Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

How do adolescents define depression? Links with depressive symptoms, self-recognition of depression,… Fuks Geddes, Czesia 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_fuks_geddes_czesia.pdf [ 15.11MB ]
Metadata
JSON: 24-1.0066481.json
JSON-LD: 24-1.0066481-ld.json
RDF/XML (Pretty): 24-1.0066481-rdf.xml
RDF/JSON: 24-1.0066481-rdf.json
Turtle: 24-1.0066481-turtle.txt
N-Triples: 24-1.0066481-rdf-ntriples.txt
Original Record: 24-1.0066481-source.json
Full Text
24-1.0066481-fulltext.txt
Citation
24-1.0066481.ris

Full Text

HOW DO ADOLESCENTS DEFINE DEPRESSION? LINKS WITH DEPRESSIVE SYMPTOMS, SELF-RECOGNITION OF DEPRESSION, AND SOCIAL AND EMOTIONAL COMPETENCE by CZESIA FUKS GEDDES R.N., Dawson College, 1973 B.S.W., University of British Columbia, 1987 M.S.W. by Major Research, University of Melbourne, 1997 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Interdisciplinary Studies) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) February 2008 © Czesia Fuks Geddes, 2008 ABSTRACT Depression in adolescents is a ubiquitous mental health problem presenting ambiguities, uncertainties, and diverse challenges in its conceptualization, presentation, detection, and treatment. Despite the plethora of research on adolescent depression, there exists a paucity of research in regards to obtaining information from the adolescents themselves. In a mixed method, cross-sectional study, adolescents (N= 332) in grades 8 and 11 provided their conceptions of depression. Adolescents' self-recognition of depression was examined in association with depressive symptomatology and reported pathways to talking to someone. Adolescents' social and emotional competence was also examined in association with severity of their depressive symptomatology. Developed categories and subcategories of adolescent depression were guided by the Diagnostic and Statistical Manual for Mental Disorders (DSM-IV-TR) criteria for Major Depressive Episode (MDE) (American Psychiatric Association [APA], 2000). Adolescents' definitions of depression were dominated by subjective, holistic interpretations and add new information and depth to the previous research on adolescent depression. Depressed Mood and Social Impairment were the core categories, both contained intricate subcategories. The frequencies of these constructs provide a map of the themes and subthemes that pervade adolescents' personal philosophies regarding adolescent depression. About half of the adolescents who self-recognized depression within two weeks (45%), qualify into screened depression (Reynolds Adolescent Depression Scale -2" version [RADS-2]; Reynolds, 2002) criteria based on the DSM-IV-TR for MDE (APA, 2000). However, this study's findings showed that the mean for screened Depression Total Score (RADS-2; Reynolds, 2002) was significantly higher in those adolescents who self-recognized versus those who did not self-recognize depression. The majority of lifetime self-recognizers of depression thought that they needed to talk to someone and reported that they talked to someone when feeling depressed. Poor Emotion Awareness was a strong contributor to increasing vulnerability to depressive symptomatology. This study provides new theoretical insights regarding the concept and detection of adolescent depression, and links between social and emotional competence and depressive symptomatology. These findings extend previous research (APA, 2000), provide new ii understanding to guide future research, and have direct implications for research, policy, and practice strategies aimed to better communicate with and help young people with and without depression. iii TABLE OF CONTENTS ABSTRACT^ ii TABLE OF CONTENTS^ iv LIST OF TABLES xi LIST OF FIGURES^ xiv ACKNOWLEDGEMENTS xv DEDICATION^ xvi CHAPTER I: INTRODUCTION^ 1 Adolescents' Understanding of Depression^  3 Significance of the Study^ 4 CHAPTER II: LITERATURE REVIEW 6 Organization of Literature Review^ 6 Depression^ 6 Defining Depression^ 6 Classification of Depression 7 Depression in Adolescence  8 Introduction^ 8 History of Adolescent Depression^ 9 Why Examine Depression During Adolescence?