pilot Sample About 300 students from three public secondary schools in Western Canada were invited to participate in the pilot study. One hundred and ninety-one students who had parental permission volunteered to take part in the survey. Excluding incomplete (e.g., missing data on one-third of a section or more) and questionable protocols (i.e., failing the validity check), the pilot sample included 179 participants, with 56.2% girls, 43.8% boys. The age range was from 13 to 19 years (M = 15.31, SD = 1.34). Over half of the participants (60.9%) were in Grade 10, 22.3% in Grade 8, 0.6% Grade 9, 5.0% Grade 11, and 11.2% Grade 12. The sample included 77.6% Asian and 22.4% other ethnicities (e.g., White, mixed ethnicities, Latin American, Black, and Middle Eastern) and diverse religious beliefs: 37.9% no religion, 20.9% Buddhism, 27.1% Christianity/Catholic, and 14.1% other religious beliefs (e.g., Shinto, Hinduism, and Islam). Participants who circled two or more ethnic groups as their cultural backgrounds were considered as having mixed ethnicities. The High School Student Sample About 1,500 Grades 8-12 students from four public secondary schools in Western Canada were invited to participate in the study. Only 734 students who had parental permission volunteered to answer the survey. Excluding incomplete and questionable protocols, the final student sample included 640 participants. The age range was from 12 to 19 years (M = 14.97, SD  60  = 1.39). As indicated in Table 3, girls (55.3%) and boys (44.8%) were almost equally distributed in each grade. Table 3 Age Means in Years for Gender and Grade for the Student Sample Boys Grade 8 9 10 11 12 Total  n  M  65 71 66 55 28  13.17 14.30 15.10 16.25 17.36  285  14.90  Girls % of Total  Total  n  M  % of Total  n  M  %  10.2 11.1 10.3 8.6 4.5  70 77 80 82 44  13.11 14.23 15.15 16.16 17.14  11.0 12.1 12.5 12.9 6.9  135 148 146 137 72  13.14 14.26 15.13 16.20 17.22  21.2 23.2 22.9 21.5 11.3  44.8  353  15.03  55.3  638  14.97  100  Note. Two cases were excluded because of missing value on gender.  The student sample consisted of multiple ethnic groups: 40.3% White, 40.4% Asian, and 19.3% others (e.g., mixed ethnicities, Latin American, First Nations, Middle Eastern, and Black), as well as diverse religious groups: 36.9% no religion, 40.4% Christianity/Catholic, and 22.7% other religious beliefs (e.g., Buddhism, Sikhism, Hinduism, and Islam). Girls and boys were almost equally distributed in each ethnic group and religious group, as indicated in Tables 4 and 5 respectively. Table 4 Age Means in Years for Gender and Ethnic Groups for the Student Sample Boys Ethnic Groups  n  White Asian Others Total  Girls  M  % of Total N  124 107 49  15.06 14.73 14.97  19.7 17.0 7.8  280  14.92  44.4  n  Total  M  % of Total N  130 148 73  15.01 15.11 14.90  20.6 23.5 11.6  254 255 122  15.04 14.95 14.93  40.3 40.4 19.3  351  15.03  55.6  631  14.98  100  Note. Nine cases were excluded because of missing value on gender or ethnicity.  n  M  % of Total N  61  Table 5 Age Means in Years for Gender and Religious Groups for the Student Sample Boys Religious Groups  Girls  n  M  % of Total  n  M  No Religion Christianity Others  116 102 66  14.83 15.18 14.62  19.4 16.9 8.5  119 156 74  14.99 15.19 14.78  Total  284  14.91  44.8  349  15.04  Total % of Total  n  M  % of Total  20.5 23.0 11.7  235 258 140  14.91 15.19 14.71  39.9 39.9 20.2  55.2  633  14.98  100  Note. Seven cases were excluded because of missing value on gender or religion.  The Adult Sample (Comparison Sample) Close to 1,400 survey forms were distributed to adults in Western Canada. Only 495 participants voluntarily returned their survey response forms. Excluding incomplete and questionable protocols, the final adult sample of 472 included 127 (26.9%) males, 326 (69.1%) females, and 19 (4.0%) participants who did not report their gender. The age of the participants ranged from 16 to 78 years (M = 31.57, SD = 11.86). Almost half of the sample (49.0%) reported having attended or were attending graduate or professional school (mostly in the education faculty); 16.3% had graduated from university; 29.1% had some college/university education; 5.5% high school education; and one missing case. The sample consisted of multiple ethnic groups: 49.7% White, 40.3% Asian, 10.0% other ethnicities (e.g., mixed ethnicities, Latin American, First Nations, Black, and Middle Eastern), and one missing case, as well as multiple religious groups: 41.5% no religion, 40.2% Christianity/Catholic, 18.3% other religious beliefs (e.g., Buddhism, Sikhism, Hinduism, and Islam), and two missing cases. This was not a balanced sample. With its bias towards higher education and female, it may reflect the high end of moral development, providing pertinent information and some advantage in model comparison.  62  Data Collection Procedures The High School Student Sample Eighth to twelfth grade students in BC secondary schools were solicited for participation. Students were initially contacted in their classrooms. They were given a recruitment letter (Appendix D) and asked to take home a parent consent form (Appendix E) and to return a signature page (Appendix F) with their and their parents‘ signatures to the school. Only students who consented and obtained parental consent to participate in the study were allowed to take part in the survey. Participants‘ rights and the confidentiality and anonymity of their answers were emphasized during the whole data collection process. After signature pages were signed and collected, participating students were groupadministered the questionnaires in their classrooms within 2 weeks. Participants typically completed all of the questionnaires in 40-65 minutes. The questionnaires asked about participants‘ background, their relationships with others, their view of the world and themselves, how they generally make decisions, and their feedback on the survey. Students who did not participate were given regular classroom work to do during data collection. The Adult Sample By sending out introductory letters (Appendix G) and emails to contacts (e.g., friends, colleagues, relatives) and potential participants, adults from all walks of life and university (graduate and undergraduate) students in Western Canada were solicited for participation. The questionnaires along with a free pen was placed inside an envelope with a cover letter (Appendix H) stapled on the front of the envelope. The envelopes were then distributed to interested potential participants. The participants completed the questionnaires individually at home. Upon completion, the participants were asked to keep the pen and mail the questionnaires back in the  63  envelope to me. To facilitate the mailing back, an envelope with prepaid postage was provided, except for UBC student participants who could send the sealed envelopes via Campus Mail to me. Measures There were six questionnaires in the survey and they were all pilot-tested with the pilot sample. Based on their feedback, changes were made with regard to the measures used in the present study. The order of the questionnaires was the same for both samples. The first questionnaire was a demographic data sheet and the last questionnaire was a validity check. The rest of the questionnaires were randomly ordered in the pilot study, as follows: the worldview scales, the relationship closeness scale, the self-view scales, and then the moral orientation scale. In the main study, however, the order of the worldview questionnaire and the relationship closeness questionnaire was switched, because the precedence of the relationship closeness items makes worldview and self-view items involving the word ‗group‘ easier to understand. In this section, I present the questionnaires in the order they appeared in the final survey, describe the measure(s) used in each questionnaire, and report the changes that were made due to the pilot study. Owing to these changes, the pilot data was not included in the analysis. Besides the demographic data sheet and the validity check, 10 measures, which involved a total of 25 subscales, were used in this study: two were published measures, five were adapted from published measures, and three were developed for the purpose of the study. Internal consistency reliability was determined by Cronbach‘s (1951) coefficient alpha for all measures. Items on each subscale were averaged to compute a composite score, after reverse scoring items where appropriate.  64  The construct validity for the developed and the adapted measures was established through factor analysis. The factor structures were examined through maximum-likelihood factor analysis with orthogonal (varimax) rotation and a loading of .35 or greater for an item was viewed as meaningful (cf. Comrey & Lee, 1992). Maximum-likelihood factoring was the preferred method because it maximizes the canonical correlations between the factors and the variables. Varimax rotation was chosen because it can simplify factors by maximizing the variance of the loadings with factors across variables and make interpretation easier (Tabachnick & Fidell, 1996). Demographic Data To provide a detailed description of the sample studied, participants were asked to fill out a demographic data sheet. Personal information, such as age, gender, education, ethnicity, and religious affiliation was collected. These data were collected for descriptive purposes only. A full copy of this questionnaire for the student sample and the adult sample can be found in Appendix I and Appendix J, respectively. Moral Inclusiveness Circles of Relationship Closeness. To quantify moral inclusiveness, a Venn diagram Circles of Relationship Closeness (CRC) response scale was used to assess relationship closeness at each level of inclusiveness. The CRC was adapted from the Circles of Closeness Scale (Uleman et al., 2000), which was in turn adapted from the Inclusion of Other in the Self Scale (Aron et al., 1992). Like Aron et al., Uleman et al. used the term closeness rather than interdependence or connectedness to describe interdependence, and considered it a central aspect of collectivism. They employed the spatial metaphor of Venn diagrams to represent one‘s closeness to other people or groups (family, relatives, and friends) to assess different types of  65  interdependence to the various ingroups (e.g., general closeness, emotional closeness, supportive closeness, identity closeness, reputational closeness, similarity to others, and harmony with the group). Items were organized in the scale by closeness type, not by ingroup, but the results of their study indicated three factors with items loading on three ingroups: family, relatives, and friends. Cronbach alpha was high, .91 for the total scale, and .88 to .89 for the three subscales. For the purpose of this study, only emotional closeness, supportive closeness, and identity closeness were used to assess participants‘ relationship closeness with 10 ingroups, including closest family member, family, close friend, peer group, school, local community, nation, nature, the universe, and God or the higher beings. That is, each group had three items. For the adult sample, the ingroup ―school‖ was changed to ―school (or workplace)‖ because it was expected that some adult participants might have left school and be working. Supportive closeness was defined as ―how supportive the group is to you‖ and identity closeness as ―how much you identify yourself as a member of each group.‖ Participants were asked to describe how close they were to the group on a 7-point Likert-type scale from 1 (Completely Detached) to 7 (Concentric) (see Appendix B). Based on feedback from pilot participants, identity closeness was further defined as ―how much you identify yourself (in your self-identity or who you are as a person) as a member of each group‖ in the current study. The 10 groups were intended to represent moral inclusiveness at Levels 2, 3, 4, 5, 6, and 7 respectively. Level 2 inclusiveness is indicated by closeness with family and the closest family member; Level 3 inclusiveness is indicated by closeness with friend and peer group; and Level 4 inclusiveness is indicated by closeness with school (or workplace) and community. Closeness with nation reflects nationalism or ethnocentrism and thus represents anti-Level 5 inclusiveness, because individuals at Level 5 embracing all human beings will not be confined by their  66  nationality or ethnicity. Closeness with nature symbolizes Level 6 inclusiveness. A combination of closeness with the universe and God represents Level 7 inclusiveness, whereas closeness with God alone represents a pseudo or religion-based Level 7 inclusiveness. Reliability and construct validity. To investigate whether the 30 CRC items were tapping relationship closeness at Levels 2, 3, 4, 5, 6, and 7, maximum-likelihood factor analysis (MLFA) was performed. A varimax rotation revealed six factors with items clearly loading on different ingroups, which was supported by a scree test (Cattell, 1988) as well. As can be seen in Table 6, six factors with eigenvalues between 9.64 and 1.27 accounted for 72.19% of the total variance: Factor 1, Closeness to Nature; Factor 2, Closeness to Family; Factor 3, Closeness to Society; Factor 4, Closeness to Friends; Factor 5, Closeness to God; and Factor 6, Closeness to Nation. All relevant items loaded at least .54 on the appropriate factor, and most items had high communalities (> .60). Items C6, C16, and C26 were dropped due to significant cross-loadings on Factor 1 in the student sample, and on Factor 6 in the adult sample. Items of universe merged with items of nature rather than with items of God in both samples, suggesting that Closeness to God represents more of a religion-based Level 7 inclusiveness. As shown in Table 7, each scale had high coefficient alpha reliability (≥ .86), mean interitem correlation (≥ .51), mean item-total scale correlations (≥ .65) and range of item-total scale correlations (.60 - .71 or higher), supporting that each factor was independently meaningful. The moderate factor correlations (≥ .40) in Table 7 should thus be interpreted as a tendency for students who felt close to nature also felt close to society, nation, and God, and so on. The adult sample also had high coefficient alphas (.87 - .96) and similar factor correlation patterns, except that closeness to God had a higher correlation with closeness to nature (r = .41) but lower (or negligible) correlations with other factors in the adults than it did in the students.  67  Table 6 Item Summary and Rotated Factor Loadings for Relationship Closeness for the Student Sample (N = 640) Factor loadings 1  2  3  4  5  6  h2  C19. Support, Universe C18. Support, Nature C29. Identity, Universe C28. Identity, Nature C8. Emotion, Nature C9. Emotion, Universe  .827 .784 .781 .761 .751 .754  .057 .145 .067 .106 .111 .037  .182 .197 .198 .235 .151 .146  .022 .116 .024 .109 .098 -.000  .126 .004 .136 .048 .011 .097  .104 .125 .170 .124 .114 .097  .747 .704 .701 .675 .621 .611  C12. Support, Close Family  .100  .796  .064  .130  .016  .032  .665  C14. Support, Family C2. Emotion, Close Family C22. Identity, Close Family C4. Emotion, Family C24. Identity, Family  .059 .104 .084 .083 .070  .779 .776 .758 .751 .746  .170 .031 .131 .101 .252  .075 .075 .155 .046 .118  .088 .003 .036 .117 .121  .116 .063 .025 .122 .075  .665 .623 .625 .612 .659  C25. Identity, School  .170  .149  .757  .178  .089  .043  .666  C15. Support, School C5. Emotion, School C26. Identity, Community C16. Support, Community C6. Emotion, Community  .168 .184 .362 .371 .354  .191 .153 .113 .118 .175  .731 .650 .655 .619 .552  .162 .177 .120 .129 .098  .099 .101 .040 .009 .006  .081 .123 .165 .235 .181  .641 .536 .617 .607 .503  C11. Support, Friend  .088  .162  .070  .774  .019  -.025  .638  C21. Identity, Friend C1. Emotion, Friend C13. Support, Peer C3. Emotion, Peer C23. Identity, Peer  .053 .076 .050 .027 .051  .179 .126 .015 .065 .013  .029 -.018 .349 .287 .307  .768 .778 .601 .600 .544  .037 -.023 .069 .069 .099  -.049 -.030 .158 .129 .179  .630 .628 .515 .469 .434  C20. Support, God  .097  .099  .103  .080  .941  .149  .943  C10. Emotion, God C30. Identity, God  .129 .101  .121 .096  .064 .099  .056 .060  .926 .909  .130 .153  .914 .883  C7. Emotion, Nation  .207  .177  .156  .068  .180  .754  .704  C17. Support, Nation C27. Identity, Nation  .331 .321  .148 .128  .251 .278  .079 .090  .187 .208  .724 .689  .759 .722  Eigenvalues  9.64  3.50  2.86  2.58  1.80  1.27  32.13  11.68  9.52  8.60  6.02  4.24  No. Item Description  % of variance  Note. Item numbers in bold were dropped. Loadings of .35 and above are in bold. Description of items can be found in Appendix B. Factor 1 = Closeness to Nature; Factor 2 = Closeness to Family; Factor 3 = Closeness to Society; Factor 4 = Closeness to Friends; Factor 5 = Closeness to God; Factor 6 = Closeness to Nation; h2 = communality.  68  Table 7 Factor Score Correlation Matrix (N = 636) and Internal Consistency Reliability Estimates (N = 625 - 635) for Relationship Closeness for the Student Sample Factors Closeness to Nature Closeness to Family Closeness to Society Closeness to Friends Closeness to God Closeness to Nation Mean inter-item correlation Range of item-total scale correlations Mean item-total scale correlation Coefficient Alpha Reliability Note. *p < .05.  1  2  3  4  5  6  .244* .413* .217* .236* .503* .63 .73 - .78 .75 .91  .356* .274* .225* .325* .51 .60 - .71 .65 .86  .391* .245* .430* .69 .73 -.78 .75 .87  .171* .263* .73 .76 - .80 .78 .89  .406* .67 .74 - .82 .78 .93  .91 .93 - .95 .93 .97  Worldview Seven aspects of worldview were assessed. Four were assessed with the Measure of Horizontal and Vertical Individualism and Collectivism (HVIC; Triandis & Gelfand, 1998). Three were assessed with Social Darwinism, Anthropocentrism, and Spiritual Worldview subscales. The items of the three worldview subscales and the HVIC were randomly ordered as a single scale asking about the participants‘ view of the world. Respondents were asked to rate each item on a 7-point Likert-type scale ranging from 1 (Strongly Disagree) to 7 (Strongly Agree) to indicate how much they agreed or disagreed with the statement. Visual representations of the 7-point rating scale the same as those for the Moral Orientation Index (see Appendix C) were provided to make the questionnaire more fun to do. The Measurement of Horizontal and Vertical Individualism and Collectivism. The Measurement of Horizontal and Vertical Individualism and Collectivism scale (HVIC) was first developed by Singelis et al., (1995) to assess horizontal and vertical individualism-collectivism. It was later refined by Triandis and Gelfand (1998) and then by Gelfand (personal  69  communication, January 31, 2006). In this study, the 2006 version of the HVIC was used. The 2006 version of the HVIC had 27 items and four subscales, all with acceptable internal consistency reliability: Horizontal Individualism (HI; 5 items; Cronbach α = .81), Vertical Individualism (VI; 8 items; Cronbach α = .82), Vertical Collectivism (VC; 6 items; Cronbach α = .73), and Horizontal Collectivism (HC; 8 items; Cronbach α = .80). Because the scales were designed for adult samples, words that do not apply to or may be uncommon to secondary students, such as ―coworker‖ and ―aroused‖ were changed to ―group member‖ and ―stirred up‖ respectively (see Appendix A). Based on the analysis for the current study, coefficient alphas for the four factors were also in the acceptable range, .79, .64, .73, and .66 for VI, VC, HC, and HI for the student sample and .79, .67, .73, and .68 for VI, VC, HC, and HI for the adult sample. Items on each subscale were averaged to compute a composite score, after reverse scoring items where appropriate: VI, VC, HC, and HI. Four new composite scores were created by averaging two related composite scores: Vertical Worldview (VER) = Mean (VI, VC), Horizontal Worldview (HOR) = Mean (HI, HC), Collectivism (COL) = Mean (VC, HC), and Individualism (IND) = Mean (VI, HI). Social Darwinism, anthropocentrism, and spiritual worldview. Items for these worldview subscales are listed in Table 8. Four items were used to assess Social Darwinism and they were drawn from the Hypercompetitive Attitude Scale (HCA; Ryckman, Hammer, Kaczor, & Gold, 1990), because they seem to capture well the spirit of social Darwinism: W34, W27, W15, and W7. According to Ryckman et al., the 26-item HCA had a high alpha reliability of .91 and was positively related to dogmatism, mistrust, and a tendency to use violence when threatened.  70  Three simple and straightforward items were created to assess anthropocentrism: W30, W22, and W37 (cf. the New Ecological Paradigm Scale; Dunlap, Van Liere, Mertig, & Jones, 2000). Four items were used to assess spiritual worldview in a general way (i.e., not pertaining to any specific religious/spiritual beliefs), W29, W33, W8, and W14. These items were adapted from the Royal Free Questionnaire for Spiritual and Religious Beliefs (RFQSR; King, Speck, & Thomas, 2001), an instrument designed to differentiate between people with high and low spiritual beliefs. The 18-item RFQSR had satisfactory reliability, with alpha ranging from .74 to .89. The RFQSR was found to correlate positively with an intrinsic religious motivation scale and self-reported spiritual experience (e.g., feeling at one with the universe, feelings of ecstasy). In total, there were 11 items, including a reverse scoring item. To assess whether the 11 non-HVIC worldview items reflected social Darwinism, anthropocentrism, and spiritual worldview, maximum-likelihood factor analysis with varimax rotation was performed on the student data. Three factors with eigenvalues between 2.83 and 1.31 accounted for 58.83% of the total variance. Inspection of the scree plot also yielded clear evidence for a three-factor solution. The first factor was characterized as spiritual worldview, the second as social Darwinism, and the third as anthropocentrism or human-centered worldview. As shown in Table 8, most items demonstrated strong loadings on the appropriate factors with minimal shared variance with other factors. The factors had coefficient alpha reliabilities ranging from .63 to .83; mean inter-item correlations ranging from .31 to .56; Factor 1 had a higher range of item-total scale correlations, .49 - .78, than Factors 2 and 3 (.35 - .47); and mean of item-total scale correlations ranging from .42 to .66, which were all within the acceptable range. Factor score correlations were low, below .32, supporting the uniqueness of each factor.  71  Table 8 Summary of Items, Rotated Factor Loadings (N = 640), and Internal Consistency Reliability Estimates (N = 622 631) for Worldview Subscales for the Student Sample Factor loadings 1  2  3  h2  W29. There is a spiritual power or force influencing what happens to people in their daily life.a  .902  .012  .039  .815  W33. There is a spiritual power or force influencing what happens in the world. a  .882  -.040  .035  .781  W8. There are some spiritual laws or being(s) governing the universe. a  .706  .025  -.078  .505  .514  -.127  -.033  .281  -.046  .616  .074  .387  W27. It is okay to disturb one‘s opponent in some way in order to get the edge in competition.b  -.064  .611  .093  .386  W15. I compete with others even if they are not competing with me. b  -.045  .508  .092  .269  W7. It‘s a dog-eat-dog world. If you don‘t get the better of others, they will sure get the better of you. b  .006  .466  .138  .237  W30. Humans are the center of the universe.  .081  .083  .778  .619  -.185  .185  .641  .479  W37. Humans should have control over other animals.  .042  .294  .362  .219  Eigenvalues  2.83  2.33  1.31  25.76  21.18  11.89  .56  .31  .36  No.  Item Description  a  W14. There is no life after death. (reverse scoring) W34. I see my opponents in competition as my enemies.  W22. Humans are the masters of the earth.  % of variance Mean inter-item correlation Range of item-total scale correlations  b  .49 - .78 .37 - .47 .35 - .49  Mean item-total scale correlation  .66  .42  .45  Coefficient Alpha Reliability  .83  .64  .63  Note. Loadings of .35 and above are in bold. Factor 1 = Spiritual Worldview; Factor 2 = Social Darwinism; Factor 3 = Anthropocentrism; h2 = communality. a Scale items are adapted with permission from ―The Royal Free Interview for Spiritual and Religious Beliefs: Development and Validation of a Self-Report Version‖ by M. King, P. Speck, & A. Thomas, 2001, Psychological Medicine, 31, p. 1022. b Scale items are from ―Construction of a Hypercompetitive Attitude Scale‖ by R. M. Ryckman, M. Hammer, L. M. Kaczor, & J. A. Gold, 1990, Journal of Personality Assessment, 55, p. 633. Copyright 1990 by Routledge. Reprinted with kind permission from Routledge.  72  The results of maximum-likelihood factor analysis for the adults were similar to those obtained for the students, except that the second and third factors were switched. The scales had higher coefficient alpha reliabilities in the adult data than the student data, with .88 for spiritual worldview, .66 for social Darwinism, and .75 for anthropocentrism. Factor score correlations for the adult data were also low (< .33), but the factor correlation pattern in the adult data was slightly different from that in the student data. For the adults, spiritual worldview had a positive correlation with anthropocentrism (r = .13), whereas for the students, spiritual worldview had a negative correlation with anthropocentrism (r = -.08), suggesting that spiritual worldview and anthropocentrism were more compatible for the adults, but incompatible for the students. Self-View Five aspects of self-view were assessed. Two were assessed with the Self-Construal Scale (SCS; Singelis, 1994), and three with the Moral Self, Ecological Self, and Religious/Spiritual Self subscales. The items of the three self-view subscales and the SCS were randomly ordered as a single scale asking about the participants‘ view of themselves. Respondents were asked to tell us about themselves by rating each item on a 7-point Likert-type scale ranging from 1 (Strongly Disagree) to 7 (Strongly Agree) to indicate how much they agreed or disagreed with the statement. Visual representations of the 7-point rating scale the same as those for the Moral Orientation Index (see Appendix C) were also provided to make the questionnaire more interesting. The Self-Construal Scale. The Self-Construal Scale (SCS; Singelis, 1994) was designed to measure levels of independence and interdependence in self-construal. The original SCS had 24 items, 12 assessing independence, and 12 assessing interdependence. A 2006 version of the SCS consists of 15 items reflecting independence and 15 items reflecting interdependence  73  (Singelis, personal communication, January 10, 2006). Cronbach Alpha reliabilities with either the 12 or 15 items, ranging from the high .60s to the middle .70s, were considered adequate given the broadness of the constructs. Because the scales were designed for adult samples, words that seemed uncommon or difficult for secondary students were changed; for example, ―forthright‖ was changed to ―upfront.‖ Two items in the interdependence subscale were excluded because they did not apply to every participant, as indicated by the pilot participants: ―If my brother or sister fails, I feel responsible‖ and ―I would offer my seat in a bus to my professor/teacher/boss.‖ Two items, ―I do my own thing, regardless of what others think‖ and ―I should take into consideration my parents‘ advice when making education/career plans‖ were removed to avoid inflating the correlation of self-view with moral orientation, because they were very similar in content with moral orientation items, M1 and M23 in the Moral Orientation Index. As a result, there were 14 items for the independence subscale and 12 items for the interdependence subscale in the current study (see Appendix K). Five items in the interdependence subscale and three items in the independence subscale were not included in the self-view questionnaire to avoid overlap with items in the HVIC items presented in the worldview questionnaire. They were items W4, W10, W13, W26, W35, and W12, W17, W28 in the independence and interdependence subscales, respectively. However, these items were included when computing Cronbach alpha reliabilities to get accurate estimates to reflect the ―true‖ reliabilities of the scales. Based on the analysis in the current study, coefficient alphas for the two subscales were in the acceptable range: 72 for independence and .74 for interdependence in the student sample, .76 for independence and .74 for interdependence in the adult sample.  74  Moral self, ecological self, and religious/spiritual self. Items for these self-view subscales are listed in Table 9. For the assessment of moral self, four items were constructed based on the Self-Importance of Moral Identity Measure (Aquino & Reed, 2002). The Moral Identity Measure was designed to assess two dimensions, symbolization and internalization, of moral self, by listing a set of moral traits (caring, compassionate, fair, friendly, generous, hardworking, helpful, honest, and kind) and asking respondents to rate items showing the degree to which these traits were central to their self-concept (internalization) or their actions in the world (symbolization). Six items reflected symbolization of moral self (e.g., ―I often wear clothes that identify me as having these characteristics‖) and five items internalization (e.g., ―it would make me feel good to have these characteristics‖). The two scales had acceptable internal consistency reliabilities of .77 and .71 respectively. By incorporating the moral traits into the internalization items, four items were created to assess intrinsic moral self: S2, S8, S12, and S18. Six simple items were created to assess ecological self (cf. the Environment Identity Scale, Clayton, 2003): S6, S23, S26, S29, S30, and S34. Another four items were used to assess religious/spiritual self-concept. Three items, S2, S8, and S21, were drawn from the Santa Clara Strength of Religious Faith (SCSORF; Plante & Boccaccini, 1997), and one item, S18 was constructed based on personal experience. The 10-item SCSORF was designed to measure strength of religious faith. It had high internal reliability (Cronbach alpha = .95). High faith individuals tended to have higher self-esteem, interpersonal sensitivity, and adaptive coping. To assess whether the 14 non-SCS self-view items measuring moral self, ecological self, and religious/spiritual self, maximum-likelihood factor analysis with varimax rotation was performed on the student data. As seen in Table 9, three factors with eigenvalues between 4.64 and 1.60 accounted for 66.01% of the total variance: the first was religious self; the second was  75  ecological self; and the third was moral self. This solution was also supported by the scree plot. Most items had strong loadings on their factors with minimal shared variance with other factors. Table 9 Summary of Items, Rotated Factor Loadings (N = 640), and Internal Consistency Reliability Estimates (N = 622 631) for Self-View Subscales for the Student Sample Factor loadings No. Item Description S21. I look to my faith as a source of comfort. a  1  2  3  h2  .889  .033  .149  .814  S8. My faith is an important part of who I am as a person. a S2. I look to my faith as providing meaning and purpose in my life. a  .878 .866  .025 .006  .190 .117  .808 .764  S18. I feel that I am being protected by God or other higher beings.  .775  .056  .141  .623  S34. Interacting with nature makes me feel connected with the universe.  .067  .820  .101  .688  S30. I am environmental. S26. I am my true self when I am having close, quiet interactions with nature. S23. I feel connected with animals and plants. S29. I enjoy watching sceneries of the earth and the sky, including those in pictures and paintings.  .054  .714  .201  .553  .043  .726  .149  .551  .015  .665  .037  .444  .098  .602  .206  .415  S6. I feel relaxed and calm when I am in a natural environment.  .024  .491  .222  .291  .115  .202  .743  .605  .163  .147  .727  .576  .151  .163  .681  .513  .130  .193  .596  .410  4.64 33.14  3.00 21.43  1.60 11.44  .75 .76 - .86 .82 .92  .47 .50 - .73 .63 .84  .52 .58 - .67 .63 .82  S27. It would make me feel good to be a person who is kind and generous.b S14. It makes me feel good to be a person who is helpful and friendly. b S24. Being someone who is honest and trustworthy is an important part of who I am.b S35. Being someone who is just and fair is an important part of who I am.b Eigenvalues % of variance Mean inter-item correlation Range of item-total scale correlations Mean item-total scale correlation Coefficient Alpha Reliability  Note. Loadings above .35 are in bold. Factor 1 = Religious Self; Factor 2 = Ecological Self; Factor 3 = Moral Self; h2 = communality. a Scale items are adapted with permission from ―The Santa Clara Strength of Religious Faith Questionnaire,‖ by T. G. Plante & M. T. Boccaccini, 1997, Pastoral Psychology, 45, p. 385. b Scale items are adapted with permission from ―The Self-Importance of Moral Identity,‖ by K. Aquino & A. Reed, II, 2002, Journal of Personality and Social Psychology, 83, p. 1428.  76  As presented in Table 9, the internal consistency of the subscales in the present study was high with coefficient alphas between .82 and .92. Mean inter-item correlation between .47 and .75 and mean item-total scale correlations between .63 and .82 were also high for the subscales. The range of item-total scale correlations for each subscale was within the acceptable range. The maximum-likelihood factor analysis for the adult sample replicated the results from the student sample. For adults, coefficient alpha reliabilities for the scales were even higher, with .95 for religious self, .88 for ecological self, and .78 for moral self. For both samples, factor score correlations were low, between .09 and .36, supporting the uniqueness of each factor. It was interesting that ecological self had higher correlations with moral self (around .35) than with religious self (around .10) for both samples, suggesting that ecological self had more in common with moral self than with religious self. Moral Orientation The development of the Moral Orientation Index. The Moral Orientation Index (MOI) aims to assess moral orientation from a multidimensional perspective. The MOI assesses how adolescents and adults generally solve problems or make decisions. It was constructed by revising and expanding the Objective Measure of Dispositional Moral Orientation (Yu, 2001), which was adapted from the self-description care and justice rating scales of the Measure of Moral Orientation (Liddell, 1990). The MOI has seven subscales, representing Levels 1-7 moral orientations as described in Table 1 in Chapter 2. Given the number of subscales, I needed to use the fewest possible items to measure each type of moral orientation while still maintaining high reliability. Thus, five to eight items were first created for each subscale. Items were developed based on the operational definitions provided in Table 1. A few reverse scoring items were written to control for acquiescence.  77  The readability level of the items was a great concern, especially for the student sample. I consulted high school students while writing the preliminary items and then tried out the items on a small group (about 30) of adolescents from different ethnic backgrounds. Based on their feedback, revisions of the items were made, before pilot testing the items with a larger group of high school students. All together, 45 items were included in the initial MOI and randomly ordered as a single scale. Respondents were asked to indicate how much they agreed or disagreed with each statement on a 7-point Likert scale from 1 (Strongly Disagree) to 7 (Strongly Agree). Items on each subscale were averaged to compute a composite score, after reverse scoring items where appropriate. A group of 179 high school students responded to the initial MOI in the pilot study. Maximum-likelihood factor analysis with varimax rotation was conducted on the items. Items that did not cluster with other items of their orientation were eliminated or rewritten. The final MOI consisted of 40 items and they were randomly ordered as a single scale (see Appendix C). Reliability and construct validity. Using the student sample data, maximum-likelihood factor analysis was undertaken to determine how well the 40 MOI items measured the seven moral orientations. As planned, a seven-factor solution with varimax rotation was imposed on the items. Inspection of the scree plot yielded evidence for a seven-factor solution. In Table 10 are the rotated factor loadings for the MOI items. As can be seen, Items M30R, M25, M35R, and M21 were dropped due to low loadings (< .35) on the appropriate factors. Item M15 was dropped because of its significant cross-loadings on Factor 4, Biocentric Orientation, despite its high loading on Factor 3, Justice Orientation. Despite their significant loadings on Factor 5, Care Orientation in the student data, Items M3 and M5 were also dropped because they failed to load  78  significantly on Care Orientation in the adult data. The remaining items had moderate to high communalities (.27 to .84). Table 10 Rotated Factor Loadings (N = 636) and Internal Consistency Reliability Estimates (N = 613 - 635) for Moral Orientation for the Student Sample. Factor loading  Communality  1 RO  2 EO  3 BO  4 JO  5 CO  6 NO  7 FO  M22 M34 M16 M40  .843 .774 .680 .701  .898 .851 .810 .806  .006 -.023 .067 -.041  .071 .026 .038 -.039  -.013 -.073 -.052 .026  .080 .097 .101 .112  .054 .099 .059 .123  .148 .152 .023 .140  M32 M20 M33 M1 M14 M26 M36 M30R M25  .507 .437 .352 .319 .273 .329 .281 .362 .104  .012 -.010 .060 .017 -.064 -.057 .012 -.174 .025  .698 .653 .568 .552 .501 .525 .460 .288 .073  -.013 .034 -.004 .015 -.041 -.087 .017 -.266 .037  .043 .090 -.082 -.038 -.108 -.108 -.139 -.114 .143  .001 -.038 -.052 -.069 .000 -.088 .025 -.351 .136  .132 .013 .069 .074 -.027 -.144 .013 -.173 .198  .019 -.010 .104 .041 .060 -.047 -.221 -.112 .134  M27 M17 M2 M9 M11 M15  .519 .618 .492 .330 .351 .380  .029 .061 .000 .045 -.038 -.072  .134 -.113 -.149 -.002 -.046 -.045  .696 .677 .650 .537 .449 .403  .006 .266 .118 .014 .318 .436  .055 .038 .058 .046 .141 .095  -.108 .266 .153 .051 .076 .100  .031 .007 .081 .187 .137 .046  M19 M38 M12 M18 M35R  .498 .490 .465 .473 .162  -.103 .008 -.003 .066 -.056  -.077 -.114 -.119 -.165 -.330  .112 .072 .055 .232 .060  .678 .655 .655 .534 .145  .081 .125 .094 .216 -.092  .032 .165 .085 .237 -.090  -.052 -.014 .054 .020 -.094  M6 M7 M37 M24 M3 M5  .524 .356 .284 .357 .334 .331  .029 .112 .043 .049 .107 .113  -.031 -.087 .212 -.001 -.171 -.177  .051 .137 -.051 -.017 .190 .118  .219 .044 -.014 .217 .104 .198  .652 .543 .478 .445 .457 .417  .145 .086 .074 .261 .147 .204  .159 .110 -.027 .204 .126 .138  Item No.  79  Table 10 (Continued) Rotated Factor Loadings for Moral Orientation (N = 636) and Internal Consistency Reliability Estimates (N = 613 635) for the Student Sample. Factor loading  Communality  1 RO  2 EO  3 BO  4 JO  5 CO  6 NO  7 FO  M39 M13 M31 M28 M21  .563 .350 .311 .280 .192  .033 .108 .055 .024 .142  .014 -.038 .068 .106 .205  .020 .114 .016 .213 .050  .199 .050 .058 .094 .052  .169 .105 .189 .142 .083  .701 .525 .450 .424 .328  .042 .186 .249 .116 .104  M8 M10 M4 M23 M29  .443 .443 .484 .388 .296  .105 .117 -.010 .150 .149  .034 .044 -.037 -.081 .164  .065 .196 .057 .040 .190  -.009 -.097 .116 .109 -.018  -.004 .139 .312 .174 .143  .131 .098 .279 .260 .087  .640 .592 .539 .497 .427  6.80 17.01  3.92 9.79  3.04 7.60  2.14 5.35  1.90 4.75  1.48 3.69  1.28 3.21  .74  .32  .39  .48  .33  .34  .36  .78 - .86  .43 - .60  .50 - .62  .51 - .63  .35 - .51  .38 - .53  .42 - .55  .82  .48  .53  .59  .44  .45  .50  .92  .77  .77  .79  .66  .67  .74  Item No.  Eigenvalues % of Variance Mean inter-item correlation Range of item-total scale correlations Mean item-total scale correlation Coefficient Alpha Reliability  Note. Item numbers in bold were dropped. Loadings of .35 and above are in bold. Description of items can be found in Table 1. Factor 1 = Religious/Spiritual Orientation (RO); Factor 2 = Egocentric Orientation (EO); Factor 3 = Biocentric Orientation (BO); Factor 4 = Justice Orientation (JO); Factor 5 = Care Orientation (CO); Factor 6 = Norm Orientation (NO); Factor 7 = Family Orientation (FO).  As shown in Table 10, seven factors with eigenvalues between 6.80 and 1.28 accounted for 51.40% of the total variance. Mean inter-item correlations ranging from .32 to .74 and mean item-total scale correlations ranging from .44 to .82 for the student sample were in the acceptable range. Coefficient alphas for the factors were also in the acceptable range from .66 to .92 (mean alpha = .76) for students and from .62 to .93 (mean alpha = .74) for adults.  80  As can be seen in Table 11, factor score correlations were low for both samples, mostly below .40, suggesting that each factor was independently meaningful. Egocentric orientation having negative correlations with justice and biocentric orientations but negligible correlations with family and care orientations in both samples further supported the meaningfulness of these factors. The uniqueness of egocentric orientation from family and care orientations, as indicated by the data refuted the suggestion of Rest et al. (1999) that family and care orientations are part of the self-interest schema. Religious orientation having higher correlations with family orientation than with norm orientation in both samples also challenged the idea of grouping religious orientation as part of norm-maintaining schema of Rest et al. A negative correlation between egocentric and religious orientations was found in the adult data but not in the student data, suggesting differences between the two age groups. Table 11 Moral Orientation Index Factor Score Correlation Matrix for the Student Sample (N = 638; lower-left) and the Adult Sample (N = 471; upper-right). Factor Religious Orientation  1  2  3  4  -.11*  .05  .00  -.14*  5  6  7  .14*  .22*  .36*  Egocentric Orientation  -.01  Biocentric Orientation  -.19*  -.09  .06  -.08  .07  -.09*  .25*  .08  -.02  -.01  Justice Orientation  .00  -.22*  .34*  Care Orientation  .20*  -.04  .19*  .28*  Norm Orientation  .20*  .03  .27*  .28*  .37*  Family Orientation  .29*  .01  .28*  .12*  .35*  .21*  .16*  .10*  .27*  .30* .40*  .44*  Note. * p < .05.  Validity Scale A short ―honesty check‖ was used to assess the validity of the respondent‘s answers. At the end of the survey, participants were asked about their feedback on the survey. They were  81  asked to answer a few questions by circling either ―Yes‖ or ―No.‖ The aim of this honesty check was to identify uncooperative participants who cannot be trusted to provide valid information. The first two items, ―When answering the questions in the above questionnaires, have you followed all the directions carefully?‖ and ―Have you answered all the questions honestly?‖ were positive questions. A ―Yes‖ response was expected. The third item, ―Have you purposefully skipped any of the questions?‖ was a negative question so that a ―No‖ response should be typical. Although these direct questions may not capture subtle attempts at distortion, a ―No‖ to Question 2 would clearly indicate a questionable protocol. For the student participants, the final question asked if they still consented to participate in the study after they had completed the survey. This question was designed for students who were not interested in the survey but participated just because they did not want to do class work. Protocols with a ―No‖ response to this question were dropped from the study. Issues in Model Testing Before using structural equation modeling (SEM), a few issues must be discussed: sample size, model identification, model estimation, and fit indices. The issue of sample size will be discussed in the next chapter. In this section, I describe how the other issues were addressed within the current study. Model Identification The basic principle of model identification is to determine whether constraints placed on the model are sufficient to determine unique estimates of the structural coefficients. A model is considered identified if there is a unique solution for each of the parameters in the model. The general structural equation model is composed of two parts. The first is the measurement part that links observed variables to latent variables via a confirmatory factor model. The second is  82  the structural part that links latent variables to each other via simultaneous equations (Kaplan, 2000). Therefore, I followed the Two-Step Rule to facilitate model identification (Bollen, 1989). The first step requires measurement parameters be identified, by rearranging the structural model into a confirmatory factor analysis model. When this condition is met, the second step requires the latent variables model parameters be identified, by focusing only on the structural portion of the model that deals with the regression coefficients relating latent variables to one another and rearranging the latent variable equation so that it is a structural equation made up of observed variables. If the second step is met, the model is then considered as identified and only identified models can be estimated. The Two-Step Rule was applied to the proposed covariance structure models at all seven levels of moral development. Model Estimation After a model is specified, population parameters are estimated with the goal of minimizing the difference between the observed and the estimated population covariance matrices. Model estimation is the process of obtaining an estimated population covariance matrix, ∑ that is as close as possible to the observed matrix, S. When there is little difference between the observed and the estimated population matrices, the residual matrix will be close to zero and the goodness-of-fit indices will indicate a good fit between the hypothesized model and the data. To identify a model, the scale of a factor (latent construct) must be defined. This can be done by fixing to 1 the regression or path coefficient from the factor to one of the measured variables or indicators. Fixing the path coefficient to 1 gives the factor the same variance as the indicators. The other parameters can be set as free parameters or equal parameter (i.e., the path  83  coefficients are equal and set to an unknown value to be estimated). Owing to the age and education differences between the two samples in this study, having equal parameters in both samples was rare, however. After all the parameters have been set, estimates of the free parameters including the equal parameters are then derived from the matrix of observed variables. That is, the estimates for the matrix ∑ are acquired from the observed matrix S. Most commonly used estimation methods, such as GLS (Generalized Least Squares), ULS (Unweighted Least Squares), ADF (Asymptotically Distribution Free), and ML (Maximum Likelihood) are iterative fitting functions designed to reduce differences between the observed and the estimated population covariance matrices. Each estimation method produces parameter estimates, standard error of estimates for the free parameters, and a test statistic to evaluate the null hypothesis and each has its own criterion to begin estimation of the free parameters of the ∑ matrix. After each iteration, the ∑ matrix is compared to the S matrix and a residual matrix is obtained. Iterations continue until the residual matrix cannot be further reduced. The model is said to have converged and a set of fitting function values are given to assess the fit of the model. Of the commonly used estimation methods, ML has been found to be the most consistent, unbiased, scale-invariant, scale-free, and normally distributed if the observed variables are multivariate normal and independent (Schumacker & Lomax, 1996) and even to be quite robust to the violation of normality (Chou & Bentler, 1995). Because the observed variables in this study were multivariate normal and independent, ML estimation was used. Goodness-of-Fit Indices There are many published goodness-of-fit indices for structural equation models. According to Hu and Bentler (1995), indices of fit can be classified loosely into two broad categories, the absolute versus the incremental fit indices. The absolute indices, such as chi-  84  square (χ2), chi-square to degrees-of-freedom ratio or relative chi-square (χ2 : df), the root mean square error of approximation (RMSEA), and expected cross-validation index (EVCI), directly assesses how well a target model reproduces the sample data, by comparing it (implicitly or explicitly) to a saturated model that exactly reproduces the observed covariance matrix. In contrast, incremental fit indices measure the proportionate improvement in fit by comparing a target model with a more restricted, nested baseline model, such as the independence model in which all the observed variables are uncorrelated. There are three types of incremental indices of fit, such as the Bentler-Bonett (1980) Normed Fit Index (NFI; Type I), the Tucker-Lewis Index (TLI; Type II), and the Bentler (1990) Comparative Fit Index (CFI; Type III). Based on guidelines in the literature, the above-mentioned goodness-of-fit indices were chosen for use in this study. The fit indices and their associated critical values for interpretation are listed in Table 12. Table 12 Indices of Model Fit and Comparison Absolute Fit Indices  Critical Values   Chi-square table value (p > α)  Relative chi-square (χ : df)    Less than 5  Root mean square error of approximation (RMSEA)    Less than .08  Expected cross-validation index (EVCI)    The smaller the value the better the fit  Bentler-Bonett (1980) normed fit index (NFI)    0 (no fit) to 1 (perfect fit)  Tucker-Lewis index (TLI)    The greater the value the better the fit  Bentler (1990) Comparative fit index (CFI)    0 (no fit) to 1 (perfect fit)  Chi-square goodness-of-fit index (χ2) 2  Incremental Fit Indices  Absolute fit indices. The Chi-square goodness-of-fit index (χ2) and chi-square to degrees-of-freedom ratio (χ2: df) test the fit between the specified model and sample data. A  85  nonsignificant chi-square value is the desired outcome, because it means that the difference between ∑ and S covariances matrix is due to sampling fluctuations and that the two matrices are not significantly different. However, when sample size increases, so does chi-square, which tends to result in greater probability of a significant chi-square, indicating a poor model. If the sample size is small enough (below 100), a good fit using χ2 is very likely; if the sample size is large enough (above 200), a good fit is difficult to achieve (Schumacker & Lomax, 1996). Thus, Wheaton, Muthén, Alwin, and Summers (1977) suggest that researchers compute a relative chisquare (χ2: df) and a ratio of approximately five or less as indicating a reasonable fit. Wheaton et al.‘s suggestion was particularly pertinent to the current study, due to the relatively large size of the two samples. RMSEA (root mean square error of approximation) is less affected by sample size than the chi-square, because it is an index based on the population discrepancy. The population discrepancy function, F0, is the value of the discrepancy function obtained by fitting a model to the population moments rather than to sample moments. F0 tends to favor models with many parameters. To counteract the effect of model complexity, F0 is divided by the number of degrees of freedom for testing the model. Taking the square root of the resulting ratio gives the population RMSEA (Arbuckle, 2006). Browne and Cudeck (1993) suggested that a value of 0.08 or less for the RMSEA would indicate a reasonable error of approximation. ECVI (expected cross-validation index) is a composite measure of badness of fit and model complexity and is valuable in assessing the relative fit of structural equation models that differ with respect to restrictiveness. It is thus intended for model comparisons and not for the evaluation of an isolated model (Browne & Cudeck, 1993). Simple, good fitting models receive  86  low scores, whereas, complicated, poorly fitting models get high scores. The smaller the ECVI the better the model is expected to cross-validate on a new sample (Arbuckle, 2006). Incremental fit indices. Bentler-Bonett‘s (1980) normed fit index (NFI) is based on comparing the minimum discrepancy of the target model to the minimum discrepancy of the baseline model, which is the independence model. NFI is guaranteed to be between 0 and 1, with 1 indicating a perfect fit (Arbuckle, 2006). The Tucker-Lewis Index (TLI) is the same as the Bentler-Bonett (1980) non- normed fit index and is calculated using chi-square. As a non-normed index, TLI can make interpretation difficult because it does not have a maximum and minimum value. In most cases, TLI values will fall between 0 and 1, but it is not limited to that range, and TLI with values of .90 or greater reflect good fit. Bollen (1990) found that the calculation of the TLI is affected by sample size, whereas the means of the sampling distributions are mostly unaffected by N. That is, TLI values are more stable when comparing samples with different Ns, although they are not normed. The comparative fit index (CFI; Bentler, 1990) is the same as McDonald and Marsh‘s (1990) relative noncentrality index. The CFI is truncated to fall in the range from 0 to 1; that is, values bigger than 1 are reported as 1 and values less than 0 are reported as 0 (Arbuckle, 2006). Hu and Bentler (1995) evaluated various measures of model fit and recommended the use of CFI and TLI as alternative measures of model fit when sample size is large (above 250) and a fit index exceeding .90 as indicating good fit. Summary The student sample of 640 secondary students was used to develop the models. The adult sample of 472 adults was used as a comparison sample. A comparison of the models obtained  87  from the student data with those obtained from the adult data can inform us not only about age or educational differences but also the validity of the models. The research design was a purpose driven sample survey. Participants were asked to complete six sets of questionnaires. Ten measures were used to assess the four latent constructs in this study. Of the 10 measures, two were published measures; five were adapted from published measures; and three were developed specifically for this study. All the measures used in the study had acceptable internal consistency reliability. For the developed and adapted measures, construct validity was established through maximum likelihood factor analysis. Structural equation modeling was used to test the theory-based causal assumptions included in the hypothesized structural model presented in Figure 3. The structural model was identified with the two-step rule and maximum likelihood estimation. Model fit was determined through the chi-square statistics, chi-square to degrees-of-freedom, the root mean square error of approximation, expected cross-validation index, the Normed Fit Index, the Tucker-Lewis Index, and the Comparative Fit Index.  88  CHAPTER FOUR: RESULTS In this chapter, I describe the results of the preliminary analysis, followed by a description of the modeling process. Then I report the subsidiary structural model tests from Level 1 to Level 7, without repeating the modeling process, because all the structural models are estimated using the same procedures. Based on the test results from all seven levels, the five individual path hypotheses are evaluated. Preliminary Analysis Sample Size and Missing Data Sample size is an important issue in SEM, because too few cases relative to the number of observed variables can lead to problems of nonconvergence and negative error variances (Tabachnick & Fidell, 1996). Sample size can also affect the goodness-of-fit indices when N is used directly in the calculation of fit indices or associated with the means of the sampling distributions of the fit indices (Bollen, 1990). Schumacker and Lomax (1996) reviewed the sample size literature and reported that larger samples are better, and that most published studies had from 250 to 500 cases. Bentler and Chou (1987) suggested that five cases per observed variable would be sufficient if the sample distribution was normal and if there were multiple indicators per latent variable. Based on this information, the sample sizes of 640 students and 472 adults in the current study with 27 latent variables and 70 observed variables were considered adequate. Cases to latent-variables ratios were 23:1 and 17:1 and cases to observed variables ratios were 9:1 and 7:1 for the student and adult samples, respectively. Missing observations ranged from 1.56 % to 0.16% for the student sample and from 2.33% to 0.21% for the adult sample, which is considered low enough to have minimal effects on  89  the results. Moreover, cases with missing values seemed to be randomly distributed throughout the data. Thus, missing data for each observed variable were replaced with the variable mean. Outliers and Normality The data were first screened for univariate outliers. Using SPSS, I calculated the standardized scores for each case. An outlier was indicated by a Z score in excess of 3.29. Variables with extreme outliers also had higher skewness and kurtosis. To reduce the influence of extreme outliers, I replaced them with a raw score one unit larger or smaller than the next most extreme score in the distribution for both the student and adult samples (Tabachnick & Fidell, 1996). Next the data were screened for multivariate outliers. I calculated the Mahalanobis distance (D2) for each case in SPSS. For Level 1, there were eight observed variables, the chisquare critical value at p < .001 was approximately 26.1; for Level 3, there were 9 observed variables, the chi-square critical value at p < .001 was approximately 27.9; for Level 4, there were 11 observed variables, the chi-square critical value at p < .001 was approximately 31.3; for Level 5, there were 10 observed variables, the chi-square critical value at p < .001 was approximately 29.6; and for Levels 2, 6, and 7, there were 12 observed variables, the chi-square critical value at p < .001 was approximately 32.9. The multivariate outliers were then examined individually and as a group to see if there were any variables that separated the group of outliers from the rest of the cases. This procedure helped me decide whether the outlying case(s) were accurate observations or should be deleted from further analysis. I performed this procedure seven times separately for each group of variables at each level. As a result, three cases at Level 1, 14 cases at Level 2, two cases at Level 3, 10 cases at Level 4, and 15 cases at Level 5 were identified as outliers. These outlying cases  90  were excluded from the analysis, because they differed from the total sample on skewness, kurtosis, and means. In contrast, 13 cases at Level 6 and 27 cases at Level 7 were also identified as outliers, but none were deleted from further analysis, because a comparison indicated that skewness, kurtosis, and means were comparable between the outlier cases and the total sample, suggesting that all were acceptable observations. The univariate summary statistics for the groups of outlying cases and the total sample with the outlier cases removed (where appropriate) at each level for the student sample are listed in Appendix L. Normality of variables was assessed by looking at the values for skewness and kurtosis and graphical outputs, such as histrograms, boxplots, normal probability plots, and detrended normal probability plots. As can be seen in Appendix L, none of the univariate analyses of the observed variables indicated severe skewness or kurtosis. The Modeling Process Step 1: Fitting the Measurement Model To fit the measurement model, each set of observed variables was fitted to its original latent constructs. When fitting the measurement model, I attempted to retain at least three marker variables per latent construct. For example, the latent construct independent self was measured by 14 items. Among these 14 items, three items that had the highest loadings and the best fit with the construct Independent Self were chosen as the set of marker variables for that particular latent construct (cf. Gable & Wolf, 1993). This process of determining marker variables was repeated for each latent variable. The latent variables and corresponding marker variables were then combined into a measurement model and retested for overall fit. For the latent construct worldview at Level 1 that combines vertical individualism (VI) with social Darwinism (SD), however, it was more appropriate to use composite variables rather  91  than item variables to fit the model, because the use of composite variables results in better fit as compared with item solutions (Nasser & Wisenbaker, 2003). Vertical individualism was measured with the Measurement of Horizontal and Vertical Individualism and Collectivism (HVIC; Triandis & Gelfand, 1998) and Social Darwinism was assessed with four items of the Hypercompetitive Attitude Scale (HCA; Ryckman et al., 1990). Both HVIC and HCA were measures with well-known psychometric properties, so composite scores are more reliable than single items. Composite variables, VI and SD, were used for the latent construct worldview at Level 1. That is, Level 1 worldview was indicated by VI and SD. For the same reasons, the latent construct worldview at Levels 4, 5, and 6 that was composed of two or more worldview variables was fitted with composite variables. On the contrary, the latent construct worldview at Levels 2, 3 and 7 was fitted with (item) marker variables, because there was only one worldview variable at Levels 2, 3, and 7, namely, vertical collectivism, horizontal collectivism, and spiritual worldview, respectively. There were seven measurement models for the current study, one at each level of inclusiveness. All the observed variables included in the measurement models are described in Appendix M. Step 2: Comparing the Measurement Model to Alternative Models When the measurement model was determined to have a good fit, it was compared to a series of alternative models. The ideal situation is to have the hypothesized model fit the data as well as the measurement model (not significantly worse) and better than the alternative models. Six structural models were included in this study: (a) the original hypothesized target model as depicted in Figure 3, (b) a self-view model, (c) a correlational model, (d) a mediational model that starts with relationship closeness, (e) a theorized model for egocentric orientation without  92  relationship closeness, and (f) a moral orientation model (see Figure 4). The alternative models represent possible relational structures among worldview, self-view, moral inclusiveness, and moral orientation that are different from and even contradictory to the hypothesized model. Most of the alternative models were based on theory, except the moral orientation model. The hypothesized model was developed based on a review of the literature. The hypothesized model being of most interest was labeled Model A. Self-View  Worldview Moral Orientation  Worldview  Moral Orientation  Self-View  Moral Inclusiveness  Moral Inclusiveness  (B) Alternative self-view model  (A) Hypothesized model  Self-View  Self-View Moral Orientation  Worldview  Moral Orientation  Worldview Moral Inclusiveness  Moral Inclusiveness  (C) Alternative correlational model  (D) Alternative mediational model  Self-View  Self-View  Worldview  Moral Orientation  Moral Orientation  Worldview Moral Inclusiveness  (E) Theorized model for egocentric orientation  (F) Alternative moral orientation model  Figure 4. Hypothesized and alternative models.  93  The recent fascination with the self in our society can be exemplified by the thousands of empirical studies on the self. Self and identity development is now seen as the key to personal success. Failing to know or be oneself or having low self-esteem has been blamed for many personal and interpersonal problems. Given such obsession with the self in the Western culture, it is possible that it is the self that affects worldview and relationship closeness rather than worldview affecting self-view and relationship closeness. Thus, self-view, as the exogenous variable, positively predicts worldview and moral inclusiveness, and together they positively predict moral orientation, as portrayed in Model B. In line with the holistic view, moral development is a dynamic system characterized by patterns of reciprocal feedback, so that changes in one part of the system or model can cause changes in other parts of the model. Therefore, worldview, self-view, and relationship closeness could be so closely correlated that it would be difficult or even futile to try to find the cause or origin of development. That is, these three constructs are correlated, and they all positively predict moral orientation, as illustrated in Model C, a correlational model. Another alternative model for moral orientation is a mediational model that starts with relationship closeness, as shown in Model D. As the exogenous variable, relationship closeness influences moral orientation through the mediation of worldview and self-view. Similarly, worldview influences moral orientation through the mediation of self-view. Model E is a theorized model of moral development for egocentric orientation at Level 1. Because there is only the self in Level 1 moral inclusiveness, there will be little or no relationship closeness to any groups. Thus, moral inclusiveness is removed from the model. Model E would not normally be estimated unless there was evidence that relationship closeness  94  was not related to moral orientation. If relationship closeness was unrelated to moral orientation, Model E should have better fit than other alternative models. The final alternative model starts with moral orientation, which positively predicts worldview, self-view, and relationship closeness, as depicted in Model F. Model F is thus the exact opposite of the hypothesized model. Given the cross-sectional and correlational nature of the current data, it is very difficult to claim causal relations among the latent constructs, even when the hypothesized model fits as well as the measurement model. If the data are merely correlational, however, Model F should fit the data as well as Model A or any other alternative models. A comparison of the hypothesized model with the alternative models including Model F could thus provide more evidence for the hypothesized pseudo-causal relations among worldview, self-view, moral inclusiveness, and moral orientation. Comparing alternative models. The alternative models were compared using chi-square difference tests. This process determined whether additional constraints imposed by the alternative models significantly worsened the fit compared to the fit of the measurement model. The chi-square difference test is calculated by subtracting the measurement model chi-square from an alternative model chi-square; the degrees of freedom are also subtracted. The result is a chi-square difference based on the degrees of freedom difference. A significant chi-square difference indicates that the constraints added by the alternative model create a fit that is significantly worse than the fit of the measurement model. Conversely, a nonsignificant chisquare difference indicates that the alternative model fits the data as well as the measurement model; has additional degrees of freedom; and is accepted as being more parsimonious (Arbuckle, 2006). The best fitting model at each level was then validated on data collected from the adult sample.  95  Step 3: Validating a Structural Model The best way to validate a structural model is by testing invariance of the structural model across samples. In the current study, the adult sample provided the data for model validation. Because the adult sample differed vastly from the student sample in education and gender ratio, the structural models developed from the student data but fitting well on the adult data could be taken as strong evidence for the validity of the structural models, stronger than cross-validating the models across samples that are similar in education and gender-ratio. In AMOS software (Arbuckle, 2006), invariance is tested sequentially as shown below from Model 1 (the least degree of invariance) to Model 7 (the most constrained model and the highest degree of invariance): (1) Unconstrained: Factor patterns are invariant across groups. (2) Measurement weights: Factor patterns and loadings are invariant across groups. (3) Measurement intercepts: All of the above and measurement intercepts (in the equations for predicting measured variables) are invariant across groups. (4) Structural weights: All of the above and the structural (regression) weights (for predicting the DVs) are invariant across groups. (5) Structural covariances: All of the above and the structural covariances (the variance of IVs) are invariant across groups. (6) Structural residuals: All of the above and the structural residuals (the variance of the error variable of the DV or simply the error of the DV) are invariant across groups. (7) Measurement residuals: All parameters are constant across groups. There were seven analyses in total for each of the seven models. For each analysis, covariances matrices for both samples were analyzed simultaneously, with one indicator fixed at  96  1.0 as a means of establishing a common scale between the observed and latent variables. Model fit was decided by the fit indices chosen for the study (see Table 12 in Chapter 3). That is, degree of invariance was determined by all the fit indices. AMOS also produces nested model comparisons after the model fit indices. By assuming one model to be correct, AMOS performs chi-square difference tests to see if the next models with more constraints are correct (i.e., not fit worse than the previous one). A nonsignificant chisquare difference suggests that the next model is likely to be invariant and expected to crossvalidate on a new sample (Arbuckle, 2006). Thus, nonsignificant chi-square difference between two adjacent models can be taken as evidence for some degrees of invariance. Tests of Structural Models In this section, I present the results of the structural modeling tests from Levels 1 to 7. The five individual path hypotheses in the structural model of moral development are then examined based on the test results from all seven levels. Predicting Egocentric Orientation at Level 1 As hypothesized, at Level 1, vertical individualism (VI) and social Darwinism (SD) positively predict independent self and egocentric orientation. Independent self also positively predicts egocentric orientation. No relationship closeness to any groups are positively correlated with vertical individualism, social Darwinism, and egocentric orientation. The estimated measurement model was good; so was the theorized model for egocentric orientation. Both models had exactly the same fit indices: χ2 (20, N = 637) = 35.839, p = .016, χ2: df = 1.792, NFI = .967, TLI = .973, CFI = .985, RMSEA = .035, and ECVI = .132. The theorized model was then fitted on the adult data and underwent invariance test. The estimated adult model was also acceptable, χ2 (20, N = 468) = 51.572, p = .000, χ2: df = 2.579,  97  NFI = .935, TLI = .926, CFI = .959, RMSEA = .058, and ECVI = .213. However, the chi-square in Table 13 shows that invariance tests were all statistically significant, indicating that the model was not invariant across students and adults. Despite the significant chi-squares, Models 1 and 2 had acceptable relative chi-square, NFI, TLI, and CFI values. Besides, if students and adults had the same factor patterns, then a nonsignificant Δχ2 (5) = 5.032, p = .412, would allow accepting the hypothesis of equal measurement weights (i.e., Model 2 would be invariant across the age groups). Similarly, if Model 4 (Structural weights) were correct, Models 5 to 7 would be invariant, as each showed a nonsignificant chi-square difference: Δχ2 (1) = .233, p = .629 for Model 5; Δχ2 (3) = 1.836, p = .607 for Model 6; and Δχ2 (11) = 11.145, p = .436 for Model 7. Moreover, the RMSEAs and the ECVIs for the seven models were all within the acceptable range. Particularly, the ECVIs ranging from .171 to .406, when compared to the ECVI of 1.736 for the independence model, indicated that these models are likely to be stable in the population and to replicate, because small ECVIs indicate that the model is expected to cross-validate on a new sample. Thus, there was evidence for some kind of invariance across groups. Table 13 Invariance Tests of Factor Structures for Egocentric Orientation Between Students (n = 637) and Adults (n = 468) χ2  df  p  χ2 : df  NFI  TLI  CFI  RMSEA  ECVI  1. Unconstrained  85.151  34  .000  2.504  .955  .940  .972  .037  .175  2. Measurement weights  90.182  39  .000  2.312  .952  .948  .972  .034  .171  3. Measurement intercepts  363.558  47  .000  7.735  .807  .732  .825  .078  .404  4. Structural weights  371.604  50  .000  7.432  .803  .744  .822  .076  .406  5. Structural covariances  371.836  51  .000  7.291  .802  .750  .823  .076  .404  6. Structural residuals  373.439  53  .000  7.046  .802  .760  .823  .074  .402  7. Measurement residuals  382.749  61  .000  6.275  .797  .790  .822  .069  .396  .000  0  1882.637  72  Model  8. Saturated model 9. Independence model  1.000 .000  26.148  .000  1.000 .000  .000  .160 .151  1.736  Note. Acceptable values are in bold. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; TLI = Tucker-Lewis index.  98  Figure 5 is the final structural model for egocentric orientation for both students and adults. As seen in Figure 5, Level 1worldview and independent self accounted for 54% of the variance of egocentric orientation, which was considered moderate. The total standardized effects of VI and SD on egocentric orientation were around .51 in both samples, but the effects of independent self on egocentric orientation were higher in adults (.63) than in students (.49).  Figure 5. Theorized model for egocentric orientation at Level 1. Bold-faced quantities close to the latent variables are their squared multiple correlations. Quantities near paths are standardized path coefficients at p < .001. SD = Social Darwinism; VI = Vertical Individualism.  The correlation matrix for the observed variables is presented in Table 14. As seen in Table 14, VI and SD had higher correlations with independent self in students (mean r = .14) than in adults (mean r = .06), suggesting that VI and SD had little influence on independent self in this adult sample. This difference could be the main reason for lack of invariance of the model. The higher positive correlations between horizontal individualism and independent self in adults (.55) than in students (.42) further suggested that horizontal individualism had more influence on independent self than did VI and SD in the adults than in the students.  99  Because horizontal individualism had more influence on independent self, I replaced social Darwinism with horizontal individualism in the model in Figure 5. However, the model was not successfully fitted because worldview (horizontal and vertical individualism) failed to predict egocentric orientation. These findings further supported my theory of extreme vertical individualism combining with social Darwinism, not individualism per se, being the cause of egocentric orientation. Table 14 Correlation Matrix for Observed Variables at Level 1 for the Student Sample (N = 626, the lower-left) and the Adult Sample (N = 462, the upper-right) Variables  VI  SD  S3  S9  S20  M1  M20  M32  RC2  RC3  RC4  RC5  VI .709 .037 -.015 .157 .348 .200 .290 -.024 .087 .139 .042 SD .706 .045 .001 .124 .365 .232 .337 -.056 .057 .052 .087 S3 .155 .130 .132 .064 .146 .287 .216 -.104 -.052 .000 -.013 S9 .099 .120 .193 .235 .131 .101 .091 .002 .062 .009 .085 S20 .173 .169 .099 .149 .158 .175 .214 .068 .017 .006 .039 M1 .242 .278 .162 .207 .051 .414 .431 -.018 .023 .020 -.028 M20 .320 .301 .269 .148 .083 .441 .495 -.113 -.041 -.027 .000 M32 .340 .285 .212 .137 .130 .435 .540 -.005 .049 -.006 .074 RC2 -.021 -.068 .017 .100 .095 .032 .007 .056 .436 .354 .312 RC3 -.026 -.048 .076 .146 .133 .083 .039 .091 .265 .569 .347 RC4 .051 -.052 .061 .083 .094 .053 .044 .051 .341 .392 .425 RC5 .008 .012 .033 .137 .081 .095 .015 .107 .330 .253 .432 Note. RC2 = Closeness to Family; RC3 = Closeness to Friends; RC4 = Closeness to Society; RC5 = Closeness to Nation; SD = Social Darwinism; VI = Vertical Individualism.  As can be seen in Table 14, vertical individualism and social Darwinism had high positive correlations and they both had little or negligible correlations with any relationship closeness, suggesting overlap between the two factors and that individuals embracing vertical individualism and social Darwinism have little relationship closeness with other people or groups. The highest positive correlation found between vertical individualism and closeness to society (RC4) in adults suggested that, unlike social Darwinism, vertical individualism is  100  compatible with or acceptable in our hierarchical society. Egocentric orientation also had negligible correlations with most relationship closeness, supporting that egocentric orientation reflects Level 1 inclusiveness and thus has little or no relationship closeness to any groups. Despite lack of full invariance across age groups, the hypothesized relationships at Level 1 were fully supported. Together, vertical individualism and social Darwinism positively predicted independent self and egocentric orientation. Independent self also positively predicted egocentric orientation. Vertical individualism, social Darwinism, and egocentric orientation had negligible correlations with relationship closeness, supporting that Level 1 inclusiveness includes no others but the self. Predicting Family Orientation at Level 2 At Level 2, higher vertical collectivism was hypothesized to predict higher interdependent self and relationship closeness to family. Together, higher vertical collectivism, interdependent self, and closeness to family predict higher family orientation. The estimated measurement model was acceptable, as indicated by the fit indices listed in Table 15, despite a significant chi-square. Among the alternative models, the solution for Model B was not admissible. Model D had the highest squared multiple correlations (SMC) for family orientation, but also the largest chi-square, so large that it created a fit significantly worse than the fit of the measurement model, as indicated by a significant Δχ2 (2) = 16.820, p < .05. Model F had the same degree of freedom as did Model A, but higher chi-square, Δχ2 = 3.208. Moreover, in Model F, closeness to family and interdependent self both failed to predict vertical collectivism statistically significantly. Model A, the target model had the best fit, because it had one more degree of freedom than the measurement model and Model C, but lower chi-square, Δχ2 = .326. Thus, Model A was the chosen model for invariance testing.  101  Table 15 Goodness-of-Fit of Alternative Models Predicting Family Orientation at Level 2 (N = 626) χ2  df  p  χ2 : df  NFI  TLI  CFI  RMSEA  ECVI  Measurement  143.787  53  .000  2.713  .940  .943  .961  .052  .348  Model A  144.113  54  .000  2.669  .940  .944  .961  .052  .346  .80  Model C  143.787  53  .000  2.713  .940  .943  .961  .052  .348  .80  Model D  160.607  55  .000  2.920  .933  .936  .955  .055  .369  .87  Model  SMC  Model F 147.321 54 .000 2.728 .939 .942 .960 .053 .351 Note. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; SMC = squared multiple correlations for family orientation; TLI = Tucker-Lewis index.  The estimated adult model was acceptable, χ2 (55, N = 464) = 98.955, χ2 : df = 1.799, p = .000, NFI = .946, TLI = .964, CFI = .975, RMSEA = .042, and ECVI = .365. Nevertheless, the chi-squares listed in Table 16 show that invariance tests across students and adults were all significant, indicating that the model was not invariant across the two age groups. Table 16 Invariance Tests of Factor Structures for Family Orientation Between Students (n = 626) and Adults (n = 464) χ2  df  p  χ2 : df  NFI  TLI  CFI RMSEA  1. Unconstrained  242.976  109  .000  2.229  .943  .953  .967  .034  .354  2. Measurement weights  253.791  111  .000  2.286  .940  .951  .965  .034  .360  3. Measurement intercepts  481.071  123  .000  3.911  .887  .889  .912  .052  .547  4. Structural weights  496.295  128  .000  3.877  .883  .890  .910  .051  .552  5. Structural covariances  496.372  129  .000  3.848  .883  .891  .910  .051  .550  6. Structural residuals  504.228  132  .000  3.820  .881  .892  .909  .051  .552  7. Measurement residuals  649.472  144  .000  4.510  .847  .866  .876  .057  .663  .000  0  4248.133  156  Model  8. Saturated model 9. Independence model  1.000 .000  27.232  .000  ECVI  1.000 .000  .000  .331 .155  3.949  Note. Acceptable values are in bold. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; TLI = Tucker-Lewis index.  However, if students and adults had the same structural weights (i.e., assuming Model 4 was correct), then it would be acceptable to assume equal structural covariances, as indicated by  102  a nonsignificant Δχ2 (1) = .077, p = .783, and equal structural residuals, as indicated by a nonsignificant Δχ2 (4) = 7.932, p = .094. Moreover, many of the fit indices, such as the relative chi-square, CFI, RMSEA, and ECVI were in the acceptable range, suggesting that these models are pretty good models and likely to cross-validate on a new sample. The final structural model predicting family orientation for both students and adults is presented in Figure 6. The total effects of Level 2 worldview (vertical collectivism) on family orientation were .84, higher than that of interdependent self and closeness to family on family orientation. Together they accounted for 76% of the variance in family orientation, which was considered moderate to high.  Figure 6. Hypothesized model for family orientation at Level 2. Bold-faced quantities close to the latent variables are their squared multiple correlations. Quantities near paths are standardized path coefficients at p < .001, except *p < .05 and **p < .01.  The path predicting family orientation from interdependent self was nonsignificant, r = .09, p = .502, in the adult model, whereas in the student model, the path predicting family  103  orientation from interdependent self was small but statistically significant, r = .20, p = .028. This difference between the adult and student data could be the main reason for failure of the full invariance tests. As can be seen in Table 17, interdependent self had the lowest positive correlations with other variables in both samples, suggesting that interdependent self was not as effective as vertical collectivism and closeness to family in predicting family orientation. Table 17 Correlation Matrix for Observed Variables at Level 2 for the Student Sample (N = 587, the lower-left) and the Adult Sample (N = 450, the upper-right) Variables W11  W16  W24  S1  S7  S28  C4  C14  C24  M4  M8  M23  W11 W16 W24 S1 S7 S28 C4 C14 C24 M4 M8 M23  .480  .481 .452  .111 .135 .157  .264 .243 .185 .166  .254 .143 .188 .099 .145  .354 .321 .390 .029 .139 .030  .342 .265 .344 .016 .128 .089 .746  .377 .297 .380 .076 .127 .116 .670 .712  .451 .390 .493 .137 .235 .189 .386 .375 .380  .422 .300 .373 -.015 .127 .155 .344 .296 .368 .390  .336 .312 .421 .112 .187 .188 .343 .387 .411 .527 .417  .446 .400 .073 .270 .114 .407 .490 .458 .407 .391 .365  .366 .156 .206 .132 .304 .313 .261 .337 .307 .295  .219 .171 .214 .214 .286 .292 .375 .312 .324  .282 .078 -.020 .034 .030 .160 .072 .169  .083 .191 .233 .263 .315 .139 .272  .027 .057 .056 .166 -.015 .123  .729 .668 .357 .326 .356  .753 .439 .362 .421  .427 .351 .438  .367 .481  .316  As expected, vertical collectivism positively predicted interdependent self, closeness to family, and family orientation. Interdependent self and closeness to family positively predicted family orientation. Thus, the hypothesized causal relationships at Level 2 were considered fully supported, in spite of lack of full invariance across the two samples. Predicting Care Orientation at Level 3 At Level 3, it was hypothesized that horizontal collectivism positively predicts interdependent self, closeness to friends, and care orientation. Interdependent self and closeness to friends also positively predict care orientation.  104  The estimated measurement model was acceptable, as indicated by the fit indices in Table 18. Again, the solutions for Model B and Model D were not admissible, because the standardized path coefficients predicting care orientation from interdependent self were greater than 1. The solution for Model F was not proper because the path predicting horizontal collectivism from care orientation was greater than 1 and interdependent self failed to predict horizontal collectivism. As seen in Table 18, Model A was worse than the fit of the measurement model, as indicated by a significant Δχ2 (1) = 5.705, p < .05. Model C had the same fit indices as the measurement model. However, in Models C and A, the path predicting care orientation from closeness to friends was nonsignificant, rs = .07 and -.01 respectively, suggesting that closeness to friends was unrelated to care orientation and should be removed. After deleting closeness to friends, the modified model, called Model E, was a better model, as indicated by the fit indices in Table 18, with no loss of SMC for care orientation. Thus, Model E was the chosen model for invariance tests. Table 18 Goodness-of-Fit of Alternative Models Predicting Care Orientation at Level 3 (N = 638) χ2  df  p  χ2 : df  NFI  TLI  CFI  RMSEA  ECVI  Measurement  104.860  53  .000  1.978  .949  .962  .974  .039  .281  Model A  110.565  54  .000  2.048  .946  .959  .972  .041  .287  .94  Model C  104.860  53  .000  1.978  .949  .962  .974  .039  .281  .94  Model  Model E 47.818 28 .011 1.708 .952 .967 .979 .033 .157 Note. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit  SMC  .95  index; RMSEA = root-mean-square error of approximation; SMC = squared multiple correlations for care orientation; TLI = Tucker-Lewis index.  The chi-squares in Table 19 show that all invariance tests across students and adults are significant, indicating lack of invariance across age, despite an acceptable adult model, χ2 (27, N  105  = 468) = 84.36, χ2 : df = 3.124, p = .000, NFI = .891, TLI = .868, CFI = .921, RMSEA = .067, and ECVI = .296. However, most of the other fit indices were within the acceptable range, such as the relative chi-square, CFI, RMSEA, and ECVI. Moreover, if students and adults had the same factor patterns, a nonsignificant Δχ2 (6) = 4.645, p = .590, would allow accepting the hypothesis of equal measurement weights. Similarly, if Model 3 was assumed to be correct, then Models 4 to 6 would be invariant, as each resulted in a nonsignificant chi-square difference: Δχ2 (3) = 1.135, p = .769 for Model 4; Δχ2 (4) = 7.212, p = .125 for Model 5; and Δχ2 (6) = 7.557, p = .272 for Model 6. That is, the model could be invariant across age groups and would likely crossvalidate on an independent sample, given similar factor patterns. Table 19 Invariance Tests of Factor Structures for Care Orientation Between Students (n = 638) and Adults (n = 468) Model  χ2  df  p  χ2 : df  NFI  TLI  CFI RMSEA  ECVI  1. Unconstrained  125.991  48  .000  2.625  .929  .913  .954  .038  .223  2. Measurement weights  130.636  54  .000  2.419  .926  .924  .955  .036  .216  3. Measurement intercepts  191.784  63  .000  3.044  .892  .891  .924  .043  .255  4. Structural weights  192.919  66  .000  2.923  .891  .897  .925  .042  .251  5. Structural covariances  198.995  67  .000  2.970  .888  .895  .922  .042  .255  6. Structural residuals  199.341  69  .000  2.889  .888  .899  .923  .041  .251  7. Measurement residuals  242.988  78  .000  3.115  .863  .887  .902  .044  .274  .000  0  8. Saturated model  1.000  1.000  .196  9. Independence model 1775.271 90 .000 19.725 .000 .000 .000 .130 1.641 Note. Acceptable values are in bold. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; TLI = Tucker-Lewis index.  The final structural model predicting care orientation for both students and adults is presented in Figure 7. The total effects of horizontal collectivism on care orientation (.67 in students and .57 in adults) were less than that of interdependent self on care orientation (.80 in students and .85 in adults). Moreover, the path predicting interdependent self from horizontal collectivism was higher in students (.45) than in adults (.33). Given all these differences between  106  students and adults, the final model had achieved only weak invariance. Anyway, horizontal collectivism and interdependent self accounted for 97% of the variance in care orientation, which is very high.  Figure 7. Modified model for care orientation at Level 3. Bold-faced quantities close to the latent constructs are their squared multiple correlations. Quantities near paths are standardized path coefficients at p < .001.  When horizontal collectivism was replaced with vertical collectivism (Level 2 worldview), the SMC for care orientation dropped to .78, and vertical collectivism failed to predict care orientation, supporting my theory of connecting horizontal collectivism with care orientation, and vertical collectivism with family orientation. Although closeness to friends (C1, C11, and C21) failed to predict care orientation, it had higher positive correlations with horizontal collectivism than with interdependent self, as shown in Table 20. The small positive correlations between closeness to friends and care orientation could be due to their correlations with horizontal collectivism. Nonetheless, these correlations provided evidence that higher horizontal collectivism was related to higher closeness to friends, though higher closeness to friends did not lead to higher care orientation.  107  Table 20 Correlation Matrix for Observed Variables at Level 3 for the Student Sample (N = 612, the lower-left) and the Adult Sample (N = 458, the upper-right) Variables W9 W10 W13 S16 S28 S33 M6 M7 M37 C1 C11 C21  W9 .559 .366 .140 .116 .181 .323 .208 .175 .315 .377 .361  W10  W13  S16  S28  S33  M6  M7  M37  C1  C11  C21  .551  .288 .221  .052 .113 .259  .121 .169 .226 .231  .095 .087 .222 .316 .364  .219 .285 .309 .285 .301 .325  .135 .182 .262 .325 .303 .329 .451  .149 .167 .235 .256 .322 .322 .239 .253  .280 .228 .141 .095 .052 .093 .197 .163 .150  .257 .289 .112 .048 .048 .089 .165 .157 .136 .732  .315 .254 .211 .074 .042 .136 .191 .204 .191 .654 .699  .359 .201 .144 .115 .368 .240 .192 .257 .311 .310  .172 .130 .180 .311 .221 .158 .093 .107 .153  .247 .276 .280 .323 .278 .113 .056 .040  .219 .265 .277 .232 .033 .089 .048  .256 .277 .293 -.026 -.012 .005  .406 .291 .130 .134 .137  .263 .185 .159 .163  .092 .071 .064  .699 .666  .681  The above findings suggested that higher horizontal collectivism predicted higher interdependent self, closeness to friends, and care orientation. Higher interdependent self also predicted higher care orientation, though higher closeness to friends failed to predict higher care orientation. The hypothesized relationships at Level 3 were thus partially supported. Predicting Norm-Maintaining Orientation at Level 4 At Level 4, it was hypothesized that vertical worldview (VER) and horizontal collectivism (HC) positively predict interdependent self, closeness to society, and normmaintaining orientation. Then, interdependent self and closeness to society positively predict norm-maintaining orientation. As presented in Table 21, the estimated measurement model was acceptable by most of the fit indices, despite a significant chi-square. The solution for Model B was again not admissible. Model A did not fit worse than the measurement model, as indicated by a nonsignificant Δχ2 (1) = 3.648, critical chi-square = 3.841, p > .05, but the path predicting norm  108  orientation from closeness to society and worldview were both nonsignificant. Thus, Model E was estimated, but the solution was not proper because the standardized path coefficient predicting norm orientation from interdependent self was greater than 1. Neither was the solution for Model C, because two standardized path coefficients were greater than 1. As seen in Table 21, Model F had the same degrees of freedom as did Model A, but higher chi-square, Δχ2 = 5.065. Finally, Model D did not fit worse than the measurement model, as indicated by a nonsignificant Δχ2 (3) = 3.738, critical chi-square = 7.81, p > .05, and seemed to fit better than Model A. Model D had two more degrees of freedom, but lower chi-square than Model A, despite a drop in the SMC for norm orientation by 2%. Model D was chosen because most of the fit indices also favored Model D over Model A. Table 21 Goodness-of-Fit of Alternative Models Predicting Norm Orientation at Level 4 (N = 630) χ2  df  p  χ2 : df  NFI  TLI  CFI  RMSEA  ECVI  Measurement  71.650  41  .002  1.748  .963  .973  .983  .034  .228  Model A  76.816  42  .001  1.829  .960  .970  .981  .036  .233  .62  Model D  76.881  44  .002  1.747  .960  .973  .982  .034  .227  .60  Model  SMC  Model F 81.881 42 .000 1.950 .957 .966 .978 .039 .241 Note. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; SMC = squared multiple correlations for norm orientation; TLI = Tucker-Lewis index.  