^  10 How do Adolescents' Define Depression?  11 Exploring the Meaning of Concepts^  12 Exploring Piaget's Stages of Cognitive Development^  13 Burden of Depression in Adolescents  16 Defining Adolescent Depression^  16 Factors Associated with Depression in Adolescents^  17 Gender^  18 Biology and Environment^  19 Cognitions, Emotions, Context, and Social Factors^ 20 Emergence, Prevalence, and Recurrence of Depression in Adolescence^ 22 Tools Used in Assessing, Measuring, and Diagnosing Depression in Adolescents ^ 24 Rationale for Choice of Measure for this Study^ 25 Treatment for Depression in Adolescents 26 Summary of Findings^ 28 Adolescents' Self-Recognition of Depression and Pathways to Choosing to Talk to Someone When Feeling Depressed 29 iv Introduction^ 29 Adolescents' Self-Recognition of Depression^ 30 Do Adolescents Think About Own Need to Talk to Someone When Feeling Depressed?32 Do Adolescents Talk to Someone When Feeling Depressed?^ 34 Social Cognition^ 36 Depression and Links to Social and Cognitive Development 37 Summary of Findings 37 Precursors to the Development of Depressive Symptomatology^ 38 Introduction^ 38 The Development of Social Perspective Coordination: Its Link to Adaptive Social Functioning and to Depressive Symptomatology^ 38 Understanding Emotions and Links to Depressive Symptomatology^ 41 Defining Emotions^ 42 Development of Emotions 43 Emotional Competence 45 Summary of Findings^ 46 Cumulative Summary of the Literature^ 46 Research Questions and Design 47 Research Questions^ 47 CHAPTER III: METHODOLOGY^ 49 Participants^ 49 Measures 52 Demographic Information^ 52 Social Perspective Coordination and Emotion Understanding^ 52 Social Perspective Coordination^ 52 Self-Awareness of Emotion and Expression of Emotion 54 Emotion Regulation: Sadness 55 Emotion Regulation: Anger^ 56 Depression: Symptomatology, Conceptions, Self-Recognition, Thinking About Own Need to Talk to Someone, and Talking to Someone when Feeling Depressed^ 57 Depressive Symptomatology^ 57 Adolescents' Conceptions of Adolescent Depression^  58 Self-Recognition of Adolescent Depression^  59 Thinking About Own Need to Talk to Someone when Depressed^ 60 Talking to Someone When Feeling Depressed 60 Content Analysis of Participants' COAD^ 61 Coding Schemes^ 61 Classification of Participants' Units of Adolescent Depression^ 62 Intercoder Reliability  68 Procedures^ 75 v Follow-Up Protocol Post Data Administration^ 78 Protocol for Participants Who Were Screened as Depressed^ 79 CHAPTER IV: RESULTS^ 81 Categories and Subcategories of Adolescents' Conceptions of Depression and their Frequencies^ 81 Gender and Grade Frequency Distributions of Conceptions of Adolescent Depression Unit Definitions Generated Per Individual^ 85 Frequency Distributions of Conceptions of Adolescent Depression Units Generated At Least Once Per Individual^ 86 Percentages of COAD Units Generated by Boys and Girls At Least Once in Each Construct of Adolescent Depression^ 91 Gender and Grade Frequency Distributions of COAD Unit Definitions Generated At Least Once Per Individual^ 95 Construct Reduction 96 Summary^ 97 Association of Adolescent Depression Constructs to Grade, Gender, Depressive Symptomatology, Self-Recognition of Depression, and a Pathway to Talking to Someone at the Time of Feeling Depressed^ 98 Association between Adolescent Depression Constructs and Grade^ 99 Association between Adolescent Depression Constructs and Gender  100 Association between Adolescent Depression Constructs and Gender in Grade 8, and in Grade 11^  101 Association between