Table 22 lists the results of the invariance tests on Model D across students and adults. The chi-squares for the invariance tests were all significant, indicating that the model was not invariant across students and adults, although the estimated Model D for adults was acceptable, χ2 (44, N = 472) = 92.153, χ2: df = 2.094, p = .000, NFI = .916, TLI = .930, CFI = .953, RMSEA = .048, and ECVI = .336. However, if the two groups had equal measurement intercepts, they would probably have equal structural weights, as indicated by a nonsignificant Δχ2 (3) = 3.456, p  109  = .327. Besides, many of the fit indices, such as the relative chi-square and RMSEA were within the acceptable range and the ECVIs for the seven models were rather low as compared to the independence model, suggesting model invariance in a weak sense. Table 22 Invariance Tests of Factor Structures for Norm Orientation Between Students (n = 630) and Adults (n = 472) Model  χ2  df  p  χ2 : df  NFI  TLI  CFI RMSEA  ECVI  1. Unconstrained  185.899  89  .000  2.089  .938  .950  .966  .031  .287  2. Measurement weights  211.030  92  .000  2.294  .930  .941  .959  .034  .305  3. Measurement intercepts  461.978  103  .000  4.485  .847  .841  .876  .056  .513  4. Structural weights  465.434  106  .000  4.391  .846  .845  .876  .056  .510  5. Structural covariances  473.490  107  .000  4.425  .843  .843  .873  .056  .516  6. Structural residuals  485.800  110  .000  4.416  .839  .844  .870  .056  .522  7. Measurement residuals  540.920  121  .000  4.470  .821  .841  .855  .056  .552  .000  0  8. Saturated model  1.000  1.000  .280  9. Independence model 3019.884 132 .000 22.878 .000 .000 .000 .141 2.785 Note. Acceptable values are in bold. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; TLI = Tucker-Lewis index.  The final structural model predicting norm orientation for both students and adults is presented in Figure 8, and the correlation matrix for the observed variables for both samples is in Table 23. As can be seen in Table 23, closeness to society had higher positive correlations with norm orientation in students than in adults. HC (horizontal collectivism) had higher positive correlations with norm orientation than did VER (vertical worldview) in students, suggesting that horizontal collectivism was more important than vertical worldview for the development of norm orientation in students, whereas the opposite was true for adults. This may reflect the positive correlation between vertical individualism and closeness to society found in adults in Table 14, because vertical worldview is the mean of vertical individualism and vertical collectivism. Moreover, the indirect effects of closeness to society on interdependent self and norm orientation (.46 and .36 respectively) in students were similar to those in adults (.41 and .37 respectively),  110  but the indirect effects of worldview on norm orientation were larger for adults (.71) than that for students (.61). Given so many group differences, the model failed to achieve full invariance. Table 23 Correlation Matrix for Observed Variables at Level 4 for the Student Sample (N = 607, the lower-left) and the Adult Sample (N = 455, the upper-right) Variables  HC  VER  S1  S7  S33  C5  C15  C25  M13  M31  M39  HC .225 .258 .246 .129 .210 VER .251 .124 .254 .259 .224 S1 .263 .180 .168 .031 .038 S7 .421 .193 .313 .150 .229 S33 .142 .135 .051 .139 .171 C5 .375 .217 .090 .250 .113 C15 .398 .178 .129 .279 .097 .733 C25 .354 .232 .105 .276 .111 .668 M13 .305 .183 .156 .381 .079 .260 M31 .211 .160 .155 .326 .147 .109 M39 .299 .143 .215 .384 .112 .141 Note. HC = Horizontal Collectivism; VER = Vertical Worldview.  .219 .260 .061 .267 .150 .680  .178 .191 .053 .234 .126 .563 .630  .094 .289 .125 .391 .115 .088 .153 .119  .112 .289 .062 .336 .101 .125 .138 .070 .293  .148 .110 .144 .322 .068 .012 .041 .042 .172 .352  .731 .236 .184 .143  .210 .188 .098  .377 .382  .401  Figure 8. Modified model for norm orientation at Level 4. Bold-faced quantities close to the latent variables are their squared multiple correlations. Quantities near paths are standardized path coefficients at p < .001. HC = Horizontal Collectivism; VER = Vertical Worldview.  111  Although Level 4 worldview did not predict closeness to society, it did positively predict interdependent self, which in turn positively predicted norm orientation. Moreover, closeness to society and Level 4 worldview also positively predicted norm orientation, albeit indirectly through the mediation of independent self. The total effect of closeness to society, Level 4 worldview, and interdependent self on norm orientation was moderate, .67. Because the modified model suggested that norm orientation starts with closeness to society that gives rise to Level 4 worldview, the hypothesized relations at Level 4 were only partially supported. Predicting Justice Orientation at Level 5 It was hypothesized that, at Level 5, higher horizontal worldview (HOR) and collectivism (COL) predict higher moral self and justice orientation, and some closeness to nation, which represents nationalism. Moral self in turn positively predicts justice orientation, but closeness to nation negatively predicts justice orientation. As presented in Table 24, the estimated measurement model predicting justice orientation at Level 5 was acceptable, despite a significant chi-square, because most of the other fit indices were within the acceptable range. The alternative models were estimated. The solution for Model B was not proper, because the standardized path coefficient predicting justice orientation from moral self was greater than 1. Both having the same degrees of freedom, Model A was better than Model F, which had much higher chi-square, Δχ2 = 10.709. Model A was also better than Model C because it had one more degree of freedom than the measurement model and Model C, but slightly higher chi-square, Δχ2(1) = .150, p > .05. Model D did not fit worse than Model A, as indicated by a nonsignificant Δχ2(1) = 3.704, p > .05, but the SMC for justice orientation dropped from .66 to .57. The 9% drop in SMC suggested that closeness to nation, mediated through Level 5 worldview and then moral self, decreased rather than increased justice  112  orientation, thus providing evidence for my hunch that nationalism or feeling too close to one‘s nation can have a negative effect on justice orientation. Thus, Model A was the chosen model for invariance tests. Table 24 Goodness-of-Fit of Alternative Models Predicting Justice Orientation at Level 5 (N = 625) χ2  df  p  χ2 : df  NFI  TLI  CFI  RMSEA  ECVI  Measurement  45.594  24  .005  1.900  .980  .982  .990  .038  .169  Model A  45.744  25  .007  1.830  .980  .983  .991  .036  .166  .66  Model C  45.594  24  .005  1.900  .980  .982  .990  .038  .169  .66  Model D  49.448  26  .004  1.902  .978  .981  .989  .038  .165  .57  Model  SMC  Model F 52.865 25 .001 2.115 .976 .977 .987 .042 .173 Note. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; SMC = squared multiple correlations for justice orientation; TLI = Tucker-Lewis index.  The significant chi-squares of the invariance tests listed in Table 25 indicated that the model was not invariant across students and adults, despite acceptable adult model, χ2 (26, N = 462) = 54.162, χ2: df = 2.083, p = .001, NFI = .958, TLI = .961, CFI = .978, RMSEA = .048, and ECVI = .239. If the two groups had equal factor patterns, they would probably have equal factor loadings, as indicated by a nonsignificant Δχ2 (5) = 2.942, p = .709. Many of the fit indices, such as NFI, CFI, RMESA, and ECVI values were all within the acceptable range. Particularly, when compared to the ECVIs of the saturated and the independence models (.199 and 3.297), the ECVIs for the seven models ranging from .207 to .491 were rather small, closer to the saturated model than to the independence model. Because small ECVIs indicate that the model is expected to cross-validate on a new sample, it was expected that these models are likely to be stable in the population and to replicate. That is, there was some degree of invariance across the two samples, despite large chi-squares and relative chi-square values for Models 3 to 7.  113  Table 25 Invariance Tests of Factor Structures for Justice Orientation Between Students (n = 625) and Adults (n = 462) χ2  df  p  χ2 : df  NFI  TLI  CFI RMSEA  1. Unconstrained  96.107  44  .000  2.184  .973  .969  .985  .033  .207  2. Measurement weights  99.049  49  .000  2.021  .972  .973  .985  .031  .200  3. Measurement intercepts  293.494  58  .000  5.060  .917  .894  .932  .061  .363  4. Structural weights  315.072  63  .000  5.001  .911  .896  .927  .061  .373  5. Structural covariances  337.466  64  .000  5.273  .905  .889  .921  .063  .392  6. Structural residuals  345.704  67  .000  5.160  .902  .892  .919  .062  .394  7. Measurement residuals  469.000  76  .000  6.171  .868  .865  .886  .069  .491  .000  0  3541.429  90  Model  8. Saturated model 9. Independence model  1.000 .000  39.349  .000  ECVI  1.000 .000  .000  .199 .188  3.297  Note. Acceptable values are in bold. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; TLI = Tucker-Lewis index.  The final structural model predicting justice orientation for both students and adults is presented in Figure 9. As can be seen in Figure 9, collectivism (COL) was more influential than horizontal worldview (HOR) in predicting Level 5 worldview.  Figure 9. Hypothesized model for justice orientation at Level 5. Bold-faced quantities close to the latent variables are their squared multiple correlations. Quantities near paths are standardized path coefficients at p < .001, except *p < .05, nsp > .05. COL = Collectivism; HOR = Horizontal Worldview.  114  The path predicting closeness to nation from worldview was small, accounting for only 5% of the variance of closeness to nation, suggesting that Level 5 worldview did not lead to strong closeness to nation. The negative small path coefficient predicting justice orientation from closeness to nation provided further evidence for the negative influence of nationalism on justice orientation. However, the effects of worldview and moral self on justice orientation differed between the two samples, which could be the cause of failing the full invariance tests. More specifically, moral self had a total effect of .97 for students and .54 for adults; Level 5 worldview had a total effect of .60 for students and .46 for adults. The nonsignificant path predicting justice orientation from Level 5 worldview indicated that the influence of worldview on justice orientation was mediated mainly through moral self. The correlation matrix for the observed variables is presented in Table 26. As seen in Table 26, collectivism had higher correlations with closeness to nation than did horizontal worldview, suggesting that it was collectivism, not horizontal worldview, predicting closeness to nation. Table 26 Correlation Matrix for Observed Variables at Level 5 for the Student Sample (N = 610, the lower-left) and the Adult Sample (N = 453, the upper-right) Variables  HOR  COL  S27  S35  HOR .495 .334 .339 COL .559 .305 .253 S27 .371 .509 .456 S35 .403 .469 .496 C7 .057 .196 .137 .150 C17 .069 .212 .083 .142 C27 .100 .224 .110 .134 M18 .373 .346 .419 .402 M38 .268 .264 .343 .319 Note. COL = Collectivism; HOR = Horizontal Worldview.  C7  C17  C27  M18  M38  .070 .277 .046 .043  .098 .247 .066 .069 .750  .009 .227 .047 .082 .720 .694  .264 .219 .243 .270 .030 .060 .030  .194 .145 .312 .217 -.005 -.039 -.030 .369  .747 .733 .025 -.018  .774 .050 .011  .072 .009  .433  115  Because collectivism predicted closeness to nation, I investigated whether horizontal worldview alone would have higher predictive power on justice orientation, by using horizontal collectivism and horizontal individualism as observed variables for Level 5worldview. However, the new model was not successfully fitted, because moral self failed to predict justice orientation, suggesting that horizontal worldview alone is not enough to cultivate a moral self that endorses justice orientation and that collectivism is nesscessary for the development of moral self and justice orientation. Level 5 worldview (collectivism and horizontal worldview) had a statistically nonsignificant effect on justice orientation, but a moderate positive effect on moral self and a very small positive effect on closeness to nation. That is, Level 5 worldview indirectly influenced justice orientation through the mediation of moral self and closeness to nation, while the former enhanced but the latter reduced justice orientation. Thus, the hypothesized causal relationships at Level 5 were largely supported. Predicting Biocentric Orientation at Level 6 As hypothesized, at Level 6, collectivism (COL), anti-anthropocentrism (RH; reversed HW scores), and anti-social Darwinism (RSD; reversed SD scores) positively predict ecological self, closeness to nature, and biocentric orientation. Then ecological self and closeness to nature positively predict biocentric orientation. The estimated measurement model was acceptable, as indicated by most of the fit indices listed in Table 27, despite a significant chi-square. Among the alternative models, the solution for Model A was not admissible because worldview and self-view both failed to predict biocentric orientation. Neither was the solution for Model F, because the SMC for worldview was greater than 1. Model D fitted worse than the measurement model, Δχ2 (3) = 45.054, p > .05,  116  and accounted for 20% less of the variance for biocentric orientation than did Models B and C. Model B was better than Model C because it had one more degree of freedom than the measurement model and Model C, but slightly lower chi-square, Δχ2 = .087. Table 27 Goodness-of-Fit of Alternative Models Predicting Biocentric Orientation at Level 6 (N = 640) χ2  df  p  χ2 : df  NFI  TLI  CFI  RMSEA  ECVI  Measurement  199.473  52  .000  3.836  .936  .928  .952  .067  .431  Model B  199.560  53  .000  3.765  .936  .929  .952  .066  .428  .88  Model C  199.473  52  .000  3.836  .936  .928  .952  .067  .431  .88  Model D  244.527  55  .000  4.446  .922  .912  .938  .073  .492  .68  Model  SMC  Model E 133.565 26 .000 5.137 .911 .872 .926 .080 . 297 Note. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit  .99  index; RMSEA = root-mean-square error of approximation; SMC = squared multiple correlations for biocentric orientation; TLI = Tucker-Lewis index.  However, in both Models B and C, the path predicting biocentric orientation from closeness to nature was statistically nonsignificant. Thus, Model E (without closeness to nature) was estimated, but the solution was not acceptable, because the relative chi-square was greater than 5. Thus, Model B was the chosen model to fit on the adult data and to undergo invariance testing. The estimated Model B for adults was acceptable, χ2 (53, N = 472) = 213.971, χ2: df = 4.037, p = .000, NFI = .912, TLI = .899, CFI = .932, RMSEA = .080, and ECVI = .611. As shown in Table 28, the chi-squares for the invariance tests were all significant, indicating that the model was not invariant across students and adults. However, some of the indices were within acceptable range, such as the RMSEA. Compared with the ECVI of independence model (5.057), the ECVIs for the seven models ranging from .531 to .984 did not seem so bad.  117  Table 28 Invariance Tests of Factor Structures for Biocentric Orientation Between Students (n = 640) and Adults (n = 472) χ2  df  p  χ2 : df  NFI  TLI  CFI RMSEA  1. Unconstrained  440.893  106  .000  4.159  .921  .909  .938  .053  .531  2. Measurement weights  457.553  109  .000  4.198  .918  .908  .936  .054  .540  3. Measurement intercepts  853.774  121  .000  7.056  .847  .825  .865  .074  .875  4. Structural weights  871.717  126  .000  6.918  .843  .829  .862  .073  .883  5. Structural covariances  876.382  127  .000  6.901  .843  .830  .861  .073  .885  6. Structural residuals  891.395  130  .000  6.857  .840  .831  .859  .073  .893  1015.852  142  .000  7.154  .817  .823  .838  .074  .984  .000  0  5565.816  156  Model  7. Measurement residuals 8. Saturated model 9. Independence model  1.000 .000  35.678  .000  ECVI  1.000 .000  .000  .324 .177  5.057  Note. Acceptable values are in bold. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; TLI = Tucker-Lewis index.  The final structural model predicting biocentric orientation for both students and adults is presented in Figure 10. As can be seen in Figure 10, the path predicting biocentric orientation from closeness to nature was nonsignificant, despite positive correlations between closeness to nature and biocentric orientation listed in Table 26, . I thus deleted closeness to nature from the model in Figure 10 and found that the SMC for biocentric orientation dropped from .86 to .76, suggesting that closeness to nature had a positive effect (about .10) on biocentric orientation, which was suppressed by the presence of ecological self. Together, ecological self, worldview, and closeness to nature had a total effect of .86 on biocentric orientation, which was large. As seen in Table 29, the correlation between collectivism and biocentric orientation was higher in students than in adults, suggesting that collectivism played a more important role in the development of biocentric orientation in the students than in the adults. Collectivism (COL) and anti-anthropocentrism (RH) had a negligible correlation in students but a negative correlation in adults, whereas collectivism and anti-social Darwinism had a positive correlation in students but  118  a negligible correlation in adults. Given so many differences between the students and adults, it is logical that the model was lack of invariance.  Figure 10. Modified model for biocentric orientation at Level 6. Bold-faced quantities close to the variables are their squared multiple correlations. Quantities near paths are standardized path coefficients at p < .001, except nsp > .05. COL = Collectivism; RH = Anti-Anthropocentrism; RSD = Anti-Social Darwinism. Table 29 Correlation Matrix for Observed Variables at Level 6 for the Student Sample (N = 616, the lower-left) and the Adult Sample (N = 454, the upper-right) Variables  RH  RSD  COL  S23  S30  S34  C8  C18  C28  RH .348 -.185 .135 .118 .047 -.017 -.028 -.013 RSD .303 -.018 .213 .173 .143 .052 .124 .090 COL .032 .125 .084 .078 .093 .169 .143 .150 S23 .139 .011 .170 .571 .610 .556 .519 .537 S30 .110 .062 .274 .505 .564 .520 .477 .541 S34 .084 .027 .272 .574 .602 .522 .509 .506 C8 .114 .023 .229 .496 .509 .501 .797 .784 C18 .084 -.032 .234 .503 .412 .460 .768 .816 C28 .056 .043 .231 .520 .450 .497 .738 .792 M2 .246 .106 .254 .398 .497 .448 .379 .335 .291 M9 .184 .023 .189 .362 .350 .417 .346 .333 .316 M17 .230 .134 .290 .360 .496 .397 .372 .314 .284 Note. COL = Collectivism; RH = Anti-Anthropocentrism; RSD = Anti-Social Darwinism.  M2  M9  M17  .300 .254 .005 .412 .527 .390 .354 .320 .355  -.036 .005 .162 .200 .217 .273 .324 .307 .313 .270  .186 .202 .075 .310 .459 .312 .303 .316 .334 .506 .184  .430 .544  .369  119  Although it was ecological self that predicted Level 6 worldview and closeness to nature, ecological self, Level 6 worldview, and closeness to nature did positively predict biocentric orientation. Level 6 worldview and closeness to nature also positively predicted biocentric orientation, but the effect of closeness to nature on biocentric orientation was mediated by ecological self. Thus, the hypothesized causal relationships at Level 6 were partially supported. Predicting Religious Orientation at Level 7 At Level 7, higher spiritual worldview was hypothesized to predict higher religious self, closeness to God, and religious orientation. Higher religious self and closeness to God were also hypothesized to predict higher religious orientation. As shown in Table 30, the estimated measurement model at Level 7 was acceptable. Alternative models were estimated. With one or two more degrees of freedom, Models A, B, D, and F obviously fitted worse than the measurement model and Model C, due to their large chisquare differences, ranging from 17.05 to 271.52. Model C was the chosen model because it had the same fit indices as the measurement model. Table 30 Goodness-of-Fit of Alternative Models Predicting Religious Orientation at Level 7 (N = 640) χ2  df  p  χ2 : df  NFI  TLI  CFI  RMSEA  ECVI  Measurement  134.058  54  .000  2.483  .983  .985  .990  .048  .322  Model A  265.563  55  .000  4.828  .967  .962  .973  .077  .522  .80  Model B  155.859  55  .000  2.834  .980  .982  .987  .054  .353  .80  Model C  134.058  54  .000  2.483  .983  .985  .990  .048  .322  .81  Model D  405.577  56  .000  7.242  .949  .938  .956  .099  .741  .77  Model  Model F 151.108 55 .000 2.747 .981 .983 .988 .052 .346 Note. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; SMC = squared multiple correlations for religious orientation; TLI = Tucker-Lewis index.  SMC  120  The estimated Model C for adults was also acceptable, χ2 (33, N = 472) = 81.55, χ2: df = 2.471, p = .000, NFI = .946, TLI = .945, CFI = .967, RMSEA = .056, and ECVI = .309. Table 31 lists the results of the invariance tests on Model C across age groups. Although the chi-squares for the invariance tests were significant, all other fit indices were within the acceptable range, providing evidence for model invariance across students and adults. Compared to the ECVIs of the saturated and the independence models, .324 and 13.104, the ECVIs for the seven models ranging from .348 to .474 were considered small, suggesting model invariance across the two samples and even new samples. Moreover, if Model 5 was correct, Models 6 would be invariant, as indicated by a nonsignificant chi-square difference, Δχ2 (1) = .318, p = .573. Table 31 Invariance Tests of Factor Structures for Religious Orientation Between Students (n = 640) and Adults (n = 472) χ2  df  p  χ2 : df  NFI  TLI  CFI RMSEA  1. Unconstrained  234.600  104  .000  2.256  .984  .986  .991  .034  .348  2. Measurement weights  243.749  108  .000  2.257  .983  .986  .991  .034  .349  3. Measurement intercepts  355.831  120  .000  2.965  .975  .979  .984  .042  .429  4. Structural weights  366.361  123  .000  2.979  .975  .978  .983  .042  .433  5. Structural covariances  391.044  129  .000  3.031  .973  .978  .982  .043  .444  6. Structural residuals  391.250  130  .000  3.010  .973  .978  .982  .043  .443  7. Measurement residuals  449.611  142  .000  3.166  .969  .976  .979  .044  .474  .000  0  14497.730  156  Model  8. Saturated model 9. Independence model  1.000 .000  92.934  .000  .000  ECVI  1.000  .324  .000  .288 13.104  Note. Acceptable values are in bold. CFI = comparative fit index; ECVI = expected cross-validation index; NFI = Bentler-Bonett‘s normed fit index; RMSEA = root-mean-square error of approximation; TLI = Tucker-Lewis index.  The final structural model for religious orientation for both students and adults is in Figure 11. As seen in Figure 11, the effects of closeness to God on religious orientation were larger than that of religious self and spiritual worldview on religious orientation. As seen in Table 32, there were strong positive correlations among the observed variables, but more so in adults than in students. These differences might be the cause of the significant chi-squares.  121  Figure 11. Modified model for religious orientation at Level 7. Bold-faced quantities close to the latent variables are their squared multiple correlations. Quantities near paths are standardized path coefficients at p < .001. Table 32 Correlation Matrix for Observed Variables at Level 7 for the Student Sample (N = 609, the lower-left) and the Adult Sample (N = 454, the upper-right) Variables W8 W29 W33 S2 S8 S21 C10 C20 C30 M16 M22 M34  W8 .623 .634 .526 .499 .538 .582 .574 .553 .558 .552 .519  W29  W33  S2  S8  S21  C10  C20  C30  M16  M22  M34  .715  .736 .832  .605 .623 .589  .594 .584 .578 .874  .624 .635 .610 .876 .878  .640 .617 .617 .697 .697 .708  .666 .643 .625 .714 .701 .747 .904  .641 .612 .597 .694 .680 .726 .882 .901  .577 .592 .558 .653 .623 .663 .637 .663 .638  .642 .619 .595 .733 .733 .763 .723 .730 .702 .721  .571 .580 .573 .747 .730 .755 .709 .712 .696 .748 .838  .798 .602 .601 .603 .544 .557 .540 .537 .579 .555  .577 .590 .577 .573 .572 .542 .597 .587 .554  .809 .790 .658 .670 .665 .600 .665 .641  .822 .648 .659 .646 .590 .652 .626  .666 .679 .666 .624 .696 .672  .926 .898 .699 .727 .702  .912 .727 .743 .714  .703 .710 .701  .741 .698  .818  Although there were no predictive relations among spiritual worldview, religious self, and closeness to God, the three latent variables did positively predict religious orientation and  122  accounted for 80% of the variance in religious orientation, which is high. Thus, the hypothesized causal relationships at Level 7 were only partially supported. Testing Individual Path Hypotheses Path 1. Path 1 in the hypothesized structural model in Figure 3 illustrates the hypothesis that worldview shapes self-view. Based on the structural modeling tests, it was considered largely supported because worldview did positively predict self-view at most levels, except Levels 6 and 7. At Level 7, where closeness to God had the largest effect on religious orientation, and spiritual worldview was found to correlate with religious self. However, Model B, in which self-view positively predicts worldview, was chosen only once at Level 6, but not admissible at Levels 2, 3, 4, and 5. In contrast, the hypothesized model, in which worldview positively predicts self-view, was not admissible only once at Level 6. It could be due to many reasons. First, the anti-social Darwinism (RSD) and anti-anthropocentricism (RH) subscales that were used for Level 6 worldview had low alpha coefficient reliabilities, in the middle .60s. Second, the collectivism items (see Appendix A) are more anthropocentric rather than biocentric, because they describe interdependence between people, not between people and other species or nature. Thus, collectivism (COL) had low correlations with RSD and RH, as shown in Table 29. Consequently, Level 6 worldview also had low internal consistency reliability, below .50. Owing to these measurement issues, the current data could not rule out the possibility of worldview positively predicting ecological self at Level 6, if a more reliable measure of biocentric worldview is used. Thus, Path 1 was considered largely supported because worldview positively predicted self-view most of the time. Path 2. Path 2 shows the hypothesis that worldview influences moral inclusiveness. The results of the structural modeling tests indicated that worldview positively predicted relationship  123  closeness at Levels 2, 3, and 5, although closeness to friends failed to predict care orientation at Level 3. Relationship closeness was not assessed at Level 1, but the correlations listed in Table 14 supported the hypothesis of Level 1 worldview predicts little or no relationship closeness, reflecting exclusion of others in moral inclusiveness at Level 1. At Level 6, current data suggested that higher ecological self predicted higher closeness to nature. However, due to the reliability issues, it was still inconclusive to say that worldview does not positively predict closeness to nature at Level 6 when a more appropriate measure of biocentric worldview is used. At Level 7, worldview was found to correlate with closeness to God. Only at Level 4, closeness to society positively predicted Level 4 worldview. Because worldview positively predicted relationship closeness most of the times, but relationship closeness positively predicted worldview only once, Path 2 was considered supported to some extent. Path 3. Path 3 presents the hypothesis that worldview shapes moral orientation. The results of the structural modeling tests fully supported this hypothesis, because worldview positively predicted moral orientation directly at Levels 1, 2, 3, 6, and 7 and indirectly at Levels 4 and 5. Path 4. Path 4 in the structural model in Figure 3 denotes the hypothesis that self-view induces moral orientation. As indicated by the structural modeling tests, self-view positively predicted moral orientation directly at all seven levels, although the path coefficient from interdependent self to family orientation was small. That is, Path 4 was totally supported. Path 5. Path 5 illustrates the hypothesis that moral inclusiveness as represented by relationship closeness influences moral orientation. Relationship closeness was not assessed at Level 1, but the negligible correlations between relationship closeness and egocentric orientation listed in Table 14 supported that Level 1 inclusiveness that includes no others is related to  124  egocentric orientation. Besides, the structural modeling tests indicated that relationship closeness predicted moral orientation directly at Levels 2, 5, and 7 and indirectly at Levels 4 and 6. The small negative path coefficient predicting justice orientation from closeness to nation at Level 5 was particularly telling of the effects of moral inclusiveness on moral orientation. Although closeness to friends failed to predict care orientation at Level 3, Path 5 was mostly supported. Overall, Paths 1 to 5 were supported based on the test results at Levels 1 to 7. Thus, the main hypothesis in this study was also considered supported in general. Summary The current samples were large enough for the purpose of the present study. First, using the student data, the measurement model was fitted with two to three observed variables for each latent construct. Next, the target model was compared to the measurement model and a series of alternative models. The best fitting model was chosen based on fit indices and chi-square difference tests. The chosen model was then fitted to the adult data. To validate the chosen model, invariance of the model across the student and the adult samples was tested. This modeling process was repeated seven times, one at each level of inclusiveness. Results of SEM at the seven levels of inclusiveness largely supported the main hypothesis that worldview shapes self-view, moral inclusiveness, and moral orientation and that self-view and moral inclusiveness also influence moral orientation. There were exceptions, however. Closeness to friends was unrelated to care orientation at Level 3. A mediational model at Level 4, a self-view model at Level 5, and a correlational model at Level 7 were chosen over the hypothesized model. These exceptions and the general results will be discussed in the next chapter.  125  CHAPTER FIVE: DISCUSSION The primary objective of this study was to examine the causal relations among worldview, self-view, moral inclusivess, and moral orientation, as stated in the hypothesized structural model in Figure 3. The second objective of this study was to construct a multidimensional scale, named Moral Orientation Index to measure different moral orientations that reflect the diverse aspects of worldview and self-view at various levels of inclusiveness. The two objectives complemented each other; so did the structural model of moral development and the Moral Orientation Index. The present data generally support the hypothesized structural model that worldview influences self-view, moral inclusiveness, and moral orientation and that self-view and moral inclusiveness also influence moral orientation. However, the relational structures among worldview, self-view, moral inclusiveness, and moral orientation at different levels may be more intricate than previously thought. In this chapter, I first address the limitations of the study and then discuss the implications of the results and future research directions. Limitations of the Study Some limitations of the present study require some reflective consideration. First, all measures were of self-report type. Although some of the associations may be partly due to shared method variance, the relationships obtained in this study are not due to one single higher-order factor relating to participants‘ general response tendencies. As indicated by exploratory factor analyses, the measured variables used in this study are distinct constructs that have good internal structures and satisfactory reliability coefficients. As indicated by the SEM results, different levels of moral development having different relational structures further preclude the conclusion of artificial relationships among these variables resulting from shared method variance, which  126  should produce similar relationships at each level. Moreover, worldview, self-view, relationship closeness, and moral orientation are all very subjective experiences that only the participants themselves could know best, making triangulation with other self-report instruments somewhat senseless. Besides, self-report instruments may be more reliable than face-to-face interview, which is more susceptible to social desirability responding (Martin & Nagao, 1989). Second, the data were cross-sectional, which prevents one from drawing causal conclusions. Fortunately, I could rely on SEM analyses to support directional relations among the variables. By comparing the hypothesized model with the alternative models that represent diverse and even contradictory views to the main hypothesis at each level of inclusiveness, the SEM results provide strong evidence for pseudo-causal relations among the constructs at most levels of inclusiveness. Third, although the student sample was balanced in terms of age, gender, grade-level, ethnicity, and religion, the adult sample skewed heavily toward females and higher education, particularly in the education faculty. It is obvious that the current adult data do not represent adult development in general, thus limiting the generalizability of the findings of the adult data. The differences between the students and adults, as discussed in the following, should be considered differences in education level rather than developmental differences. Given the vast differences between the two samples, the structural models developed with the student data still fitted well with the adult data provide further evidence for the validity of the models, notwithstanding the stringent invariance tests. Despite all these limitations, the current study has provided an important first empirical test of the causal relations among worldview, self-view, moral inclusiveness, and moral orientation for the first part of the comprehensive model of moral development. There is also  127  evidence for the hypothesized specific aspects of worldview, self-view, moral inclusiveness, and moral orientation at each of the seven levels of moral inclusiveness. The Implications of the Results The Multiplicity of Moral Orientation As the centerpiece of the comprehensive model of moral development, moral orientation also plays a crucial role in other theories of morality. Many theorists have distinguished between spontaneous, habitual, intuitive morality and deliberate, infrequent, reflective morality (e.g., Davidson & Youniss, 1991; J. S. Walker, 2000). The former, which is habitual, immediate, and nonmediated thought, ensues from the interplay between the individual and his or her social experience and is ―universal (i.e., equivalent across persons given equivalent social stimulus event)‖ (Davidson & Youniss, 1991, p. 120). As interpretive habits, it plays a critical role in the interpretation of moral situations and in guiding most of our social behavior. In contrast, the latter, involving thorough appraisal and rationalization, comes into play relatively infrequently, when two or more interpretations might fit the situation (J. S. Walker, 2000). D. Hart and Killen (1995) concur that even Kant, Piaget, and Kohlberg recognized the existence of habitual moral thoughts and reflexive action and agreed that habit and reflective judgment can be part of a single dialectical process. How the two are involved in the process and how to examine their relations empirically is a significant and necessary but difficult task (p. 10). By definition, moral orientation represents the former and moral judgment the latter, which is then connected to the former as situational moral orientation. The proposed comprehensive model of moral development is the first attempt (as far as I know) to delineate the roots (worldview, self-view, and social relationships) of the former (moral orientation) and its relations with the latter (moral judgment) and behavior. As indicated by the data, moral orientation is  128  multidimensional, reflecting different aspects of worldview, self-view, and social relationships. To understand human morality, we must first understand the multiplicity of moral orientation. The distinctiveness of egocentric, family, and care orientations. The present research endeavored to distinguish family orientation from care orientation and egocentric orientation while the DIT data of Rest et al. (1999) suggest that the traditional Kohlbergian Stages 2 and 3 items fused into one factor, namely, the primitive Personal Interest schema. The current data, however, provide evidence for the distinctiveness of egocentric orientation, family orientation, and care orientation, because these orientations are affected by and reflect different worldviews, self-views, and moral inclusiveness. First, egocentric orientation arises from vertical individualism, social Darwinism, and independent self, whereas care orientation stems from horizontal collectivism and interdependent self, and family orientation is mainly the result of vertical collectivism and closeness to family. Second, justice orientation being correlated positively with family orientation and care orientation, but negatively with egocentric orientation in both samples (see Table 11), confirms that family orientation and care orientation as defined in this study are not the same as the personal interest schema of Rest et al. (1999). Finally, egocentric orientation is correlated positively with independent self (see Table 14) but negatively with interdependent self (-.02 for students and -.13 for adults). Quite the opposite, care orientation is correlated strongly with interdependent self (.57 and .61) but weakly and even negatively with independent self (.16 and -.07). Unlike care and egocentric orientations, family orientation‘s correlations with interdependent self (.31 and .37) do not differ much from its correlations with independent self (.26 and .13). These correlations suggest that individuals endorsing egocentric or care orientation higher than family orientation are less  129  balanced or mature in their personality and development than are individuals who endorse family orientation higher than egocentric and care orientations and that egocentric and care orientations are less sophisticated than is family orientation, according to the complementary perspective. These findings not only verify the distinctiveness of egocentric, family, and care orientations, but also provide preliminary evidence for the validity and importance of family orientation. The findings of interdependent self as the stronger predictor than horizontal worldview for care orientation in both samples concur with research (e.g., Lyons, 1983; Pratt et al, 1988; Skoe & Marcia, 1991), supporting Gilligan‘s (1982) assertion of the relations of interdependent self and care. However, it is counter-intuitive that the influence of interdependent self on family orientation was weak in the student sample and insignificant in the mostly female adult sample. Perhaps a different self-concept, such as a family person or self, to be measured with items such as ―I enjoy family life‖ and ―I am a family person,‖ may be more relevant than the concept of interdependence for Level 2 moral orientation. If family life is the molding ground of social relationships and personality (c.f. Bowlby, 1969), a family self could be the precursor of an interdependent self. The finding of relationship closeness with friends having no influence on care orientation is contrary to the peer-group socialization theory that peer relationships are the most influential sources shaping individuals‘ personality and development (Piaget, 1932/1965; J. R. Harris, 1998). This finding occurring in both samples that differed in age and education level seems compelling and demands explanation. I speculate that first, care orientation may be seen as part of the individuals‘ self-concept or identity (e.g., I am a caring person), as suggested by the strong correlations between interdependent self and care, so strong that caring individuals will practice care orientation regardless of relationship closeness with friends and peers.  130  Second, Leman (2002) reported that the cognitive-developmental benefits of peer interaction might not accrue when children often influence their peers to accept their less advanced position by persistently returning to it. When a more advanced position is adopted, the acceptance of the more advanced position seems to be a consequence of an appreciation of the legitimacy of that position rather than a consequence of vigorous arguments. That is, children accept the advanced position because they find something about it convincing and credible (the truth, I guess) and there is no need to argue for it. Finally, peers do play an important role because individuals‘ concept of their strategies is dependent on their social acceptance (Keltikangas-Järvinen, 2001). However, the outcomes of friends‘ influence can be positive or negative, depending on the friends‘ characteristics (Berndt, 1999). As a result, the effects of relationship closeness with friends and peers on care orientation can be neutralized or minimized. Quite the reverse, the positive influence of family interaction, including parental warmth and support, on child and adolescent development has been consistently demonstrated by researchers (e.g., Palmer & Hollin, 1996, 2001; Speicher, 1992, 1994; White & Matawie, 2004). If a family self is the precursor of an interdependent self, family orientation (e.g., self-sacrifice, commitments, etc.) could be the precursor of norm orientation, justice orientation, and so forth, because morality requires sacrificing self-interest or the self for others and the greater good. Thus, family orientation marks the beginning of morality and vertical collectivism is the foundation of interdependence. Although family is the breeding ground of morality, it has been neglected by researchers, as Ochs and Kremer-Sadlik (2007) lamented. This neglect may be in part due to the overemphasis on peer relations as the foundation of moral development by the cognitive approach (e.g., Piaget, 1932/1965) and the care-justice debate in the last few decades.  131  Differentiation between family and care orientation is thus important in facilitating investigation into the influence of family on moral development. However, extreme family orientation can deter moral development by limiting perspective taking; extreme forms of care such as self-silencing and submissiveness can become signs of pathology rather than moral maturity (cf. Hennig, 2004). Likewise, aggression, if regulated and balanced with cooperation and consideration, can be adaptive and functional. The individualism-independence-egocentrism configuration found in the current highly educated, mostly female adult sample seems to attest to the pervasiveness of egocentric orientation. It also suggests that egocentrism need not deter moral development if it is balanced with collectivism-interdependence based moral orientations, such as family and care orientations. That is, not only are vertical relationships with parents and horizontal relationships with peers both important for the development of moral maturity (L. J. Walker, Hennig, & Krettenauer, 2000), but vertical individualism can also contribute to moral development, if it is balanced with collectivism, as can be seen in the development of norm orientation. The distinctiveness of norm and religious orientations. The current study also endeavored to separate religious orientation from norm orientation. Most social institutions, including religious communities, are vertical in structure and collectivist in nature. Thus, a religious orientation could be a form of norm orientation that aims at maintaining the religious norms (Rest et al., 1999) and was treated as such in the Defining Issues Test (Rest, 1993). The present data, however, indicate that norm and religious orientations are distinct concepts reflecting different aspects of worldview, self-view, and relationship closeness, and having diverse configurational patterns.  132  Regarding the formation of norm orientation, the current data support a mediational model that starts with closeness to society (school or workplace), which promotes a balanced worldview that acknowledges the hierarchical structure of most social institutions and the collective spirit within. Such a balanced worldview mediated through interdependent self-view gives rise to norm orientation. It seems logical that norm orientation starts with closeness to society, because the incentive to follow the rules and maintain the norms of a social institution or community will be low if one does not feel close to or want to stay in the community. Again, norm orientation having higher correlations with closeness to society (.25 for students and .16 for adults) and closeness to nation (.14 and .18) than with closeness to friends and peers (.09 and .10) is contrary to the peer-group socialization theory (Piaget, 1932/1965; J. R. Harris, 1998). These correlations suggest that other people in the school, such as teachers and principals play a more significant role in the students‘ moral development than the peer groups do. This suggestion is in accord with Ianni‘s (1989) conclusion after a 12-year study of American youth that an adolescent‘s social conduct is influenced by the adults (e.g., parents, teachers) who share a set of common standards and expectations for the youth‘s behavior. The correlations also suggest that the binding force for moral development at Level 4 comes from relationship closeness with a collective or community beyond the context of interpersonal relationships. According to the holistic and complementary principles, to strike a balance between polarities and to integrate elements from previous levels is a necessary step in human and moral development. As indicated by the findings at Level 1, individualism as a whole is not to blame for egocentric orientation or moral decay, but social Darwinism or extreme individualism to the exclusion of others in moral inclusiveness. When vertical individualism is balanced with vertical collectivism at Level 2 and horizontal collectivism at Level 3, it results in norm orientation at  133  Level 4, thus bearing witness to the holistic and complementary stance that each level of development builds on the previous one. Combining vertical worldview with horizontal collectivism, Level 4 worldview represents a more balanced worldview and an opportunity to integrate worldviews of the first three levels. The greater effects of worldview on norm orientation found in adults than in students suggest that the adults may have developed a more balanced Level 4 worldview than the students may. The positive correlation of closeness to society with vertical individualism was higher in adults than in students (see Table 14), reflecting the entrenched vertical or competitive spirit of our society. With regard to religious orientation, the SEM results point towards a correlational model, in which spiritual worldview, religious self, and closeness to God have reciprocal or bidirectional rather than unidirectional relations, despite their unidirectional relations with religious orientation. Moreover, closeness to God has the largest positive effect and spiritual worldview has the least positive effect on religious orientation. These findings seem to concur with the literature that spirituality is a natural human quality that develops from the beginning of an individual‘s life (e.g., T. Hart, 2006; Hay et al., 2006) and may precede a cognitive appreciation of the spiritual world. These findings may also be due to some measurement issues. For example, different religions have different views of the spiritual world. The current use of just a few items that are not specific to any religious traditions may risk being too general to accurately assess the kind of spiritual worldview that is specific to a religion, thus underestimating the influence of spiritual worldview on closeness to God, religious self, and religious orientation. In any case, feeling close to some entity of supernatural power and wisdom may be sufficient to induce an individual to follow that entity‘s advice in resolving problems, because of the comfort the relationship closeness to something greater than the self may bring.  134  Norm orientation and religious orientation not only have different structural models but also have different correlations with other orientations, although the two are positively correlated (Table 11). Their positive correlations became practically insignificant (.10 for students and .09 for adults), when family orientation was partialed out, suggesting that their correlations were mainly due to a third variable, family orientation. Moreover, justice orientation has positive correlations with norm orientation but not with religious orientation in either sample, suggesting that norm orientation is based on social relations, but religious orientation is based on relations with some other worldly higher beings or laws. All these differences between norm and religious orientations further support that they are different constructs reflecting different mindsets. Religious orientation having negligible correlations with justice and biocentric orientations, but positive correlations with family orientation (the highest), then norm and care orientation in both samples (see Table 11) suggests that most religious teachings emphasize family values, obedience, and sympathy, but not justice or environmental values. These correlations are thus comparable to the extant research that many religious people are willing to help family members or persons close to them, but not outgroup members (e.g., Batson, Floyd, Meyer, & Winner, 1999; Goldfried & Miner, 2002). Without embracing Level 5 and Level 6 inclusiveness, these individuals cannot be regarded as religious because the genuineness of religious life is rooted in its moral helpfulness, as James (1902) suggested (L. J. Walker, 2003). The distinctiveness of justice and biocentric orientations. The current findings of positive correlations between justice and biocentric orientations in both samples (see Table 11) provide evidence for the continuity and overlap between justice and biocentric orientations. These findings are also consistent with the report of Clayton (2000) on the emerging issue of environmental justice (e.g., responsibility to other species and future generations, and the rights  135  of the environment) in resolving environmental conflicts. The findings of both justice and biocentric orientations having negative correlations with egocentric orientation in both samples (see Table 11) further suggest that justice and biocentric orientations are more morally advanced than other orientations. Not only do justice and biocentric orientations have higher positive correlations with each other than with other orientations, but moral self and ecological self also have higher positive correlations with each other than with other self-views in both samples. Together, the above findings suggest that Level 5 and Level 6 inclusiveness, as reflected by justice and biocentric orientations respectively, do not represent independent domains, but progressively enlarging circles of moral inclusiveness within the moral domain. While Level 6 inclusiveness is the widening of Level 5 inclusiveness from the inclusion of all human beings to the inclusion of all living and non-living beings, biocentric orientation is the extension of justice orientation to cover all life forms, including humans, and the earth. Despite having common ground, justice orientation and biocentric orientation reflect different aspects of worldview, self-view, and moral inclusiveness that support their distinctiveness. With regard to the structures of justice orientation, the SEM results fully support my theory of moral development at Level 5 that individuals must first establish a horizontal worldview founded on collectivism and yet be able to regard themselves as moral beings that are not confined by their social relationships or nationality to adhere to their justice orientation. In other words, moral self has a positive direct influence and mediates the influence of Level 5 worldview on justice orientation, but closeness to one‘s nation has a negative influence on justice orientation, as supported by the current data. Focusing on treating all people equally and fairly, the current justice orientation thus departs from Gilligan‘s (1982) ethic of justice that is characterized by objectivity, rationality, and  136  separation and focuses on rules, duties, rights, and reciprocity. Defining the ethic of justice in so many different terms makes the concept of justice orientation multidimensional rather than unitary (Yu & Kishor, 2001b). For example, reciprocity and separation seem to reflect elements of egocentric orientation at Level 1; rules and duties are features of norm-maintaining orientation at Level 4; and impartiality and rationality may be more characteristic of justice orientation at Level 5. To put elements belonging to different orientations together is unwise because they reflect different worldviews. For instance, collectivism is correlated negatively with egocentric orientation (-.15 for students and -.14 for adults), but positively with norm orientation (.41 and .40) and justice orientation (.32 and .17). On the contrary, vertical individualism has high positive correlations with egocentric orientation (.43 and .51), low positive correlations with norm orientation (.02 and .15), but negative correlations with justice orientation (-.15 and -.24). The above correlations suggest that rules and duties, objectivity and rationality, and rights are based on collectivism rather than individualism. Justice orientation having higher correlations with moral self (.46 for students and .42 for adults), but lower and somewhat equal correlations with independent self (.21 and .16) and interdependent self (.21 and .12) suggests that individuals endorsing justice orientation highly may have achieved a balance between independence and interdependence. The current findings are therefore contrary to Gilligan‘s (1982, Gilligan & Attanucci, 1988) theory that respecting the rights of others requires a separation identity, because a separation identity or independent self can be too self-centered to be objective. Regarding the relational structures of biocentric orientation, the present data indicate a self-view model that begins with ecological self, which has a positive influence on Level 6 worldview (i.e., collectivism combined with anti-anthropocentrism and anti-social Darwinism), closeness to nature, and biocentric orientation. Level 6 worldview and closeness to nature also  137  have a positive influence on biocentric orientation, but the influence of closeness to nature on biocentric orientation is mediated through ecological self. As argued previously, the failure of Level 6 worldview to predict ecological self may be due in part to the anthropocentric collectivism scale failing to tap the kind of primitive beliefs of an ecological worldview about the nature of the earth and humans‘ humble relationship with it (Dunlap et al., 2000), in addition to low alpha coefficient reliabilities. Thus, it must await future research to determine the structural model for biocentric orientation. Collectivism playing a role in Level 6 worldview in shaping biocentric orientation corroborates the continuity between Level 5 and Level 6 inclusiveness. It seems to resonate with Schultz‘s (2000) argument that environmental concerns are rooted in a person‘s connection with natural environment and other people. Opotow (2003), who developed the Environmental Identity (EID) Scale, also reported that the EID score was correlated positively with horizontal collectivism and negatively with vertical individualism in college students. In the present study, both ecological self and biocentric orientation are correlated negatively with vertical individualism and positively with horizontal collectivism as well as horizontal individualism in both samples. These correlations further confirm that vertical individualism differs from horizontal individualism for the former may deter, but the latter may foster the development of justice and biocentric orientations. Because a committed ecological identity is not yet supported by society at large (Zavestoski, 2003), horizontal individualism may help individuals hold on to their ecological self and biocentric reasoning. Thus, horizontal individualism may be involved in Level 6 worldview, although it was not included in Level 6 worldview. The positive correlations of horizontal individualism with ecological self and biocentric orientation further verify the continuity between Level 5 and Level 6 development.  138  The current data also concur with extant literature that biocentric orientation can emerge in three ways: an ecological worldview (Dunlap et al., 2000), ecological identity (Kempton & Holland, 2003; Zavestoski, 2003), and relationship closeness with nature (Kahn, 2003) or inclusion of nature in our moral community (Opotow, 2003; Schultz, 2000). The positive influence of closeness to nature on biocentric orientation mediated through ecological self seems to imply that ecological self can emerge through interaction with the natural environment (G. Myers & Russell, 2003). By incorporating all three in the same study, this research was able to provide preliminary data regarding the relational structures among them. Although Level 6 is in theory the peak of secular moral development, biocentric orientation alone may not represent the peak of secular moral development. Research has shown that people of all ages can express pro-environmental statements and behavior, but out of different reasons, such as biocentric, anthropocentric (e.g., Kahn, 1996, 1997; Jurin & Hutchinson, 2005; Kortenkamp & Moore, 2001), and even egocentric orientations (e.g., for personal health; Clark, Kotchen, & Moore, 2003). Thus, anthropocentrism could work for rather than against biocentrism, when people realize that the damage uncontrolled human activities have caused the earth will finally cause them their wellbeing. Moral development is thus a complicated, multifarious process. To unearth the underlying reasons for individuals‘ moral orientation and behavior, we have to look deeper into their worldview, self-view, and moral inclusiveness. From “Is” to “Ought” Reflecting the principles of individual and group selection, individualism and collectivism provide the rationale for moral inclusiveness and moral orientation. As indicated by the current data, individualism corresponds to Level 1 moral inclusiveness (self only) and  139  predicts egocentric orientation, whereas different levels of collectivism coincide with moral inclusiveness and moral orientation from Levels 2 to 7. Although collectivism was not used as an indicator for Level 7 worldview, spiritual worldview is correlated positively with collectivism (.32 for students and .24 for adults). Besides, closeness to God has the second highest positive correlation with collectivism (.31 and .26), next to that with spiritual worldview, but is correlated negatively with individualism (-.10 and -.17), supporting the notion of expanding moral inclusiveness at each successive level based on collectivism. Derived from worldview and giving substance to moral orientation, moral inclusiveness is an important concept that has been discussed by theorists since the 19th century (e.g., Carter, 1980; James, 1897; Schultz, 2000) and supported by recent empirical research (e.g., Reed & Aquino, 2003; Reimer & Wade-Stein, 2004). The present data have provided further evidence to justify moral inclusiveness as a legitimate psychological construct and the Circles of Relationship Closeness (CRC) as an acceptable measure of moral inclusiveness. As a pictorial measure of relationship closeness with different groups, the CRC subscales have high internal consistency reliabilities and meaningful correlation patterns with different aspects of worldview and moral orientation at different levels of inclusiveness. The overlapping circles representing self and other group seem to capture well the essence of moral inclusiveness, whether the motivation behind moral inclusiveness is self-expansion (Aron & Aron, 1986) or self-reduction. The current study has demonstrated that diversity in moral inclusiveness and moral orientation can be examined in reference to diversity in worldview, operationalized as different aspects of individualism and collectivism. Diverse worldviews then affect differentially our interpretations of the moral meaning of social facts and values, which in turn influence whom we choose to include in our moral consideration and what moral orientation(s) we favor. For  140  example, as indicated by the current data, individuals holding stronger beliefs in vertical individualism and social Darwinism (reflecting the principle of individual selection) tend to have little or no relationship closeness with any social groups and to endorse highly egocentric orientation because looking out for the self is seen as the right thing to do. In contrast, for those who believe in higher levels of collectivism, regard for the welfare and integrity of the social groups of which they form a part tends to take precedence over regard for self and others. Thus, family and care orientations require a higher regard for family and friends than self-interest, whereas norm or justice orientation requires a higher regard for the society or humanity than for oneself and one‘s family or nation. The current findings thus fully reveal the significant influence of worldview on moral inclusiveness and moral orientation. Moral orientation will in turn interact with situational factors and influence our moral judgment and behavior, as described in the comprehensive model of moral development in Figure 2. Worldview helps us interpret the situation in terms of what is important morally and what actions should be taken. What ―ought‖ to be done is therefore derived from what the world ―is‖ to us. This ―is-ought‖ association found at each level of inclusiveness makes explaining the genesis of morality easy and predicting individuals‘ behavioral decisions that they make in their everyday lives possible. My discussion on the relation of ―is‖ to ―ought,‖ however, differs from Kohlberg‘s (1981). Defining moral development as the progress of moral judgment from lower to higher stages, Kohlberg (1981) considered moral judgments as statements of value or morality and warned against the fallacy of conflating descriptions of fact and value. Kohlberg suggested that the ―is-ought‖ confusion occurs only at the lower stages and that stage 6 judgments must hinge on moral principles that are independent of factual beliefs.  