Adolescent Depression Constructs and Grade Level in Boys, and in Girls  104 Association between Adolescent Depression Constructs and Depressive Symptomatology^  106 Association between Adolescent Depression Constructs and Self-Recognition of Depression  108 Univariate Analyses of Self-Recognition of Adolescent Depression^ 109 Association between Adolescent Depression Constructs and Self-Recognized Depression within the Past Two Weeks versus Not Self-Recognized Depression ^ 109 Association between Adolescent Depression Constructs and Self-Recognition of Depression within the Past Two Weeks versus Those Who Self-Recognized Depression Beyond the Two Weeks^  111 Summary^ 112 Association between Adolescent Depression Constructs and Thinking About Own Need to Talk to Someone when Depressed^  113 Association between Adolescent Depression Constructs and Talking to Someone at the Time of Feeling Depressed^  114 Summary^  115 Depressed Mood and its Subcategories ^  115 Social Impairment and its Subcategories  116 vi Remaining Significant Categories and Subcategories^  116 Association of Self-Recognition of Depression to Depressive Symptomatology, and to Thinking About Own Need to Talk and Talking to Someone at the Time of Feeling Depressed^  121 Relations of Social Perspective Coordination, Emotion Awareness, Expression of Emotion, and Emotion Regulation, to Adolescents' Dimensions of Depressive Symptomatology^ 125 Preliminary Analyses^  126 Correlational Analyses  128 Regression Analysis  129 Hierarchical Regression Analysis for Boys Only^  131 Hierarchical Regression Analysis for Girls Only  133 CHAPTER V: DISCUSSION^ 135 Depression: A Voice of an Adolescent in Grade 8^  135 What are Adolescents' Conceptions of Depression?  137 Construction of Adolescents' COAD^ 137 Generation, Distribution, and Range of COAD Units^  138 Grade Distribution of COAD Units  138 Gender Distribution of COAD Units^  139 Generated Constructs of Adolescent Depression  140 Depressed Mood^  141 Sadness.  141 Irritability.  141 Inner Pain^  142 Other Subcategories of Depressed Mood.^  142 Social Impairment  144 The Importance of Social Impairment  144 Differentiating Qualities of Social Impairment.^  145 Other Categories Defined by Adolescents^  146 Remaining Categories and Subcategories in Descending Order of Percentages as Defined by Adolescents' COAD^ 148 Summary^ 149 Associations of Adolescent Depression Constructs to Grade, Gender, Depressive Symptomatology, Self-Recognition of Depression, and Pathway to Talking to Someone When Depressed^  150 Associations of Categories and Subcategories of Adolescent Depression to Grade and Gender  150 Association of Categories and Subcategories of Depression to Grade^ 150 Association of Categories and Subcategories of Depression to Gender  151 Association of Categories and Subcategories of Depression to Gender in Grade 8 ^ 151 Association of Categories and Subcategories of Depression to Gender in Grade 11 152 vii Association of Categories and Subcategories of Depression to Grade Levels for Boys Only  152 Association of Categories and Subcategories of Depression to Grade Levels for Girls Only  152 Association of Depressive Symptomatology to Constructs of Adolescent Depression ^ 153 Self-Recognition of Depression and Its Association with Constructs of Adolescent Depression^  154 Association of Adolescents' Thinking About Own Need to Talk to Someone When Depressed to Constructs of Adolescent Depression^  155 Association of Talking to Someone when Feeling Depressed to Constructs of Adolescent Depression^  155 Similar Percentages in the Constructs Generated by Adolescents' COAD^ 156 Associations of Self-Recognition of Depression to Depressive Symptomatology, and Pathways to Talking to Someone When Feeling Depressed^  156 Association of Self-Recognition of Depression within the Past Two Weeks to Depressive Symptomatology^  