141  Indeed, moral judgment can be divorced from factual beliefs especially when it is assessed with hypothetical dilemmas that render the individuals‘ worldviews, self-views, moral inclusiveness, and moral orientations irrelevant. Divorcing the ―ought‖ from the ―is‖ also makes predicting the individuals‘ daily behavioral decisions impossible, as indicated by the research findings of respondents scoring higher on hypothetical moral dilemmas, but lower on real-life dilemmas (e.g., Carpendale & Krebs, 1995; Krebs et al., 1991; L. J. Walker et al., 1987; Wark & Krebs, 1996, 1997). Defining moral development as the progress toward greater moral inclusiveness (James, 1897), I regard moral inclusiveness and moral orientation as descriptions of value and moral judgment as situational moral orientation that can be used to decide, rationalize or justify one‘s behavioral decisions under specific situations (cf. Bandura, 1991). Defining morality as the enlargement of moral inclusiveness seems to be a more adequate definition of morality than defining morality as the progress of moral judgment. By bridging the ―is‖ and the ―ought,‖ the former satisfies both the philosophic concern about what truly should be and the scientific concern of predicting and explaining moral behavior effectively — the two basic requirements for an adequate definition of morality, according to Kohlberg (1981, 1984). Whereas, the latter fails to satisfy either of the requirements by divorcing the ―ought‖ from the ―is.‖ Bridging the ―is‖ and the ―ought,‖ the present thesis seems to draw a picture of moral relativism that people, even within the same cultures, have somewhat different worldviews or ideas about the desirable states of moral functioning for human beings. At the same time, however, finding the prescriptive norms for development‘s end state in any society is also possible, from a holistic and complementary perspective. Derived from more holistic worldviews, Level 6 and Level 7 inclusiveness reflect the greatest circles of inclusiveness and  142  thus represent the end states of secular and religious moral development, respectively. The question of ―why be moral?‖ in an immoral world (Kohlberg & Ryncarz, 1990) becomes irrelevant in a holistic view of the world where duality ceases to exist. When all is one, there will be no ―self‖ to be sacrificed amid all the altruistic acts. Before then, however, the self is very much alive, reflecting one‘s worldview. Worldview and Self-View The current data support not only the diversity of worldview, but also the multiplicity of self-view, reflecting the multidimensionality of social relationship and moral orientation. For example, interdependent self is not a satisfactory concept for relationship closeness with family to predict family orientation, suggesting that dissimilar social relationships require dissimilar self-views to deal with the unique social situations. As the SEM results indicated, worldview generally positively predicts self-view, validating the common belief of worldview influencing self-view (e.g., Markus & Kitayama, 1991; L. J. Myers et al., 1991). The correlational relations among spiritual worldview, religious self, and closeness to God found at Level 7 do not refute such belief, but exemplify the reciprocal relations among worldview, self-view, and relationship closeness, as predicted by the holistic perspective. Two interesting patterns of results emerged from the data. The first is that self-view usually has greater influence than worldview on moral orientation, except at Level 2, where interdependent self does not seem to be a relevant concept for family relationships. The second is that the influence of worldview on moral orientation is usually mediated through self-view, except at Levels 2, 6, and 7. The indirect and hidden influence of worldview on moral orientation, in contrast to the direct and apparent influence of self-view on moral orientation,  143  seem to authenticate the individualism-collectivism elements as embodying the distal, cultural influence of ―worldview‖ on moral orientation and the distinctiveness of the constructs of worldview and self-view. These patterns also seem to back up the importance of self-concept or identity in moral development (e.g., Blasi, 1984; Clayton & Brook, 2005; Davidson & Youniss, 1991), but may also reflect our society‘s absorption with the self. An overemphasis on the self makes a sense of oneness or selflessness that signifies authentic Level 7 non-egoistic and nondualistic development even more difficult to achieve. Research has supported that identity can be a source of moral motivation (e.g., Aquino & Reed, 2002; Reed & Aquino, 2003; Reed, Aquino, & Levy, 2007) and inform our decisions about work, religious and social group affiliations (Zavestoski, 2003), and civic actions (Kempton & Holland, 2003). However, research has also indicated that moral identity does not predict later community involvement among adolescents (Pratt, Hunsberger, Pancer, & Alisat, 2003). Similarly, ecological identity can fail to predict environmental behavior because people have multiple identities that carry diverse import from one situation to another (Kahn, 2003). People‘s self-views are very personal, about who they are as persons. It is not surprising that most people want to have a positive self-image and may have ideal selves that are different from their actual selves. Owing to this possible discrepancy between ideal and actual selves, association between self-view and behavior can be further downgraded. People also have multiple worldviews for different social situations, but multiple worldviews can be summed up or reduced to an overarching individualistic or collectivistic worldview. Moreover, people‘s worldviews are only about the world around them and do not carry the kind of significance as self-view or identity, so there is no need for them to profess a worldview that is better than the one they truly embrace. When a person‘s self-view is  144  incongruent with his or her worldview, such as a moral self in an immoral world, it will be difficult for the person to hold on to his or her moral self and to act morally consistently. Thus, worldview is a more reliable predictor of behavior than is self-view. Theorists and educators may be better off to rely on the connection to a more holistic worldview rather than a moral or ecological self to fill in the missing step between moral judgment and action. Implications for Education On the basis of the current results and the extant literature, I recommend three approaches to enhance higher-level moral orientations, particularly justice and biocentric orientations. The first focuses on worldview; the second on self-view; and the third on moral inclusiveness, because current data support the joint influence of worldview, self-view, and moral inclusiveness on moral orientation. As indicated by the present data, worldview influences self-view, moral inclusiveness, and moral orientation, while self-view and moral inclusiveness also influence moral orientation. The primacy of worldview has important implications for moral education, because worldviews are not yet firmly formed during adolescence. Besides, a change in worldview could have a powerful transformative effect on individuals‘ personality and moral behavior (Yu, 1998), as I reported in Chapter 1. Moreover, changing one‘s worldview may be easier than changing one‘s self-view, as argued above. Thus, moral education should aim at educating and creating the kind of worldview that is most conducive to moral development. Of all the worldview variables, collectivism seems to have the most positive influence on moral development, whereas social Darwinism, vertical individualism, and anthropocentrism seem to exert more negative than positive influence on moral development, as indicated by the current data. In our Western, individualist cultures that stress competition, promoting  145  collectivism becomes even more critical for a more balanced development, according to the holistic and complementary point of view. As suggested earlier, people‘s worldviews may be influenced by numerous sources, including education, teachers, peers, and school practice. To promote collectivist, biocentric worldview and to facilitate downward changes in social Darwinism, vertical individualism, and anthropocentrism, worldview education may be desirable. Education and debate on worldview theories can be taken up by teachers of all disciplines, particularly teachers of science and social studies. Students need to understand that many worldviews, such as individualism and religious worldview, espoused by our culture are only perceptions of reality and not necessary the true reality. The focus of worldview discussion is not on changing worldview or behavior, but on critical and reflective thinking, so that students could understand how different worldviews promote different values and influence our interpersonal relationships and moral orientation differentially. It is the hope that through science education, students can learn the interconnectedness of all beings, so that they can form a scientifically compatible collectivist worldview that is not only ecologically friendly and spiritual but also morally responsible to humanity. Schools can help create environments that promote collectivist and horizontal worldviews, for example, by emphasizing the collective over the individual and cooperation over competition. Research has indicated that an emphasis on interpersonal competition exacerbates individuals‘ self-centeredness and results in reduced achievement and alienation from school (Nicholls, 1989). Thus, hypercompetitive attitude must be curbed. To restore healthy competition, competition should be construed as competition against one‘s own past performance or skill levels (i.e., self-growth development) rather than against others; learning from the competition process must be emphasized over winning.  146  To promote collective spirit, cooperative learning may be more desirable than individualistic and competitive learning, especially for our individualistic and pluralistic society. When students acquire academic knowledge and skills from talking, working, and cooperating with each other in a group, they may also get to know more about each other‘s cultural values and practices and try to find a common ground for the group. Thus, cooperative learning programs have been found to be successful in facilitating cross-race peer interaction and reducing prejudice (Slavin & Cooper, 1999) and should be incorporated into the classrooms to complement individualistic and competitive learning. The goal of creating a collectivistic school culture is to foster a collective sense of self (the ―We‖ rather than the ―I‖) that one is not loyal to another individual but a social group, which unites individuals and members of the community with common ends. It is because closeness to society rather than closeness to friends and peers has a more positive influence on norm orientation, as indicated by the data. Such a collective spirit engenders greater moral inclusiveness and a sense of collective or social responsibility in the self or moral self. This collective spirit can then be extended to cover all human beings, other species, and nature. Aronson (1993) suggests that ―action changes consciousness of the self and the other objects in the situation, and thus in turn affects future actions‖ (p. 76, emphasis in original). Research has also indicated that community involvement leads to subsequent moral self (Pratt et al., 2003) and that ecological self develops out of action, and then the two build on each other in a positive feedback loop (Kempton & Holland, 2003). Thus, one way to fortify moral self and ecological self is to organize volunteer service groups or programs to offer opportunities for students to help people in need, to visit sick people and the elderly, to plant trees in the community, to clean up the beaches and so on.  147  Finally, research has supported that taking the other‘s perspective may reduce the degree of separation that a person perceives between him or herself and the other (Schultz, 2000) and that role-playing trees and animals allow students to feel empathy with trees and animals (Zavestoski, 2003). Thus, to expand moral inclusiveness at Levels 5 and 6, students can be encouraged to role-play, for example, gay or lesbian persons, racial minorities, individuals with disabilities, animals in pain or captivity. By empathizing with these individuals and animals, students may come to understand the predicaments these individuals and animals are facing and identify with them. The process of empathy and identification is likely to expand the students‘ moral inclusiveness to embrace these individuals and animals. Besides, field trips to experience nature first hand can enhance feelings of closeness with nature and a sense of ecological self. Future Research Directions This study explored the relational structures among worldview, self-view, moral inclusiveness, and moral orientation, but left out moral judgment and behavior. To assess the interplay between moral orientation and moral judgment on behavior, future researchers are encouraged to incorporate moral judgment and behavior in the investigation. The incorporation of interviews, open-ended or (semi-)structured, in future research is also recommended for it can assist in validation efforts of the current paper-and-pencil measures and quantitative results. Despite evidence supporting small correlations between stages of moral judgment and measures of moral behavior, Krebs and Denton (2006) contended that these correlations could be due to a third factor. As suggested by the comprehensive model of moral development, moral orientation can be a possible candidate for such a third factor, while worldview, self-view, and moral inclusiveness may have indirect influence on moral judgment and behavior as well. Because of the strong influence of independent self in the Western cultures, even individuals  148  with the highest moral self and justice orientation will also endorse egocentric orientation to some degree. After all, the self is also part of the equally just and fair system that should not be neglected. Moral behavior is thus the sum of these conflicting orientations in real-life; moral judgments can be justifications of our behavior. Thus, research on moral judgment and behavior can be based on the whole or the second part of the comprehensive model of moral development. Either way, the results will probably reveal the conflicts between competing moral orientations across people and time. As suggested earlier, the current measure of Level 6 worldview failed to tap the kind of ecological worldview about the earth and humans‘ humble relationship with it, thus, leading to a self-view model for biocentric orientation. Likewise, the current measure of Level 7 worldview could be too simplistic, thus leading to a correlational model for religious orientation. Future research can devise better measures for ecological and spiritual worldviews, so that the relational structures for biocentric orientation and spiritual worldview can be clarified. Researchers may also be interested in the effects of gender and ethnicity, as well as marital and parental status and other personal characteristics (e.g., empathic capability, psychological needs) on these constructs, especially worldview and moral orientation. By studying different groups of individuals, such as social activists on social justice or environmental issues, moral exemplars, and delinquents, researchers may uncover individual differences in the ways these ideas are organized and interrelated. Future research can also look at the influence of parental worldview on adolescent development. For research on adult development, more balanced and representative adult samples will be desirable. The current study has covered relationship closeness with seven general groups but left out intimate, romantic relationship, which was the foundation of the self-expansion model (Aron  149  & Aron, 1986). I speculate that intimate relationships are likely to lead to feeling of selfexpansion as people often refer to their spouses or partners as ―my better half.‖ The feeling of ―I was half, now that I am whole‖ in an intimate relationship implies that the intimate other is seen and included as part of the self, a Level 1 expansion. For individuals who endorse egocentric orientation highly, egocentric orientation will probably serve to protect not only their own selfinterest, but also the interest of their intimate others. However, for individuals who endorse family orientation higher than egocentric orientation, they will probably expect their intimate others to sacrifice for their family as well. It is possible that an expanded Level 1 inclusiveness will enhance higher-level moral orientations, because the self is now feeling secured and complete, or override higher-level moral orientations, because the intimate other is not the better, but the worse half of the self. Either way, romantic, intimate relationships are usually part of Level 1 inclusiveness. When the intimate other becomes or is going to become a parent of one‘s offspring, however, the intimate other may be seen as part of the family, that is, a Level 2 expansion. An expanded Level 2 inclusiveness with renewed family responsibilities, from that of a child to that of a parent, may supersede relationship closeness at other levels, including the self at Level 1, thus leading to self-reduction rather than self-expansion. A father once said that he used to have the whole steak for himself for supper, now he just cut the steak into pieces and gave all the pieces to his children. Parenthood may be an excellent opportunity for moral development. The inclusion of people in intimate relationships and people planning to raise a family in the investigation of the comprehensive model of moral development can thus provide insightful information on human development. It will be particularly interesting if they can be followed from the beginning of their courtship to the birth of their children and then their changing  150  parenthood at different stages, as their children grow up. To study developmental change in worldview, self-view, moral inclusiveness, moral orientation, moral judgment, behavior, and their relational structures, longitudinal designs will be required. Experimental study on the effect of worldview education and programs promoting moral self or ecological self or expanding Level 5 or Level 6 inclusiveness on moral orientation as well as moral judgment and behavior will be especially desirable. Experimental designs with pre- and post measures can provide hard evidence for the causal relations of worldview, self-view, and moral inclusiveness to moral orientation. The addition of moral judgment and behavior as dependent variables in the research can further inform us about the practical value of the comprehensive model of moral development and guide education and social reform. Whether researching the whole, or the first or the second part of the comprehensive model and whether the research design is cross-sectional, longitudinal, or experimental, the multidimensional Moral Orientation Index (MOI) will become handy. Future Research with the Moral Orientation Index As the focal point of the comprehensive model for model development, moral orientation represents habitual morality that results from the multi-level person-environment interaction processes and results in individual current functioning (moral judgment and behavior) in its immediate environment. Thus, moral orientation is an important construct in moral psychology. Lacking an acceptable measure of the multidimensionality of moral orientation, the construct has been underused and misconstrued. The current data indicate that the MOI has differentiated among family, care, and egocentric orientations, between religious and norm orientations, and between justice and biocentric orientations. The MOI subscales seem to have adequate reliability and satisfactory construct validity, as supported by both factor analysis and the structural models  151  at each level of inclusiveness. Model F that starts with moral orientation positively predicting worldview, self-view, and relationship closeness has never been chosen at any levels of inclusiveness, suggesting that moral orientation is the joint function of worldview, self-view, and relationship closeness. Therefore, the MOI has practical value as a research tool. For example, the MOI can be used to validate the stage assignments of moral judgment items. When moral judgment is assessed out of context, as in the Defining Issues Test (Rest, 1993), we have no way of knowing the inner process that gives rise to that judgment. To put moral judgment into the context of worldview, self-view, and moral inclusiveness may seem a daunting task. Correlating moral judgment with moral orientation can at least tell us where the judgment came from. By correlating moral orientation scores with the measures of (e.g., environmental) moral judgment and behavior, we could get a better picture of the underlying reasons for that judgment and behavior. Group differences can be examined based on pairs or groups of moral orientation scores. For instance, four groups of participants can be distinguished based on their scores on egocentric and justice orientations: (a) those who score high on egocentric orientation but low on justice orientation, (b) those who score low on egocentric orientation but high on justice orientation, (c) those who score almost equally low on both egocentric and justice orientations, and (d) those who score almost equally high on both egocentric and justice orientations. The first group is more likely to admit openly that egotism is their moral principle and primary orientation. The third group is more likely to be indecisive and somewhat disorientated in their worldview, selfview, social relationships, and moral compass. Compared with the second and third groups, the last group is more likely to show moral hypocrisy (i.e., appear moral while trying to avoid the cost of acting moral; Batson, Kobrynowicz, Dinnerstein, Kampf, & Wilson, 1997) and to score  152  higher on hypothetical moral dilemmas, but lower on real-life dilemmas (e.g., Carpendale & Krebs, 1995; L. J. Walker et al., 1987). There are numerous ways to compare participants based on their scores on moral orientation subscales, and the comparisons could yield interesting and valuable insights into human morality. To combine information from all seven moral orientation scores of the MOI, we can develop two overall measures of moral development that characterize a participant‘s level of moral development, one for secular morality that peaks at Level 6 and one for spiritual morality that peaks at Level 7. These overall indexes can be derived from empirically weighted scoring schemes, so that longitudinal change, education change, and group differences can be readily assessed. The MOI items written in simple language can be used with participants who speak English as a second language and even Grades 6 and 7 students. However, there is room for improvement of the MOI. For instance, the coefficient alphas for care orientation subscale were not very high (.66 and .62). One or two items could be added to improve the reliability coefficients of the care subscale. As suggested by the data, the religious orientation subscale may be over-representing religiosity, but under-representing spirituality, despite high reliability coefficients. In addition to biocentric and religious orientations, another orientation that focuses on forgiveness and compassion towards all beings rather than friends and family members, as reflected in the relational care orientation, may be needed to assess spirituality at Level 6 and Level 7 inclusiveness. By comparing the scores of the forgiveness and compassion orientation(s) with the scores of other orientations such as egocentric, justice, biocentric, and religious orientations, we may be able to uncover the state of the individual‘s religious or spiritual (e.g., for self-salvation  153  or ‗all is one‘) and moral development. Thus, I do not regard the MOI as the final product and would anticipate that future work by other researchers will improve the MOI items and add important insights into its psychometric properties. Conclusions With its original empirical findings, the present research is unique in many respects. First, theorists have contended that socio-cultural influence is an essential part of mind and behavior and should not be isolated as variables to be analyzed (e.g., Damon, 1996). To avoid separating socio-cultural influence from personal meaning, the proposed comprehensive model of moral development incorporates cultural and social influences as part of the mental system by using the constructs of worldview, self-view, and moral inclusiveness. Based on a holistic and complementary perspective, the comprehensive model highlights the multi-level personenvironment interaction process as the source of individual moral functioning and development. Delineating the relations between habitual morality (moral orientation), reflective morality (moral judgment), and behavior, the model provides a useful conceptual framework that can deal with the complexities of everyday morality. Second, this study advances theory by integrating theories of morality from different approaches as different moral orientations reflecting different levels of inclusiveness, worldview, self-view, and social relationship or environment. Human morality is a complicated affair with many faces that requires investigation from various approaches. Thus, the field of moral psychology has been filled with diversities. It is time, however, to integrate the disparate knowledge gathered from various schools to provide a more comprehensive understanding of human morality. For example, research on the ethics of care and justice (e.g., Gilligan, 1982; Lyons, 1983), morality and spiritual or religious development (e.g., Batson et al., 1999;  154  MacLean, Walker, & Matsuba, 2004), and biocentric reasoning (e.g., Dunlap et al., 2000; Kahn, 1996, 1997) has informed us about different facets of morality. How these different facets of morality are interrelated, however, remains a mystery. The combination of various kinds of reasoning strategy in the same study, as this one, not only integrates diverse research literature on morality, but also offers a window into the interrelations among the various moral orientations and their developmental patterns, and a more complete picture of human morality. Third, by using individualism-collectivism as the basic worldview dimensions, this study combines cultural psychology with moral psychology. As indicated by the SEM results, worldview shapes not only self-view, but also moral inclusiveness and moral orientation. Different aspects of individualism and collectivism corresponding to different levels of moral inclusiveness and moral orientation thus builds a bridge from ―is‖ to ―ought.‖ That is, what we perceive the world to be informs us what is moral and what ought to be done. The primacy of worldview, particularly collectivism, in moral development has great implications for education. Fourth, by defining moral development as the tendency to take an increasingly wider perspective from different groups in defining social facts and moral values and by operationalizing moral inclusiveness as relationship closeness with different groups, this study has provided empirical data to justify moral inclusiveness as a valid psychological construct and a central one in moral development. Having good reliability, the Circles of Relationship Closeness is an acceptable measure of moral inclusiveness. Each level of inclusiveness reflects the social environment, in which relationship closeness with the social group in that environment is built and worldview specific to that environment develops. Finally, I have constructed the Moral Orientation Index (MOI) to assess seven commonly used moral orientations: egocentric, family, care, norm, justice, biocentric, and religious  155  orientations. The SEM results of each moral orientation reflecting different aspects of worldview, self-view, and relationship closeness corroborate with the factor analysis results that each orientation is unique and independently meaningful. The differentiation of family orientation from Kohlberg‘s (1971) social approval orientation, Gilligan‘s (1982) ethic of care, and Rest et al.‘s (1999) self-interest schema, as well as religious orientation from the Kohlbergian norm-maintaining orientation will facilitate future investigation into the influence of family and religion on moral development. Having satisfactory reliability and construct validity, the MOI will be a much needed tool in research, counseling, and moral education, in this multicultural and diversified era. In sum, the structural models at various levels of inclusiveness that were developed with the student data while making some comparative references to the adult data have provided strong evidence for the causal relations among worldview, self-view, moral inclusiveness, and moral orientation in the comprehensive model of moral development. The empirical findings of this original study suggest that the comprehensive model is holistic and complementary. It takes into account the influence of culture, social relationship, and context, without divorcing environmental influence from personal meaning. It portrays moral development as the successive expanding levels of moral inclusiveness that build on the previous one. Attesting to the complementary principles, the process of development is both interindividual and intraindividual, aspiring to achieve balance between polarities (e.g., individualism-collectivism, independence-interdependence). Countering the deficiencies of decontextualized and dichotomized theorizing, the comprehensive model is likely to influence theory and research in the field of moral psychology.  156  REFERENCES Agnew, C. R., Loving, T. J., Le, B., & Goodfriend, W. (2004). Thinking close: Measuring relational closeness as perceived self-other inclusion. In D. Mashek & A. Aron (Eds.), Handbook of closeness and intimacy (pp. 103 – 115). Mahwah, NJ: Erlbaum. Allport, G. (1937). Personality: A psychology interpretation. New York: Holt, Rinehart & Winston. Ambrose, D. (2000). World-view entrapment: Moral-ethical implications for gifted education. Journal for the Education of the Gifted, 23, 159-186. Aquino, K., & Reed, A. (2002). The self-importance of moral identity. Journal of Personality and Social Psychology, 83, 1423-1440. Arbuckle, J. L. (2006). Amos 7.0 user’s guide. Spring House, PA: Amos Development. Aron, A., & Aron, E. N. (1986). Love as the expansion of self: Understanding attraction and satisfaction. New York: Hemisphere. Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of Other in the Self Scale and the structure of interpersonal closeness. Journal of Personality and Social Psychology, 63, 596-612. Aron, A., Aron, E. N., Tudor, M., & Nelson, G. (1991). Close relationships as including other in the self. Journal of Personality and Social Psychology, 60, 241 – 253. Aron, A. P., Mashek, D. J., & Aron, E. N. (2004). Closeness as including other in the self. In D. J. Mashek & A. P. Aron (Eds.), Handbook of closeness and intimacy (pp. 27-41). Mahwah, NJ: Erlbaum. Aronson, H. (1993). Becoming an environmental activist: The process of transformation from everyday life into making history in the hazardous waste movement. Journal of Political and Military Sociology, 21, 63-80. Bandura, A. (1991). Self-regulation of motivation through anticipatory and self-reactive mechanisms. In R. A. Dienstbier (Ed.), Nebraska Symposium on Motivation, 1990: Perspectives on motivation (pp. 69-164). Lincoln, NE: University of Nebraska Press. Bandura, A. (2002). Selective moral disengagement in the exercise of moral agency. Journal of Moral Education, 31, 101-119. Batson, C. D., Floyd, R. B., Meyer, J. M., & Winner, A. L. (1999). ―And who is my neighbor?‖: Intrinsic religion as a source of universal compassion. Journal for the Scientific Study of Religion 38, 445-57.  157  Batson, C. D., Kobrynowicz, D., Dinnerstein, J. L., Kampf, H. C., & Wilson, A. D. (1997). In a very different voice: Unmasking moral hypocrisy. Journal of Personality and Social Psychology, 72, 1335–1348. Baumeister, R. F. (1997). Identity, self-concept, and self-esteem: The self lost and found. In R. Hogan, J. Johnson, & S. Briggs (Eds.), Handbook of personality psychology (pp. 681710). San Diego, CA: Academic Press. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238-246. Bentler, P. M., & Bonett, D. G. (1980). Significance test of goodness-of-fit in the analysis of covariance structures. Psychological Bulletin, 88, 588-606. Bentler, P. M., & Chou, C. (1987). Practical issues in structural modeling. Sociological Methods and Research, 16, 78-117. Bergman, L. R., Cairns, R. B., Nilsson, L.-G., & Nystedt, L. (2000). Developmental science and the holistic approach. Mahwah, NJ: Erlbaum. Berndt, T. J. (1999). Friends‘ influence on students‘ adjustment to school. Educational Psychologist, 34, 15-28. Bernstein, W. M., & Davis, M. H. (1982). Perspective-taking, self-consciousness, and accuracy in person perception. Basic and Applied Social Psychology, 3, 1-19. Blasi, A. (1984). Moral identity: Its role in moral functioning. In W. Kurtines & J. L. Gewirtz (Eds.), Morality, moral behavior and moral development (pp. 128-139). New York: Wiley. Blum, L. (1987). Particularity and responsiveness. In J. Kagan, & S. Lamb (Eds.), The emergence of morality in young children (pp. 306-377). University of Chicago Press. Bohm, D., & Hiley, B. J. (1993). The undivided universe: An ontological interpretation of quantum theory. London: Routledge. Bohr, N. (1934). Atomic theory and the description of nature. Cambridge, UK: University Press. Bollen, K. (1989). Structural equations with latent variables. New York: Wiley Bollen, K. (1990). Overall fit in covariance structural models: Two types of sample size effects. Psychological Bulletin, 107, 256-259. Bowlby, J. (1969). Attachment. New York: Basic Books. Boyatzis, C. J., Dollahite, D. C., & Marks, L. D. (2006). The family as a context for religious and spiritual development in children and youth. In E. C. Roehlkepartain, P. E. King, L.  158  Wagener, & P. L. Benson (Eds.), The handbook of spiritual development in childhood and adolescence (pp. 297-309). Thousand Oaks, CA: Sage. Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136-162). Newbury Park, CA: Sage. Campbell, E. K. (1983). Beyond anthropocentrism. Journal of the History of the Behavioral Sciences, 19, 54-67. Carpendale, J. I. M. (2000). Kohlberg and Piaget on stages and moral reasoning. Developmental Review, 20, 181-205. Carpendale, J. I. M., & Krebs, D. L. (1995). Variations in level of moral judgment as a function of type of dilemma and moral choice. Journal of Personality, 63, 289-313. Carr, D. (2002). Moral education and the perils of developmentalism. Journal of Moral Education, 31, 5-19. Carter, R. E. (1980). What is Lawrence Kohlberg doing? Journal of Moral Education, 9, 88-102. Cattell, R. B. (1988). The meaning and strategic use of factor analysis. In J. R. Nesselroade & R. B. Cattell (Eds.), Handbook of multivariate experimental psychology (2nd ed.) (pp. 131203). New York: Plenum Press. Chou, C.-P., & Bentler, P. M. (1995). Estimates and tests in structural equation modeling. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (pp. 37-55). Thousand Oaks, CA: Sage. Chovan, W., & Freeman, N. L. (1993). Moral reasoning and personality components in gifted and average students. Perceptual and Motor Skills, 77, 1297-1298. Cialdini, R. B., Brown, S. L., Lewis, B. P., Luc, C., & Neuberg, S. L. (1997). Reinterpreting the empathy-altruism relationship: When one into one equals oneness. Journal of Personality and Social Psychology, 73, 481-494. Clark, C. F., Kotchen, M. J., & Moore, M. R. (2003). Internal and external influences on proenvironmental behavior: Participation in a green electricity program. Journal of Environmental Psychology, 23, 237-246. Clayton, S. (2000). Models of justice in the environmental debate. Journal of Social Issues, 56, 459-474.  159  Clayton, S. (2003). Environmental identity: A conceptual and an operational definition. In S. Clayton & S. Opotow (Eds.), Identity and the natural environment: The psychological significance of nature (pp. 45-65). Cambridge, MA: MIT Press. Clayton, S., & Brook, A. (2005). Can psychology help save the world? A model for conservation psychology. Analyses of Social Issues and Public Policy: ASAP, 5, 87-102. Colby, A., & Damon, W. (1995). The development of extraordinary moral commitment. In D. Hart & M. Killen (Eds.), Morality in everyday life: Developmental perspective (pp. 342370). New York: Cambridge University Press. Colby, A., & Kohlberg, L. (with Speicher, B., Hewer, A., Candee, D., Gibbs, J., & Power, C.). (1987). The measurement of moral judgment: Vols. 1 and 2. New York: Cambridge University Press. Colby, A., Kohlberg, L., Gibbs, J., & Lieberman, M. (1983). A longitudinal study of moral judgment. Monographs of the Society for Research in Child development, 48, 1-124. Comrey, A. L., & Lee, H. B. (1992). A first course in factor analysis (2nd ed.). Hillsdale, NJ: Erlbaum. Cooper, W. (2003). William James‘s moral theory. Journal of Moral Education, 32, 411-422. Crick, N. R., & Dodge, K. A. (1994). A review and reformulation of social informationprocessing mechanisms in children's social adjustment. Psychological Bulletin, 115, 74101. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297-334. Cushman, P. (1990). Why the self is empty: Toward a historically situated psychology. American Psychologist, 45, 599-611. Cushman, P. (1991). Ideology obscured: Political uses of the self in Daniel Stern‘s infant. American Psychologist, 46, 206-219. Dalrymple, J. (2001). Safeguarding young people through confidential advocacy services. Child & Family Social Work, 6, 149-160. Damon, W. (1975). Early conceptions of positive justice as related to the development of logical operations. Child Development, 46, 301-312. Damon, W. (1996). Nature, second nature, and individual development: An ethnographic opportunity. In R. Jessor, A. Colby, & R. A. Shweder (Eds.), Ethnography and human development: Context and meaning in social inquiry (pp. 459-475). University of Chicago Press.  160  Danziger, K. (1997). The historical formation of selves. In R. D. Ashmore & L. Jussim (Series & Vol. Eds.), Rutgers series on self and social identity: Vol. 1. Self and identity: Fundamental issues (pp. 137-159). New York: Oxford University Press. Darley, J. M. (1995). Mutable theories that organize the world. Psychological Inquiry, 6, 290294. Davis, M. H., Conklin, L., Smith, A., & Luce, C. (1996). Effect of perspective taking on the cognitive representation of persons: A merging of self and other. Journal of Personality and Social Psychology, 70, 713-726. Davis, M. H., Luce, C., & Kraus, S. J. (1994). The heritability of characteristics associated with dispositional empathy. Journal of Personality, 62, 369-391. Davidson, P., & Youniss, J. (1991). Which comes first, morality or identity? In W. M. Kurtines & J. L. Gewirtz (Eds.), Handbook of moral behavior and development: Vol. 1. Theory (pp. 105-121). Hillsdale, NJ: Erlbaum. Dawson, T. L. (2003). A stage is a stage is a stage: A direct comparison of two scoring systems. Journal of Genetic Psychology, 164, 335-364. Dien, D. S. (1982). A Chinese perspective on Kohlberg‘s theory of moral development. Developmental Review, 2, 331-341. Dien, D. S. (1997). Worldviews and morality: How do they intersect? Human Development, 40, 345-349. Dunlap, R. E., Van Liere, K. D., Mertig, A. G., & Jones, R. E. (2000). Measuring endorsement of the New Ecological Paradigm: A revised NEP Scale. Journal of Social Issues, 56, 425442. Retrieved May 7, 2007, from the Academic Search Complete database. Gable, R. K., & Wolf, J. W. (1993). Instrumental development in the affective domain: Measuring attitudes and values in corporate and school settings (2nd ed.). Boston: Kluwer Academic. Gallup, G. G. Jr. (1997). On the rise and fall of self-conception in primates. In J. G. Snodgrass & R. L. Thompson (Eds.), Self across psychology: Self-recognition, self-awareness, and the self-concept (pp. 73-82). New York: New York Academy of Sciences. Garrod, A., & Beal, C. R. (1993). Voices of care and justice in children‘s responses to fable dilemmas. In A. Garrod (Ed.), Approaches to moral development: New research and emerging themes (pp. 59-71). New York: Teachers College Press. Gibbs, J. C. (1979). Kohlberg‘s moral stage theory: A Piagetian revision. Human Development, 22, 89-112.  161  Gibson, D. R. (1990). Personality correlates of logical and sociomoral judgment. Journal of Personality and Social Psychology, 59, 1296-1300. Gilligan, C. (1982). In a different voice: Psychological theory and women’s development. Cambridge, MA: Harvard University Press. Gilligan, C. (1990). Joining the resistance: Psychology, politics, girls and women. Michigan Quarterly Review, 29, 501-536. Gilligan, C. (1995). Hearing the difference: Theorizing connection. Hypatia, 10, 120-127. Gilligan, C., & Attanucci, J. (1988). Two moral orientations: Gender differences and similarities. Merrill-Palmer Quarterly, 34, 223-237. Gilligan, C., Brown, L. M., & Rogers, A. (1990). Psyche embedded: A place for body, relationships, and culture in personality theory. In A. Rabin, R. Zucker, R. Emmons, & S. Frank (Eds.), Studying persons and lives (pp. 86-147). New York: Springer. Gilligan, C., & Wiggins, G. (1987). The origins of morality in early childhood relationships. In J. Kagan & S. Lamb (Eds.), The emergence of morality in young children (pp. 277-305). University of Chicago Press. Goldfried, J., & Miner, M. (2002). Quest religion and the problem of limited compassion. Journal for the Scientific Study of Religion, 41, 685-695. Guba, E. G., & Lincoln, Y. S. (1994). Competing paradigms in qualitative research. In N. K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (pp. 105-117). Thousand Oaks, CA: Sage. Guisinger, S., & Blatt, S. J. (1994). Individuality and relatedness: Evolution of a fundamental dialectic. American Psychologist, 49, 104-111. Hardin, G. (1977). The limits of altruism: An ecologist’s view of survival. Bloomington, IN: Indiana University Press. Hardy, S. A., & Carlo, G. (2005). Identity as a source of moral motivation. Human Development, 48, 232-256. Harris, J. R. (1998). The nurture assumption: Why children turn out the way they do. New York: Free Press. Harris, M., Fontana, A.F., & Dowds, B.N. (1977). The World Hypotheses Scale: Rationale, reliability and validity. Journal of Personality Assessment, 41, 537-547. Hart, D., & Killen, M. (1995). Introduction. In D. Hart & M. Killen (Eds.), Morality in everyday life: Developmental perspective (pp. 1-20). New York: Cambridge University Press.  162  Hart, T. (2006). Spiritual experiences and capacities of children and youth. In E. C. Roehlkepartain, P. E. King, L. Wagener, & P. L. Benson (Eds.), The handbook of spiritual development in childhood and adolescence (pp. 163-177). Thousand Oaks, CA: Sage. Hay, D., Reich, K. H., & Utsch, M. (2006). Spiritual development: Intersections and divergence with religious development. In E. C. Roehlkepartain, P. E. King, L. Wagener, & P. L. Benson (Eds.), The handbook of spiritual development in childhood and adolescence (pp. 46-59). Thousand Oaks, CA: Sage. Hennig, K. H. (2004). Care gone awry: The role of attachment and reflective functioning. In T. A. Thorkildsen & H. J. Walberg (Eds.), Nurturing morality (pp. 61-76). New York: Kluwer Academic. Heubner, A. M., & Garrod, A. C. (1993). Moral reasoning among Tibetan monks: A study of Buddhist adolescents and young adults in Nepal. Journal of Cross-Cultural Psychology, 24, 167-185. Ho, D. Y.-F. & Chiu, C. -Y. (1994). Component ideas of individualism, collectivism, and social organization: An application in the study of Chinese culture. In U. Kim & H. C. Triandis (Eds.), Individualism and collectivism: Theory, method, and applications (pp. 137-156). Thousand Oaks, CA: Sage. Hoffman, M. L. (1976). Empathy, role-taking, guilt, and development of altruistic motives. In T. Lickona (Ed.), Moral development and behavior: Theory, research, and social issues (pp. 124-143). New York: Holt, Rinehart & Winston. Hoffman, M. L. (2000). Empathy and moral development: Implications for caring and justice. Cambridge, UK: Cambridge University Press. Hogan, R. (1969). Development of an empathy scale. Journal of Consulting and Clinical Psychology, 33, 307-316. Hogan, R. (1975). Theoretical egocentrism and the problem of compliance. American Psychologist, 30, 533-540. Holland, D. (1997). Selves as cultured: As told by an anthropologist who lacks a soul. In R. D. Ashmore & L. Jussim (Series & Vol. Eds.), Rutgers series on self and social identity: Vol. 1. Self and identity: Fundamental issues (pp. 160-190). New York: Oxford University Press. Hong, Y.-Y., Chiu, C.-Y., & Dweck, C. S. (1999). Implicit theories, attributions, and coping: A meaning system approach. Journal of Personality & Social Psychology, 77, 588-599. Hu, L.-T., & Bentler, P. M. (1995). Evaluating model fit. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications (pp.76-99). Thousand Oats, CA: Sage.  163  Hui, C. H., & Triandis, H. C. (1986). Individualism-collectivism: A study of cross-cultural researchers. Journal of Cross-Cultural Psychology, 17, 225-248. Hyland, M. E. (1985). Do person variables exist in different ways? American Psychologist, 40, 1003–1010. Imamoglou, E. O. (1998). Individualism and collectivism in a model and scale of balanced differentiation and integration. Journal of Psychology, 132, 95-105. Ingalsbee, T. (1996). Earth First! Activism: Ecological postmodern praxis in radical environmentalist identities. Sociological Perspectives, 39, 263-276. Retrieved May 7, 2007, from the PsycINFO database. Ianni, F. A. J. (1989). The search for structure: A report on American youth today. New York: Free Press. Jaffee, S., & Hyde, J. S. (2000). Gender differences in moral orientation: A meta-analysis. Psychological Bulletin, 126, 703-726. James, W. (1897). The will to believe and other essays in popular philosophy. New York: Dover. James, W. (1902). The varieties of religious experience. New York: Penguin Books. (Original work published 1982) James, W. (1983). The principles of psychology. Cambridge, MA: Harvard University Press. (Original work published 1890) Jankowski, M. S. (1991). Islands in the street: Gangs and American urban society. Berkeley, CA: University of California Press. Jenkins, T. N. (2002). Chinese traditional thought and practice: Lessons for an ecological economics worldview. Ecological Economics, 40, 39-52. Jensen, L. A. (1997). Different worldviews, different morals: America‘s culture war divide. Human Development, 40, 325-344. Johnson, D. K. (1988). Adolescents‘ solutions to dilemmas in fables: Two moral orientations— Two problem solving strategies. In C. Gilligan, J. V. Ward, J. M. Taylor, & B. Bardige (Eds.), Mapping the moral domain: A contribution of women’s thinking to psychological theory and education (pp. 49-71). Cambridge, MA: Harvard University Press. Johnson, C. N., & Boyatzis, C. J. (2006). Cognitive-cultural foundations of spiritual development. In E. C. Roehlkepartain, P. E. King, L. Wagener, & P. L. Benson (Eds.), The handbook of spiritual development in childhood and adolescence (pp. 211-223). Thousand Oaks, CA: Sage.  164  Jurin, R. R., & Hutchinson, S. (2005). Worldviews in transition: Using ecological autobiographies to explore students‘ worldviews. Environmental Education Research, 11, 485-501. Retrieved December 1, 2007, from the Academic Search Complete database. Kahn, P. H. (1996). Along the Rio Negro: Brazilian children‘s environmental views and values. Developmental Psychology, 32, 979-987. Kahn, P. H. (1997). Children‘s moral and ecological reasoning about the Prince William Sound oil spill. Developmental Psychology, 33, 1091-1096. Kahn, P. H. Jr. (1999). The human relationship with nature: Development and culture. Cambridge, MA: MIT Press. Kahn, P. H. Jr. (2003). The development of environmental moral identity. In S. Clayton & S. Opotow (Eds.), Identity and the natural environment: The psychological significance of nature (pp. 113-134). Cambridge, MA: MIT Press. Kaplan, D. (2000). Structural equation modeling: Foundations and extensions. Thousand Oats, CA: Sage. Keltikangas-Järvinen, L. (2001). Aggressive behaviour and social problem-solving strategies: A review of the findings of a seven-year follow-up from childhood to late adolescence. Criminal Behaviour and Mental Health, 11(4), 236-250. Kempton, W., Boster, J. S., & Hartley, J. A. (1995). Environmental values in American culture. Cambridge, MA: MIT Press. Kempton, W., & Holland, D. C. (2003). Identity and sustained environmental practice. In S. Clayton & S. Opotow (Eds.), Identity and the natural environment: The psychological significance of nature (pp. 317-341). Cambridge, MA: MIT Press. King, M., Speck, P., & Thomas, A. (2001). The Royal Free Interview for spiritual and religious beliefs: Development and validation of a self-report version. Psychological Medicine, 31, 1015-1023. Kohlberg, L. (1969). Stages and sequence: The cognitive-developmental approach to socialization. In D. Goslin (Ed.), Handbook of socialization theory and research (pp. 347-480). Chicago: Rand McNally. Kohlberg, L. (1971). Indoctrination versus relativity in value education. Zygon, 6, 285-310. Kohlberg, L. (1976). Moral stages and moralization: The cognitive-developmental approach. In T. Lickona (Ed.), Moral development and behavior: Theory, research, and social issues (pp. 31-53). New York: Holt, Rinehart & Winston. Kohlberg, L. (1981). Essays in moral development: Vol. I. The philosophy of moral development. New York: Harper & Row.  165  Kohlberg, L. (1984). Essays in moral development: Vol. II. The psychology of moral stages, their nature and validity. San Francisco: Harper & Row. Kohlberg, L., & Ryncarz, R. A. (1990). Beyond justice reasoning: Moral development and consideration of a seventh stage. In C. N. Alexander & E. J. Langer (Eds.), Higher stages of human development: Perspectives on adult growth (pp. 191-207). New York: Oxford University Press. Koltko-Rivera, M. E. (2004). The psychology of worldview. Review of General Psychology, 8, 3-58. Kortenkamp, K. V., & Moore, C. F. (2001). Ecocentrism and anthropocentrism: Moral reasoning about ecological commons dilemmas. Journal of Environmental Psychology, 21, 261272. Krebs, D. L., & Denton, K. (2005). Toward a more pragmatic approach to morality: A critical evaluation of Kohlberg‘s model. Psychological Review, 112, 629-649. Krebs, D. L., & Denton, K. (2006). Explanatory limitations of cognitive-developmental approaches to morality. Psychological Review, 113, 672-675. Krebs, D. L., Denton, K. L., Vermeulen, S. C., Carpendale, J. I., & Bush, A. (1991). Structural flexibility of moral judgment. Journal of Personality and Social Psychology, 61, 10121023. Lao Tzu. (1972). Tao Te Ching. (G.-F. Feng & J. English, Trans.) New York: Alfred A. Knopf. Lee, M., & Prentice, N. M. (1988). Interrelation of empathy, cognition, and moral reasoning with dimensions of juvenile delinquency. Journal of Abnormal Child Psychology, 16, 127139. Leman, P. J. (2002). Argument structure, argument content and cognitive change in children's peer interaction. Journal of Genetic Psychology, 163, 40-57. Liddell, D. L. (1990). The Measure of Moral Orientation. (Available from the author, University of Iowa, Division of Counselor Education, Iowa City, IA 52242). Lyons, N. P. (1983). Two perspectives: On self, relationships, and morality. Harvard Educational Review, 53, 125-145. MacLean, A. M., Walker, L. J., & Matsuba, M. K. (2004). Transcendence and the moral self: Identity integration, religion, and moral life. Journal of the Scientific Study of Religion, 43, 429-437. Magnusson, D., & Stattin, H. (2006). The person in context: A holistic-interactionistic approach. In W. Damon & R. M. Lerner (Series Eds.) & R. M. Lerner (Vol. Ed.), Handbook of  166  child psychology, Vol. 1: Theoretical models of human development (6th ed.). (pp.400464). Hoboken, NJ: Wiley. Maner, J. K., Luce, C. L., Neuberg, S. L., Cialdini, R. B., Brown, S., & Segarin, B. J. (2002). The effects of perspective taking on motivations for helping: Still no evidence for altruism Personality and Social Psychology Bulletin, 28, 1601-1610. Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224-253. Martin, C. L., & Nagao, D. H. (1989). Some effects of computerized interviewing on job applicant responses. Journal of Applied Psychology, 74, 72-80. Mashek, D. J., Aron, A., & Boncimino, M. (2003). Confusions of self with close others. Personality and Social Psychology Bulletin, 29, 382-392. Mashek, D., Cannaday, L. W., & Tangney, J. P. (2007). Inclusion of Community in Self Scale: A single-item pictorial measure of community connectedness. Journal of Community Psychology, 35, 257-275. Maslow, A. H. (1954). Motivation and personality. New York: Harper. McDonald, R. P., & Marsh, H. W. (1990). Choosing a multivariate model: Noncentrality and goodness-of-fit. Psychological Bulletin, 107, 247-255. McSherry, W., & Cash, K. (2004). The language of spirituality: An emerging taxonomy. International Journal of Nursing Studies, 41, 151-161. Retrieved March 6, 2008, from the Academic Search Complete database. McSherry, W., & Smith, J. (2007). How do children express their spiritual needs? Paediatric Nursing, 19 (3), 17-20. Retrieved March 6, 2008, from the Academic Search Complete database. Mead, G. H. (1934). Mind, self, and society. Chicago, IL: University of Chicago Press. Miller, D. T. (1999). The norm of self-interest. American Psychologist, 54, 1053-1060. Miller, J. G., & Bersoff, D. M. (1995). Development in the context of everyday family relationships: Culture, interpersonal morality, and adaptation. In D. Hart & M. Killen (Eds.), Morality in everyday life: Developmental perspective (pp. 259-282). New York: Cambridge University Press. Myers, G., & Russell, A. (2003). Human identity in relation to wild black bears: A natural-social ecology of subjective creatures. In S. Clayton & S. Opotow (Eds.), Identity and the natural environment: The psychological significance of nature (pp. 67-90). Cambridge, MA: MIT Press.  167  Myers, L. J., Speight, S. L., Highlen, P. S., Cox, C. I., Reynolds, A. L., Adams, E. M., & Hanley, C.P. (1991). Identity development and worldview: Toward an optimal conceptualization. Journal of Counseling and Development, 70, 54-63. Nash, R. F. (1989). The rights of nature: A history of environmental ethics. Madison, WI: University of Wisconsin Press. Nasser, F., & Wisenbaker, J. (2003). A Monte Carlo study investigating the impact of item parceling on measures of fit in confirmatory factor analysis. Educational and Psychological Measurement, 63, 729-757. Retrieved December 30, 2007, from the PsycINFO database. Newberg, A. B., & Newberg, S. K. (2006). A neuropsychological perspective on spiritual development. In E. C. Roehlkepartain, P. E. King, L. Wagener, & P. L. Benson (Eds.), The handbook of spiritual development in childhood and adolescence (pp. 183-196). Thousand Oaks, CA: Sage. Nicholls, J. G. (1989). The competitive ethos and democratic education. Cambridge, MA: Harvard University Press. Ochs, E., & Kremer-Sadlik, T. (2007). Introduction: Morality as family practice. Discourse & Society, 18, 5-10. Retrieved March 6, 2008, from the Academic Search Complete database. Opotow, S. (2003). Forging social identity and social conflict. Journal of Peace Psychology, 9, 177-179. Retrieved December 30, 2007, from the PsycINFO database. Oser, F. K., Scarlett, G., & Bucher, A. (2006). Religious and spiritual development throughout the life span. In W. Damon & R. M. Lerner (Series Eds.) & R. M. Lerner (Vol. Ed.), Handbook of child psychology, Vol. 1: Theoretical models of human development (6th ed.) (pp. 942-998). Hoboken, NJ: Wiley. Palmer, E. J., & Hollin, C. R. (1996). Sociomoral reasoning, perceptions of own parenting and self-reported delinquency. Personality and Individual Differences, 21, 175-182. Palmer, E. J., & Hollin, C. R. (2001). Sociomoral reasoning, perceptions of parenting and selfreported delinquency in adolescents. Applied Cognitive Psychology, 15, 85-100. Perloff, R. (1987). Self-interest and personal responsibility redux. American Psychologist, 42, 311. Peterson, J. B., & Flanders, J. L. (2005). Play and the regulation of aggression. In R. E. Tremblay, W. W. Hartup, & J. Archer (Eds.), Developmental origins of aggression (pp. 133-157). New York: Guilford Press. Piaget, J. (1965). The moral judgment of the child. Glencoe, IL: Free Press. (Original work published in 1932)  168  Plante, T. G. & Boccaccini, M. T. (1997). The Santa Clara Strength of Religious Faith Questionnaire. Pastoral Psychology, 45, 375-387. Pratt, M. W., Golding, G., Hunter, W., & Sampson, R. (1988). Sex differences in adult moral orientation. Journal of Personality, 56, 373-391. Pratt, M. W., Hunsberger, B., Pancer, S. M., & Alisat, S. (2003). A longitudinal analysis of personal values socialization: Correlates of a moral self-ideal in late adolescence. Social Development, 12, 563-585. Puka, B. (1991). Toward the redevelopment of Kohlberg‘s theory: Preserving essential structure, removing controversial content. In W. M. Kurtines & J. L. Gewirtz (Eds.), Handbook of moral behavior and development: Vol. 1. Theory (pp. 373-393). Hillsdale, NJ: Erlbaum. Rand, A. (1964). The virtue of selfishness. New York: Signet. Reed, A., & Aquino, K. F. (2003). Moral identity and the expanding circle of moral regard toward out-groups. Journal of Personality and Social Psychology, 84, 1270-1286. Rest, J. R. (1973). The hierarchical nature of moral judgment: A study of patterns of comprehension and preference of moral stages. Journal of Personality, 41, 86-109. Rest, J. R. (1976). New approaches in the assessment of moral judgment. In T. Lickona (Ed.), Moral development and behavior: Theory, research, and social issues (pp. 198-218). New York: Holt, Rinehart & Winston. Rest, J. R. (1979). Development in judging moral issues. Minneapolis, MN: University of Minnesota Press. Rest, J. (1993). Guide for the Defining Issues Test--Revised (Version 1.3). (Available from the Center for the Study of Ethical Development, 206-A Burton Hall, 178 Pillsbury Drive SE, University of Minnesota, Minneapolis, MN 55455). Rest, J., Cooper, D., Coder, R., Masanz, J., & Anderson, D. (1974). Judging the important issues in moral dilemmas—An objective test of development. Developmental Psychology, 10, 491-501. Rest, J., Narvaez, D., Bebeau, M. J., & Thoma, S. J. (1999). Postconventional moral thinking: A neo-Kohlbergian approach. Mahwah, NJ: Erlbaum. Rokeach, M. (1973). The nature of human values. New York: Free Press. Ryckman, R. M., Hammer, M., Kaczor, L. M., & Gold, J. A. (1990). Construction of a Hypercompetitive Attitude Scale. Journal of Personality Assessment, 55, 630-639.  169  Ryckman, R. M., Hammer, M., Kaczor, L. M., & Gold, J. A. (1996). Construction of a Personal Development Competitive Attitude Scale. Journal of Personality Assessment, 66, 374386. Sampson, E. E. (1977). Psychology and the American ideal. Journal of Personality & Social Psychology, 35, 767-782. Sampson, E. E. (1988). The debate on individualism: Indigenous psychologies and their role in personal and societal functioning. American Psychologist, 43, 15-22. Sanders, C. E., Lubinski, D., & Benbow, C. P. (1995). Does the Defining Issues Test measure psychological phenomena distinct from verbal ability?: An examination of Lykken‘s query. Journal of Personality and Social Psychology, 69, 498-504. Schultz, P.W. (2000). Empathizing with nature: The effects of perspective-taking on concern for environmental issues. Journal of Social Issues, 56, 291–406. Schumacker, R. E., & Lomax, R. G. (1996). A beginner‘s guide to structural equation modeling. Hillsdale, NJ: Erlbaum. Selman, R. L. (1976). Social-cognitive understanding: A guide to educational and clinical practice. In T. Lickona (Ed.), Moral development and behavior: Theory, research, and social issues (pp. 299-316). New York: Holt, Rinehart & Winston. Shimizu, H. (2000). Beyond individualism and sociocentrism: An ontological analysis of the opposing elements in personal experiences of Japanese adolescents. Human Development, 43, 195-211. Simpson, E. L. (1974). Moral development research: A case study of scientific cultural bias. Human Development, 17, 81-106. Singelis, T. M. (1994). The measurement of independent and interdependent self-construals. Personality & Social Psychology Bulletin, 20, 580-591. Singelis, T. M., Triandis, H. C., Bhawuk, D., & Gelfand, M. J. (1995). Horizontal and vertical dimensions of individualism and collectivism: A theoretical and measurement refinement. Cross-Cultural Research, 29, 240-275. Sinha, D. & Tripathi, R. C. (1994). Individualism in a collectivist culture: A case of coexistence of opposites. In U. Kim & H. C. Triandis (Eds.), Individualism and collectivism: Theory, method, and applications (pp. 123-136). Thousand Oaks, CA: Sage. Skoe, E. E., & Marcia, J. E. (1991). A measure of care-based morality and its relation to ego identity. Merrill-Palmer Quarterly, 37, 289-304. Slavin, R. E., & Cooper, R. (1999). Improving intergroup relations: Lessons learned from cooperative learning programs. Journal of Social Issues, 55, 647-663.  