157 Lifetime Self-Recognized Depression to Reported Pathways to Talking to Someone When Feeling Depressed^  159 Social Perspective Coordination, Emotion Understanding, and Emotion Regulation, as Predictors of Severity of Depressive Symptomatology^  161 Intercorrelations among Social Perspective Coordination, Emotion Understanding and Depressive Symptomatology^  161 Hierarchical Regression Analyses Predicting Depressive Symptomatology^ 162 Gender Differences in Social Perspective Coordination and Emotion Understanding as Predictors of Severity of Depressive Symptomatology^  164 Study Strengths and Limitations^  165 Study Strengths^  165 Limitations  166 Implications and Considerations for Research, Policy, and Practice^ 169 Research^  169 Policy  172 Practice  172 REFERENCES^ 177 APPENDIX A: Parental or Guardian Consent^ 219 APPENDIX B: Student Assent Form^ 221 APPENDIX C: Demographic Information 223 APPENDIX D: 4+ Relationship Questionnaire (4+ RE1-Q), v. 4.0^  224 APPENDIX E: Conceptions of Adolescent Depression (COAD) 225 viii APPENDIX F: Self-Recognition of Adolescent Depression (S-ROAD)^ 226 APPENDIX G: Thought About Own Need to Talk to Someone 227 APPENDIX H: Talking To Someone When Feeling Depressed^ 228 APPENDIX I: Access Columns: Id (Id Numbers Removed), Responses, and Broken Down Responses into COAD Units^ 229 APPENDIX J: Emerging Category and Subcategory of Adolescent Depression For Grades 8 and 11^ 230 APPENDIX K: Photo Example of Sheet with Clusters of Words ^ 231 APPENDIX L: Clustering COAD Units Within a Category and Subcategory 232 APPENDIX M: Intercoder Sheet With Randomly Selected COAD Units^ 233 APPENDIX N: Intercoder Instructions for Category Alphabet^ 234 APPENDIX 0: Pilot Study^ 235 APPENDIX P: Letter to School Principals^ 236 APPENDIX Q: Parent or Guardian Information Letter^ 239 APPENDIX R: Student Information Letter^ 242 APPENDIX S: Student Identification 244 APPENDIX T: Resource Pamphlet^ 246 APPENDIX U: Protocol For Referral For Suicidality^ 249 APPENDIX V: Protocol for Referral for Depressive Symptomatology^ 251 APPENDIX W: Parental Letter Informing Referral Made^ 255 APPENDIX X: Table X-1: Frequency Distribution and Percentages of Adolescents' COAD Units at Least Once in each Category and Subcategory 256 APPENDIX Y: Table Y-2: Frequency Distribution and Percentages of Greater than or Equal to 5% of Participants Generating a COAD Unit at Least Once in Each Category or Subcategory 257 APPENDIX Z: Table Z-3: Percentages and Chi-Square Tests in Adolescent Depression Constructs by Grade^ 258 APPENDIX Al: Table A1-4: Percentages and Chi-Square Tests of Adolescent Depression Constructs by Gender 259 APPENDIX Bl: Table B1-5: Percentages and Chi-Square Tests for Adolescent Depression Constructs by Gender in Grade 8, and in Grade 11^ 260 APPENDIX Cl: Table C 1-6: Percentages and Chi-Square Tests in Frequency for Adolescent Depression Constructs by Grade Level in Boys, and in Girls^  261 ix APPENDIX Dl: Table D1-7: Percentage Distribution and Chi-Square Tests in Adolescent Depression Constructs by Depressive Symptomatology^ 262 APPENDIX El: Table E1-8: Percentage Distribution and Chi-square tests in Adolescent Depression Constructs by Self-Recognition of Depression in Those Who Self-Recognized Depression in the Past Two Weeks Prior to The Survey Versus Those Who Did Not Self- Recognize Depression  263 APPENDIX Fl: Table F 1-9: Adolescent Depression Constructs by Self-Recognition of Depression Within the Past two Weeks Prior to the Survey Versus Those Beyond Two Weeks264 APPENDIX Gl: Self-Recognition of Depression Time Recall^ 265 APPENDIX Hl: Table H1-10: Percentage Distribution and Chi-Square Tests in Adolescent Depression Constructs by Thinking About Own Need to Talk to Someone^ 266 APPENDIX Il: Table I1-11: Percentages and Chi-Square Tests of Frequency in Adolescent Depression Constructs by