170  Smith, H. (2001). Why religion matters: The fate of the human spirit in an age of disbelief. San Francisco: Harper. Snarey, J. R. (1985). Cross-cultural universality of social-moral development: A critical review of Kohlbergian research. Psychological Bulletin, 97, 202-232. Snarey, J., & Keljo, K. (1991). In a Gemeinschaft voice: The cross-cultural expansion of moral development theory. In W. M. Kurtines & J. L. Gewirtz (Eds.), Handbook of moral behavior and development: Vol. 1. Theory (pp. 373-393). Hillsdale, NJ: Erlbaum. Snarey, J. R., Reimer, J., & Kohlberg, L. (1984). The socio-moral development of kibbutz adolescents: A longitudinal, cross-cultural study. Developmental Psychology, 21, 3-17. Snyder, D. M. (1983). On the nature of the relationships involving the observer and the observed phenomenon in psychology and physics. Journal of Mind and Behavior, 4, 389–400. Sober E., & Wilson, D. S. (1998). Unto others: The evolution and psychology of unselfish behavior. Cambridge, MA: Harvard University Press. Speicher, B. (1992). Adolescent moral judgment and perceptions of family interaction. Journal of Family Psychology, 6, 128-138. Speicher, B. (1994). Family patterns of moral judgment during adolescence and early adulthood. Developmental Psychology, 30, 624-632. Staub, E. (1984). Steps toward a comprehensive theory of moral conduct: Goal orientation, social behavior, kindness, and cruelty. In W. Kurtines & J. L. Gewirtz (Eds.), Morality, moral behavior and moral development (pp. 241-260). New York: Wiley. Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics (3rd ed.). New York: HarperCollins College. Tappan, M. B. (1997). Language, culture, and moral development: A Vygotskian perspective. Developmental Review, 17, 78-100. Triandis, H. C. (1989). The self and social behavior in differing cultural contexts. Psychological Review, 96, 506-520. Triandis, H. C. (1996). The psychological measurement of cultural syndromes. American Psychologist, 51, 407-415. Triandis H. C., & Gelfand, M. J. (1998). Converging measurement of horizontal and vertical individualism and collectivism. Journal of Personality & Social Psychology, 74, 118128. Tropp, L. R., & Wright, S. C. (2001). Ingroup identification as the inclusion of ingroup in the self. Personality and Social Psychology Bulletin, 27, 585-600.  171  Turiel, E. (2004). Commentary: Beyond individualism and collectivism—A problem, or progress? New Directions for Child and Adolescent Development, 104, 91-100. Uleman, J. S., Rhee, E., Bardoliwalla, N., Semin, G., & Toyama, M. (2000). The relational self: Closeness to ingroups depends on who they are, culture, and the type of closeness. Asian Journal of Social Psychology, 3, 1-17. Retrieved November 16, 2004, from the PsycINFO database. Vandenberg, B. (1991). Is epistemology enough? An existential consideration of development. American Psychologist, 46, 1278-1286. Wainryb, C. (2004). ―Is‖ and ―ought‖: Moral judgments about the world as understood. New Directions for Child and Adolescent Development, 103, 3-18. Walker, J. S. (2000). Choosing biases, using power and practicing resistance: Moral development in a world without certainty. Human Development, 43, 135-156. Walker, L. J. (1989). A longitudinal study of moral reasoning. Child Development, 60, 157-166. Walker, L. J. (1997). Is morality gendered in early parent-child relationships? A commentary on the Lollis, Ross, and Leroux study. Merrill-Palmer Quarterly, 43, 148-159. Walker, L. J., de Vries, B., & Trevethan, S. D. (1987). Moral stages and moral orientations in real-life and hypothetical dilemmas. Child Development, 58, 842-858. Walker, L. J., & Moran, T. J. (1991). Moral reasoning in a communist Chinese society. Journal of Moral Education, 20, 139-155. Walker, L. J., Hennig, K. H., & Krettenauer, T. (2000). Parent and peer contexts for children‘s moral reasoning development. Child Development, 71, 1033-1048. Walker, L. J., Pitts, R. C., Hennig, K. H., & Matsuba, M. K. (1995). Reasoning about morality and real-life moral problems. In D. Hart & M. Killen (Eds.), Morality in everyday life: Developmental perspective (pp. 371-408). New York: Cambridge University Press. Wark, G. R., & Krebs, D. L. (1996). Gender and dilemma differences in real-life moral judgment. Developmental Psychology, 32, 220-230. Wark, G. R., & Krebs, D. L. (1997). Sources of variation in moral judgment: Toward a model of real-life morality. Journal of Adult Development, 4, 163-178. Wheaton, B., Muthén, B., Alwin, D. F., & Summers, G. F. (1977). Assessing reliability and stability in panel models. In D. R. Heise (Ed.), Sociological methodology (pp. 84-136). San Francisco: Jossey-Bass. White, F. A., & Matawie, K. M. (2004). Parental morality and family processes as predictors of adolescent morality. Journal of Child and Family Studies, 13, 219-233.  172  Wygant, S. A. (1997). Moral reasoning about real-life dilemmas: Paradox in research using the Defining Issues Test. Personality and Social Psychology Bulletin, 23, 1022-1033. Yacker, N., & Weinberg, S. L. (1990). Care and justice moral orientation: A scale for its assessment. Journal of Personality Assessment, 55, 18-27. Yamada, A.-M., & Singelis, T. M. (1999). Biculturalism and self-construal. Journal of Intercultural Relations, 23, 697-709. Yu, A. (1998, November). Faith, emotion regulation, empathy, and moral behavior: A case study. Poster session presented at the annual meeting of the Association for Moral Education, Hanover, NH, USA. Yu, A. (1999, November). The Tao of morality. Paper presented at the annual meeting of the Association for Moral Education, Minneapolis, MN, USA. Yu, A. (2001). Interrelations among empathy, moral judgment, and moral orientation in adolescents. Unpublished master‘s thesis, University of British Columbia, Vancouver, British Columbia, Canada. Yu, A., & Kishor, N. (2001a, April). Is moral orientation dispositional or situational? –The Objective Measures of Dispositional and Situational Moral Orientations. Paper presented at the annual meeting of the American Educational Research Association, Seattle, WA, USA. Yu, A., & Kishor, N. (2001b, October). Sex and cultural differences in moral orientation in adolescence. Paper presented at the annual meeting of the Association for Moral Education, Vancouver, BC, Canada. Zavestoski, S. (2003). Constructing and maintaining ecological identities: The strategies of deep ecologists. In S. Clayton & S. Opotow (Eds.), Identity and the natural environment: The psychological significance of nature (pp. 297-315). Cambridge, MA: MIT Press.  173  APPENDICES Appendix A The Measurement of Horizontal and Vertical Individualism and Collectivism Vertical Individualism (VI): W2. Winning is everything. W5. I enjoy working in situations involving competition. W19. Without competition, it is not possible to have a good society. W23. Competition is the law of nature. W32. Some people emphasize winning; I am not one of them (reversed). W25. When another person does better than I do, I get tense and aroused. W36. It annoys me when other people perform better than I do. W40. It is important that I do my job better than others. Vertical Collectivism (VC): W4. It is important to me that I respect the decisions made by my groups. W6. Children should be taught to place duty before pleasure. W11. Family members must stay together, despite any hardships or sacrifices.* W16. Parents should live with their children for as long as possible.* W24. It is my responsibility to look after my family no matter what it takes.* W35. I usually sacrifice my self-interest for the benefit of my group. Horizontal Collectivism (HC): W1. If a relative were in financial difficulty, I would help within my means. W3. The well-being of my coworkers is important to me. W9. To me, pleasure is spending time with others. W10. I feel good when I cooperate with others. W13. My happiness depends very much on the happiness of those around me. W21. If a coworker gets a prize, I would feel proud. W26. It is important to me to maintain harmony in my group. W39. I like sharing little things with my neighbors. Horizontal Individualism (HI): W12. I often do ‗my own thing‘. W17. My personal identity, independent of others, is very important to me. W18. I‘d rather depend on myself than others. W20. I rely on myself most of the time; I rarely rely on others. W28. Being a unique individual is important to me. ______________________________________________________________________________ Note. Scale items are from ―Horizontal and Vertical Dimensions of Individualism and Collectivism: A Theoretical and Measurement Refinement,‖ by T. M. Singelis, H.C. Triandis, D. Bhawuk, and M. J. Gelfand, 1995, Cross-Cultural Research, 29, pp. 255-256. Copyright 1995 by SAGE Publications. Reprinted with permission. *Original items have been removed for copyright reasons; brief descriptions of the removed items have been provided instead. To locate the original material, please see ―Converging Measurement of Horizontal and Vertical Individualism and Collectivism,‖ by H. C. Triandis and M. J. Gelfand, 1998, Journal of Personality and Social Psychology, 74, p. 120.  174  Appendix B The Circles of Relationship Closeness THE CLOSENESS OF YOUR RELATIONSHIPS Please describe how close you are to the following people or groups. You will do this by using the scale below. In each of the seven pairs of circles, one circle represents you, and the other circle represents the group or person. This is the scale for your closeness to a person or group, ranging from 1 to 7:  1  2  3  4  5  6  7  (A). Please rate the emotional closeness of your relationship with each of the following groups. Circle the number that best matches your answer.  1  2  3  4  5  6  7  1.  Your closeness with… your close friend (if you have more than one close friend, just pick one and refer to the same one in Sections B and C)  1  2  3  4  5  6  7  2.  your closest family member (if you have more than one closest family member, just pick one and refer to the same one in Sections B and C)  1  2  3  4  5  6  7  3.  your peer group (if you have more than one peer group, just pick one and refer to the same one in Sections B and C)  1  2  3  4  5  6  7  4.  your family (if you have more than one family, just pick one and refer to the same one in Sections B and C)  1  2  3  4  5  6  7  5.  your school or workplace (pick one and refer to the same one in Sections B and C)  1  2  3  4  5  6  7  6.  your local community  1  2  3  4  5  6  7  7.  your nation (if you have more than one nationalities, just pick one and refer to the same one in Sections B and C)  1  2  3  4  5  6  7  8.  nature  1  2  3  4  5  6  7  9.  the universe  1  2  3  4  5  6  7  10. God or the higher being(s) in your religious/ spiritual beliefs  175  (B). Please rate the supportive closeness of your relationship with each of the following groups. That is, how supportive each group is to you. Circle the number that best matches your answer.  1  2  3  4  5  6  7  Your closeness with… 11. your close friend  1  2  3  4  5  6  7  12. your closest family member  1  2  3  4  5  6  7  13. your peer group  1  2  3  4  5  6  7  14. your family  1  2  3  4  5  6  7  15. your school or workplace  1  2  3  4  5  6  7  16. your local community  1  2  3  4  5  6  7  17. your nation  1  2  3  4  5  6  7  18. nature  1  2  3  4  5  6  7  19. the universe  1  2  3  4  5  6  7  20. God or the higher being(s) in your religious/ spiritual beliefs  (C). Please rate the identity closeness of your relationship with each of the following groups. That is, how much you identify yourself (in your self-identity or who you are as a person) as a member of each group. Circle the number that best matches your answer.  1  2  3  4  5  6  7  Your closeness with… 21. your close friend  1  2  3  4  5  6  7  22. your closest family member  1  2  3  4  5  6  7  23. your peer group  1  2  3  4  5  6  7  24. your family  1  2  3  4  5  6  7  25. your school or workplace  1  2  3  4  5  6  7  26. your local community  1  2  3  4  5  6  7  27. your nation  1  2  3  4  5  6  7  28. nature  1  2  3  4  5  6  7  29. the universe  1  2  3  4  5  6  7  30. God or the higher being(s) in your religious/ spiritual beliefs  Note. Scale items are adapted from ―The Relational Self: Closeness to Ingroups Depends on Who They Are, Culture, and the Type of Closeness,‖ by J. S. Uleman, E. Rhee, N. Bardoliwalla, G. Semin, & M. Toyama, M., 2000, Asian Journal of Social Psychology, 3, p.5. Copyright 2000 by Blackwell Publishers Ltd with the Asian Association of Social Psychology and the Japanese Group Dynamics Association. Adapted with permission.  Appendix C The Moral Orientation Index HOW YOU GENERALLY SOLVE PROBLEMS OR MAKE DECISIONS Please rate each sentence on a scale from 1 (Strongly Disagree) to 7 (Strongly Agree) to indicate how much you agree or disagree with the sentence. Circle the number that best matches your answer.                Strongly Disagree 1  Disagree  Slightly Disagree  Neutral/ Undecided  Slightly Agree  Agree  2  3  4  5  6  Strongly Agree 7        1  2  3 4 5 6 7  1. I do what I feel is best for me, regardless of how that might affect others.  1  2  3 4 5 6 7  2. Preserving nature is more important than expanding business and consumption.  1  2  3 4 5 6 7  3. I am willing to forgive others who have done me wrong.  1  2  3 4 5 6 7  4. I try not to let my family down in whatever I do.  1  2  3 4 5 6 7  5. I would like to be known as a person who is sensitive to others‘ feelings.  1  2  3 4 5 6 7  6. I try to make everyone happy when making decisions.  1  2  3 4 5 6 7  7. When my desires conflict with my friends‘ benefits, I rather give up my desires than to hurt my friends.  1  2  3 4 5 6 7  8. Family relationships are more important than friendships.  1  2  3 4 5 6 7  9. The law of nature is more fair and just than the law of any human society.  1  2  3 4 5 6 7  10. Responsibilities to my family are more important than my own rights.  1  2  3 4 5 6 7  11. All life forms, including humans, animals, and plants are equally valuable.  1  2  3 4 5 6 7  12. All individuals are equal, regardless of race.  1  2  3 4 5 6 7  13. Students should obey their teachers.  1  2  3 4 5 6 7  14. When I do people a favour, I expect them to return a favour in the future.  1  2  3 4 5 6 7  15. Animals have a right to live just as much as humans do.  1  2  3 4 5 6 7  16. Obeying God or other higher being(s) is more important than obeying one‘s parents or man-made laws.  1  2  3 4 5 6 7  17. Protecting the natural environment is our social responsibility.  1  2  3 4 5 6 7  18. In resolving conflicts, each person‘s rights are equally important.  1  2  3 4 5 6 7  19. All individuals are equal, regardless of sexual orientation.  1  2  3 4 5 6 7  20. I make decisions based on my own needs and desires.  177  HOW YOU GENERALLY SOLVE PROBLEMS OR MAKE DECISIONS (Cont.)        1  2  3 4 5 6 7  21. Workers and soldiers must show respect to their superiors.  1  2  3 4 5 6 7  22. I make decisions based on my religious or spiritual beliefs.  1  2  3 4 5 6 7  23. When I make decisions, I take my family‘s opinions and benefits into consideration.  1  2  3 4 5 6 7  24. I try to avoid making decisions that may damage my relationships.  1  2  3 4 5 6 7  25. In making decisions, I weigh out the cost and benefit.  1  2  3 4 5 6 7  26. If people are mean to me, I‘ll be mean to them.  1  2  3 4 5 6 7  27. Taking care of our natural environment is more important than following the law.  1  2  3 4 5 6 7  28. We must maintain social order to protect personal and public safety.  1  2  3 4 5 6 7  29. Responsibilities to one‘s family are more important than duties to the society.  1  2  3 4 5 6 7  30. I prefer to forgive than to revenge.  1  2  3 4 5 6 7  31. We must comply with all the laws, although some of them may seem unnecessary or unreasonable at times.  1  2  3 4 5 6 7  32. In solving problems, I try to protect my own benefits first.  1  2  3 4 5 6 7  33. I don‘t waste my time and effort on something that‘s not going to benefit me.  1  2  3 4 5 6 7  34. In resolving conflicts, I rely on the teachings of my religion or spiritual beliefs.  1  2  3 4 5 6 7  35. For the sake of national security, sometimes it is impossible to treat everybody equally and fairly.  1  2  3 4 5 6 7  36. My benefits and goals are more important than my family‘s benefits and goals.  1  2  3 4 5 6 7  37. I am willing to agree in order to have good relationships with my friends.  1  2  3 4 5 6 7  38. All individuals are equal, regardless of religious or spiritual beliefs.  1  2  3 4 5 6 7  39. I try to follow the existing rules and laws of the society.  1  2  3 4 5 6 7  40. I pray to God or other higher being(s) for guidance when solving problems.  178  Appendix D Student Recruitment Letter  179  180  Appendix E Parent Consent Form  181  182  Appendix F Signature Page  183  184  Appendix G Adult Introductory Letter  185  Appendix H Adult Cover Letter  186  Appendix I Demographic Data Sheet for Students BACKGROUND INFORMATION (1) Are you male or female? (Circle One) (2) How old are you?  Male  Female  ____________ (years)  (3) What grade are you in this year? (Circle One)  8th  9th  10th  11th  12th  (4) How do you describe your cultural background? (Circle one or more, as you see it) 1….  First Nation (e.g., Aboriginal, Native Indian, Métis, Inuit)  2….  Black (e.g., African, Caribbean)  3….  White (e.g., Anglo, Caucasian, European)  4….  East Asian (e.g. Chinese, Japanese, Korean, Polynesian)  5….  South Asian (e.g. Indian, Pakistani, Sri Lankan, Bangladeshi)  6….  Southeast Asian (e.g. Burmese, Cambodian, Filipino, Laotian, Malaysian, Thai, Vietnamese)  7….  Middle East (e.g. Arabian, Armenian, Iranian, Israeli, Lebanese, Palestinian, Syrian, Turkish)  8….  Latin American (e.g. Mexican, native Central and South American)  9….  If your cultural background is not listed above, please describe your background here: ________________________________________________________________________  (5) The most important source of your religious/spiritual beliefs is: (Circle one) 1…. No religion  2…. Animism  3…. Buddhism  4…. Catholic  5…. Christianity  6…. Confucius  7…. Hinduism  8…. Islam  9…. Judaism  10…Shinto  11... Sikhism  12…Scientology  13…Other, please specify: _________________________________________________________ (6) What language(s) do you speak at home? ________________________________________  (7) How long have you lived in Canada? ________________ (years)  187  Appendix J  Demographic Data Sheet for Adults BACKGROUND INFORMATION (1) Are you male or female? (Circle One)  Male  Female  (2) Your age: (Circle One) 1… 20 or less  2… 21-24  3… 25-29  4… 30-34  5… 35-39  6… 40-44  7… 45-49  8… 50 or more  (3) Your education level: (Circle One) 1… Some high school  2… Graduated from high school  3… Vocational school or technical school  4… Some college  5… Some university  6… Graduated from university  7… Attending/attended graduate or professional school (e.g., to be a doctor, lawyer, or teacher) 8….Don‘t know 8… Don‘t know (4) Your cultural background: (Circle one or more, as you see it) 1….  First Nation (e.g., Aboriginal, Native Indian, Métis, Inuit)  2….  Black (e.g., African, Caribbean)  3….  White (e.g., Anglo, Caucasian, European)  4….  East Asian (e.g. Chinese, Japanese, Korean, Polynesian)  5….  South Asian (e.g. Indian, Pakistani, Sri Lankan, Bangladeshi)  6….  Southeast Asian (e.g. Burmese, Cambodian, Filipino, Laotian, Malaysian, Thai, Vietnamese)  7….  Middle East (e.g. Arabian, Armenian, Iranian, Israeli, Lebanese, Palestinian, Syrian, Turkish)  8….  Latin American (e.g. Mexican, native Central and South American)  9….  If your cultural background is not listed above, please describe your background here: __________________________________________________________________________  (5) Your religious/spiritual background: (Circle one) 1…. No religion  2…. Animism  3…. Buddhism  4…. Catholic  5…. Christianity  6…. Confucius  7…. Hinduism  8…. Islam  9…. Judaism  10… Scientology  11…Shinto  12... Sikhism  13…Other, please specify: _________________________________________________________ (6) What language(s) do you speak at home? ________________________________________ (7) How long have you lived in Canada? ________________ (years)  188  Appendix K The Self-Construal Scale Independent Self-Construal: S3. S5. S9. S11. S13. S15. S17. S19. S20. S22. S25. S31.  Being able to take care of myself is a primary concern for me. I value being in good health above everything. I prefer to be direct and forthright when dealing with people I‘ve just met. Speaking up during a class or a meeting is not a problem for me.* I can talk openly with a person who I meet for the first time, even when this person is much older than I am.* I act the same way at home that I do at school.* I‘d rather say ―No‖ directly, than risk being misunderstood. Having a lively imagination is important to me. I am comfortable with being singled out for praise or rewards. I act the same way no matter who I am with. I feel it is important for me to act as an independent person.* I enjoy being unique and different from others in many respects.  Interdependent Self-Construal: S1. S4.  I respect people who are modest about themselves. I often have the feeling that my relationships with others are more important than my own accomplishments. S7. I have respect for the authority figures with whom I interact. S16. Even when I strongly disagree with group members, I avoid an argument. S28. I will stay in a group if they need me, even when I‘m not happy with the group. S32. I feel my fate is intertwined or tangled with the fate of those around me.* S33. I usually go along with what others want to do, even when I would rather do something different.* ______________________________________________________________________________ Note. Scale items are from ―The Measurement of Independent and Interdependent SelfConstruals,‖ by T. Singelis, 1994, Personality & Social Psychology Bulletin, 20, p. 585. Copyright 1994 by SAGE Publications. Reprinted with permission. *Scale items are adapted with permission from the ―2006 Version of the Self-Construal Scale‖ by T. Singelis (personal communication, January 10, 2006), Department of Psychology, California State University, Chico, USA.  189  Appendix L Univariate Summary Statistics for Observed Variables for the Student Sample Level 1 Total sample minus outlier cases (n = 637) Variables VI SD S3 S9 S20 M1 M20 M32  Mean 4.02 3.76 5.49 4.73 4.75 4.54 4.87 4.62  SD 1.06 1.22 1.20 1.46 1.61 1.54 1.42 1.20  Skewness 0.06 0.04 -0.49 -0.29 -0.37 -0.44 -0.51 -0.24  Kurtosis -0.40 -0.20 -0.62 -0.41 -0.50 -0.45 -0.02 -0.29  Outlier cases (n = 3) Mean 5.33 4.67 3.33 3.33 6.33 4.00 5.00 4.33  SD 0.62 2.52 0.58 3.22 1.16 3.00 3.46 2.08  Skewness -1.65 -0.59 1.73 1.54 -1.73 0.00 -1.73 -1.29  Kurtosis . . . . . . . .  Level 2 Total sample minus outlier cases (n = 626) Variables W11 W16 W24 S1 S7 S28 C4 C14 C24 M4 M8 M23  Mean 5.99 5.07 5.50 5.66 5.24 4.85 5.57 5.88 5.79 5.70 4.94 5.16  SD 1.31 1.56 1.35 1.13 1.43 1.63 1.47 1.38 1.43 1.29 1.56 1.42  Skewness -1.38 -0.61 -0.76 -0.77 -0.81 -0.65 -0.99 -1.22 -1.19 -1.03 -0.45 -0.89  Kurtosis 1.26 -0.31 -0.04 0.40 0.42 -0.30 0.10 0.72 0.57 0.51 -0.31 0.52  Outlier cases (n = 14) Mean 5.36 4.07 4.50 4.86 4.21 3.14 3.64 5.14 4.57 5.14 4.07 3.15  SD 2.27 2.13 1.83 2.21 2.61 2.44 1.86 2.21 2.21 2.14 2.37 2.19  Skewness -0.88 -0.05 0.00 -0.29 -0.40 0.73 0.86 -0.56 -0.01 -0.82 -0.18 0.22  Kurtosis -1.25 -1.42 -1.16 -1.82 -1.84 -1.08 -0.51 -1.71 -1.96 -1.22 -1.48 -1.57  Level 3 Total sample minus outlier cases (n = 638) Variables W9 W10 W13 S16 S28 S33 M6 M7 M37  Mean 5.72 5.75 5.42 4.37 4.81 4.36 5.70 5.17 5.12  SD 1.24 1.08 1.24 1.87 1.67 1.69 1.24 1.42 1.39  Skewness -0.98 -0.64 -0.55 -0.27 -0.65 -0.35 -1.03 -0.68 -0.78  Kurtosis 0.55 -0.28 -0.60 -1.14 -0.35 -0.74 0.70 0.17 0.49  Outlier cases (n = 2) Mean 2.00 6.50 5.00 5.00 5.50 5.00 5.50 1.00 2.00  SD 0.00 0.71 2.83 1.41 2.12 2.83 2.12 0.00 0.00  Skewness . . . . . . . . .  Kurtosis . . . . . . . . .  190  Univariate Summary Statistics for Observed Variables for the Student Sample (continued) Level 4 Total sample minus outlier cases (n = 630) Variables HC VER S1 S7 S33 C5 C15 C25 M13 M31 M39  Mean 5.49 4.57 5.65 5.23 4.36 3.90 4.18 4.08 4.95 4.70 5.51  SD 0.69 0.70 1.14 1.45 1.68 1.59 1.54 1.66 1.55 1.48 1.12  Skewness -0.66 -0.03 -0.78 -0.87 -0.36 -0.18 -0.29 -0.28 -0.89 -0.47 -0.58  Kurtosis 0.48 -0.38 0.41 0.54 -0.74 -0.79 -0.58 -0.78 0.26 -0.19 -0.32  Outlier cases (n = 10) Mean 4.79 4.56 5.11 4.44 4.50 3.40 4.70 3.50 4.80 4.20 5.10  SD 0.98 0.89 2.42 2.24 2.46 2.17 2.06 2.42 2.20 2.53 1.91  Skewness -0.28 -0.64 -0.64 -0.12 -0.28 0.50 -0.74 0.15 -0.62 -0.13 -0.18  Kurtosis 0.84 -0.01 -1.88 -1.30 -1.56 -1.34 -0.56 -1.84 -1.04 -1.93 -2.21  Level 5 Total sample minus outlier cases (n = 625) Variables HOR COL S27 S35 C7 C17 C27 M18 M38  Mean 5.38 5.31 6.15 5.52 3.95 3.72 3.96 5.93 6.16  SD 0.60 0.66 1.05 1.21 1.86 1.82 1.86 1.08 1.18  Skewness -0.28 -0.61 -1.26 -0.48 -0.12 0.02 -0.06 -0.87 -1.34  Kurtosis -0.21 0.51 0.92 -0.69 -1.09 -1.03 -1.07 -0.00 0.72  Outlier cases (n = 15) Mean 5.16 4.44 4.40 4.20 3.60 5.47 3.93 5.40 5.07  SD 0.91 1.11 1.92 1.52 2.41 2.17 2.58 1.92 1.82  Skewness -0.82 -0.59 0.74 1.02 0.23 -1.19 -0.03 -0.45 0.06  Kurtosis 0.75 0.28 -1.64 -0.52 -1.88 -0.10 -1.87 -1.97 -2.04  Level 6 Total sample minus outlier cases (n = 627) Variables RH RSD COL S23 S30 S34 C8 C18 C28 M2 M9 M17  Mean 4.79 4.23 5.30 3.97 4.28 3.78 4.27 4.00 3.97 4.74 4.44 5.36  SD 1.33 1.20 0.67 1.82 1.65 1.73 1.71 1.82 1.84 1.56 1.35 1.41  Skewness -0.26 -0.04 -0.68 -0.04 -0.27 -0.02 -0.17 -0.14 -0.10 -0.35 0.17 -0.89  Kurtosis -0.48 -0.23 0.97 -0.96 -0.70 -0.82 -0.83 -1.00 -1.00 -0.32 -0.43 0.69  Outlier cases (n = 13) Mean 4.64 4.31 4.59 3.08 3.31 3.15 4.00 3.77 3.54 5.23 4.62 4.15  SD 1.73 2.21 1.06 2.36 2.14 2.19 2.58 2.74 2.57 2.28 1.85 2.30  Skewness -0.05 -0.24 -0.54 0.79 0.07 0.22 0.03 0.17 0.33 -1.08 0.02 -0.12  Kurtosis -1.60 -1.24 -0.96 -1.02 -1.84 -1.57 -1.86 -2.06 -1.76 -0.21 -1.43 -1.62  191  Univariate Summary Statistics for Observed Variables for the Student Sample (continued) Level 7 Total sample minus outlier cases (n = 613) Variables  Mean  SD  W8 3.98 1.84 W29 4.17 1.90 W33 4.19 1.83 S2 4.32 2.04 S8 4.44 2.10 S21 4.06 2.06 C10 3.70 2.32 C20 3.78 2.40 C30 3.72 2.39 M16 3.49 2.11 M22 3.34 1.99 M34 3.29 1.97 Note. SD = standard deviation.  Outlier cases (n = 27)  Skewness  Kurtosis  Mean  SD  Skewness  Kurtosis  -0.08 -0.24 -0.32 -0.34 -0.36 -0.19 0.15 0.10 0.13 0.25 0.29 0.33  -0.86 -0.96 -0.78 -1.08 -1.14 -1.17 -1.54 -1.60 -1.58 -1.24 -1.18 -1.14  3.44 4.33 4.41 4.37 4.44 3.63 3.41 3.77 3.15 3.92 3.85 3.48  2.29 2.30 2.24 2.31 2.28 2.39 2.34 2.64 2.54 2.71 2.43 2.41  0.36 -0.30 -0.38 -0.39 -0.39 0.20 0.53 0.17 0.68 0.09 -0.06 0.26  -1.42 -1.38 -1.30 -1.34 -1.37 -1.60 -1.27 -1.81 -1.39 -1.86 -1.71 -1.67  192  Appendix M Description of Marker Variables Variables  Statement  Level 1 VI SD S3 S9 S20 M1 M20 M32  Vertical individualism Social Darwinism Being able to take care of myself is a primary concern for me. I prefer to be direct and up-front when dealing with people I‘ve just met. I am comfortable with being singled out for praise or rewards. I do what I feel is best for me, regardless of how that might affect others. I make decisions based on my own needs and desires. In solving problems, I try to protect my own benefits first.  Level 2 W11 W16 W24 S1 S7 S28 C4 C14 C24 M4 M8 M23  Family members must stay together, despite any hardships or sacrifices. Parents should live with their children for as long as possible. It is my responsibility to look after my family no matter what it takes. I respect people who are modest about themselves. I have respect for the authority figures with whom I interact. I will stay in a group if they need me, even when I‘m not happy with the group. Emotional closeness with your family Supportive closeness with your family. Identity closeness with your family. I try not to let my family down in whatever I do. Family relationships are more important than friendships. When I make decisions, I take my family‘s opinions and benefits into consideration.  Level 3 W9 W10 W13 S16 S28 S33 M6 M7 M37  To me, pleasure is spending time with others. I feel good when I cooperate with others. My happiness depends very much on the happiness of those around me. Even when I strongly disagree with group members, I avoid an argument. I will stay in a group if they need me, even when I‘m not happy with the group. I usually go along with what others want to do, even when I would rather do something different. I try to make everyone happy when making decisions. When my desires conflict with my friends‘ benefits, I rather give up my desires than to hurt my friends. I am willing to agree in order to have good relationships with my friends.  193  Description of Marker Variables (continued) Variables Level 4 HC VER S1 S7 S33  Statement  M39  Horizontal collectivism Vertical worldview (vertical individualism and vertical collectivism) I respect people who are modest about themselves. I have respect for the authority figures with whom I interact. I usually go along with what others want to do, even when I would rather do something different. Emotional closeness with your school. Supportive closeness with your school. Identity closeness with your school Students should obey their teachers. We must comply with all the laws, although some of them may seem unnecessary or unreasonable at times. I try to follow the existing rules and laws of the society.  Level 5 HOR COL S27 S35 C7 C17 C27 M18 M38  Horizontal worldview (horizontal individualism and horizontal collectivism) Collectivism (horizontal collectivism and vertical collectivism) It would make me feel good to be a person who is kind and generous. Being someone who is just and fair is an important part of who I am Emotional closeness with your nation. Supportive closeness with your nation. Identity closeness with your nation. In resolving conflicts, each person‘s rights are equally important. All individuals are equal, regardless of religious or spiritual beliefs.  C5 C15 C25 M13 M31  194  Description of Marker Variables (continued) Variables Level 6 RH RSD COL S23 S30 S34 C8 C18 C28 M2 M9 M17 Level 7 W8 W29 W33 S2 S8 S21 C10 C20 C30 M16 M22 M34  Statement  Anti-anthropocentrism Anti-social Darwinism Collectivism (horizontal collectivism and vertical collectivism) I feel connected with animals and plants. I am environmental. Interacting with nature makes me feel connected with the universe. Emotional closeness with nature. Supportive closeness with nature. Identity closeness with nature. Preserving nature is more important than expanding business and consumption. The law of nature is more fair and just than the law of any human society. Protecting the natural environment is our social responsibility.  There are some spiritual laws or being(s) governing the universe. There is a spiritual power or force influencing what happens to people in their daily life. There is a spiritual power or force influencing what happens in the world. I look to my faith as providing meaning and purpose in my life. My faith is an important part of who I am as a person. I look to my faith as a source of comfort. Emotional closeness with God. Supportive closeness with God. Identity closeness with God. Obeying God or other higher being(s) is more important than obeying one‘s parents or man-made laws. I make decisions based on my religious or spiritual beliefs. In resolving conflicts, I rely on the teachings of my religion or spiritual beliefs.  195  Appendix N UBC Research Ethics Board's Certificates of Approval for the Pilot Study  196  197  Appendix O UBC Research Ethics Board's Certificates of Approval for the Student Study  198  199  Appendix P UBC Research Ethics Board's Certificates of Approval for the Adult Study  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0054481/manifest

Comment

Related Items