Talking to Someone at the Time of Feeling Depressed^ 267 APPENDIX J1: Table J1-12: Significant Chi-square Results for Adolescent Depression Constructs by Grade, Gender, Gender in Grade 8 and in Grade 11, Grade in Boys and in Girls, Depressive Symptomatology, Self-recognition of Depression, Thought of Own Need to Talk to Someone, and Talking to Someone when Feeling Depressed^ 268 APPENDIX Kl: Table K1-13: Association of Self-recognition of Depression and Depressive Symptomatology^ 271 APPENDIX Ll: Emotion Measures Key^ 272 APPENDIX Ml: UBC Ethics Approval for Main Study^ 273 APPENDIX N1: UBC Ethics Approval for Pilot Study 274 LIST OF TABLES Table 1 Developmental Levels of Psychosocial Maturity in Social Perspective Coordination, as also Operationalized in the Relationship Questionnaire^ 39 Table 2 Distribution of Participants by Grade, Gender, and Mean Age, N= 332^ 49 Table 3 Sample Ethnicity (N= 332)^ 50 Table 4 Sample Characteristics, N= 332 51 Table 5 DSM-IV-TR Criteria for Major Depressive Episode (MDE)^ 64 Table 6 Categories, Subcategories, and Examples of Participants' Content Analyzed Conceptions of Adolescent Depression (COAD) for Depressed Mood, N= 332 70 Table 7 Categories, Subcategories, and Examples of Participants' Content Analyzed Conceptions of Adolescent Depression (COAD) for Social Impairment, N= 332^ 71 Table 8 Remaining Categories, Subcategories, and Examples of Participants' Content Analyzed Conceptions of Adolescent Depression (COAD), N= 332^ 71 Table 9 Frequency Distribution of Participants' Response Breakdown of Conceptions of Adolescent Depression (COAD), N= 332^ 82 Table 10 Frequency Distribution and Percentages of COAD Unit Definitions, Total units = 2,074 83 Table 11 Gender and Grade Differences in the Number of COAD Units Generated Per Person, Range of COAD Units, and Means of COAD units^ 85 Table 12 Frequency Distribution and Percentages of COAD Units Generated by Participants at Least Once and the Range of COAD Information Units in Each Category and Subcategory (N =- 332) 87 Table 13 Percentages of COAD Units Generated by Boys and Girls at Least Once in Each Category and Subcategory of Adolescent Depression, N= 332^ 92 Table 14 Percentages of Adolescent Depression Constructs by Grade Level in Boys, and in Girls, N= 332^ 94 Table 15 Gender and Grade Differences in the Number of COAD Units Defined at Least Once per Participant, Range of Units, and Means of COAD units^ 96 Table 16 Percentages, Significant and Near Significant (< .10) Chi-Square Tests in Adolescent Depression Categories and Subcategories by Grade, N= 332 100 Table 17 Percentages, Significant and Near Significant (< .10) Chi-Square Tests in Adolescent Depression Constructs by Gender, N = 332^  101 Table 18 Percentages, Significant and Near Significant (< .10) Chi-Square Tests of Frequency in Adolescent Depression Constructs by Gender in Grade 8, and in Grade 11, N= 332^ 103 Table 19 Percentages, Significant and Near Significant (< .10) Chi-Square Tests in Adolescent Depression Constructs by Grade Level in Boys, and in Girls, N= 332^ 105 xi Table 20 Percentage Distribution, Significant and Near Significant (< .10) Chi-Square Tests of Frequency in Adolescent Depression Constructs by Depressive Symptomatology, N= 332 ^ 108 Table 21 Percentage Distribution, Significant, and Near Significant (< .10) Chi-Square Tests of Frequency in Adolescent Depression Constructs by Self-Recognition of Depression in Those Who Self-Recognized Depression Within two Weeks Prior to the Survey versus Those Who Did Not Self-Recognize Depression, N= 174^  110 Table 22 Percentage Distribution, Significant, and Near Significant (< .10) Chi-Square Tests of Frequency in Adolescent Depression Constructs by Self-Recognition Depression in Those Who Self-Recognized Depression Beyond Two Weeks Prior to the Survey versus Those Who Self- Recognized Depression within the Past Two Weeks, N= 182^  112 Table 23 Percentage Distribution, Significant, and Near Significant (< .10) Chi-Square Tests of Frequency in Association between Adolescent Depression Constructs and Thinking About Own Need to Talk to Someone When Depressed, N= 233 114 Table 24 Percentages and Chi-Square Tests of Frequency in Adolescent Depression Constructs by Talking to Someone at the Time of Feeling Depressed, N= 180^ 115 Table 25 Significant Chi-Square Tests of Frequency in Adolescent Depression Constructs by Grade, Gender, Gender Differences in Grade 8 and in Grade 11, Grade Level Differences in Boys and in Girls, Self-Recognition of Depression, Thought about Needing to Talk to Someone, and Talking to Someone When Feeling Depressed^  118 Table 26 Association between Self-recognition of Depression Within the Past two Weeks and Depressive Symptomatology^  122 Table 27 Descriptives of Measures used in Regression analyses, N= 332^ 127 Table 28 Intercorrelations among Social Perspective Coordination, Emotion Measures, and Depressive Symptomatology^  129 Table 29 Summary of the Hierarchical Regression Analysis for Social Perspective Coordination, Emotion Awareness, Emotion Expression, and Emotion Regulation, in Predicting Depressive Symptomatology, N= 320^ 130 Table 30 Summary of the Hierarchical Regression Analysis for Social Perspective Coordination, Emotion Awareness, Emotion Expression, and Emotion Regulation, in Predicting Depressive Symptomatology in Boys Only, N= 142^ 132 Table 31 Summary of the Hierarchical Regression Analysis for Social Perspective Coordination, Emotion Awareness, Emotion Expression, and Emotion Regulation, in Predicting Depressive Symptomatology in Girls Only, N= 178^  133 Table X-1 Frequency Distribution and Percentages of Adolescents Generating a COAD unit At Least Once in Each Category and Subcategory, N= 332^ 256 Table Y-2 Frequency Distribution and Percentages of Greater than or Equal to 5% of Participants Generating a COAD Unit at Least Once in Each Category or Subcategory, N= 332^ 257 Table Z-3 Percentages, and Chi-Square Tests in Adolescent Depression Construct Categories and Subcategories by Grade, N= 332^ 258 xii Table Al -4 Percentages and Chi-Square Tests of Frequency Distribution for Adolescent Depression Constructs by Gender, N= 332^ 259 Table B 1-5 Percentages and Chi-Square Tests in Adolescent Depression Constructs by Gender in Grade 8 and in Grade 11, N= 332^ 260 Table C1-6 Percentages and Chi-Square Tests of Frequency in Adolescent Depression Constructs by Grade Level in Boys, and in Girls, N= 332^ 261 Table D1-7 Percentage Distribution and Chi-Square Tests of Frequency in Adolescent Depression Constructs by Depressive Symptomatology, N= 332 262 Table E1-8 Percentage Distribution and Chi-Square Tests of Frequency in Adolescent Depression Constructs by Self-Recognition Depression in Those Who Self-Recognized Depression Within two Weeks Prior to the Survey Versus Those Who Did Not Self-Recognize Depression, N= 174^ 263 Table F 1-9 Percentage Distribution and Chi-Square Tests of Frequency in Adolescent Depression Constructs by Self-Recognition Depression in Those Who Self-Recognized Depression Beyond two Weeks Prior to the Survey versus Those Who Self-Recognized Depression Within the Past two Weeks, N= 182 264 Table H1-10 Percentage Distribution and Chi-Square Tests of Frequency in Association Between Adolescent Depression Constructs and Thinking About Own Need to Talk to Someone When Depressed, N= 233^ 266 Table Il-11 Percentages and Chi-Square Tests of Frequency in Adolescent Depression Constructs by Talking to Someone at the Time of Feeling Depressed, N= 180^ 267 Table J1-12 Chi-Square Tests of Frequency in Adolescent Depression Constructs by Grade, Gender, Gender Differences in Grade 8 and in Grade 11, Grade Level Differences in Boys and in Girls, Self-Recognition of Depression, Thinking About Own Need to Talk to Someone, and Talking to Someone When Feeling Depressed^ 268 Table K1-13 Association of Self-recognition of Depression (Not Self-recognized, Self- recognized in the Past Two Weeks, and Self-recognized Beyond the Past Two Weeks) and Depressive Symptomatology^ 271 LIST OF FIGURES Figure 1 School Data Administration^ 78 Figure 2 Self-recognized Depression, Thinking About Own Need to Talk to Someone when Depressed, and Talking to Someone when Feeling Depressed^  124 xiv ACKNOWLEDGEMENTS There are many individuals that enabled my doctoral journey come to its fruition. I am deeply grateful to the colleagues, friends, and family who inspired, supported, and contributed in various ways throughout the process of this thesis to whom the amount of thanks that I owe is immeasurable. It has been my good fortune to work with a doctoral committee that both inspired me and demonstrated remarkable patience in waiting for my research to mutate into form. In particular, this project could not be possible without the support and input from my advisor and mentor, Kim Schonert-Reichl, who has deeply inspired my thinking. Her intellectual generosity, wisdom, and guidance were unending and are gratefully appreciated. Jim Frankish has provided continued support, guidance, and the link to population health promotion research as well as the physical space to contain the project — thank you. Jane Garland's understanding of depression in adolescents has been profound, insightful, and practical and I am deeply thankful for the advice and continued thoughtful direction in this project. I would also like to thank Larry Green who propelled me into the doctoral program and provided the initial foundation and gentle guidance, and opened the door to health promotion thinking and space in which to complete this thesis. Bill Reynolds's directive support and advice in assessment criteria for depression in adolescents at the initial phase of the process has been invaluable. There are also a number of people who have facilitated the process in different capacities to whom I extend my warm gratitude. The Interdisciplinary Studies' unique environment for graduate study made possible for my project to take on interdisciplinary flight. Bruno Zumbo has provided amazing methodological advice at a critical time early in the study. Phil Davis has provided statistical guidance and gentle support at the onset of the research journey. Amanda Walker has been my other set of eyes during data administration whose consistency, integrity, and support are appreciated. To my friends, Leslie, David, and Marnelle who would always be ready to listen and provide support particularly in the first phase of the project. I am thankful for my siblings for their pride, and of course my family, Ross, Maiya, and Rebecca who are unwavering in their belief, presence, and love. Finally, I would like to acknowledge the support of the teachers, the school personnel, and above all, the adolescents in this study, who I know, worked so hard at providing me with their written voices to help make this study a reality. XV Dedicated to my daughters, Maiya and Rebecca thank you for helping me in keeping awake the sense of wonder in the world xvi Chapter I: Introduction CHAPTER I: INTRODUCTION Depression in adolescents is a ubiquitous mental health problem presenting ambiguities, uncertainties, and diverse challenges in its conceptualization, presentation, detection, and treatment (American Academy of Child and Adolescent Psychiatry [AACAP], 1998; Birmaher et al., 2007; Cicchetti & Toth, 1998; Good & Kleinman, 1985; Kazdin, 2001, 2002; Olfson, Gameroff, Marcus, & Waslick, 2003; Rounsaville, Andrews, & Kendall, 2002). At present, the diagnosis of adolescent depression in North America is based on adult depression criteria found in the Diagnostic and Statistical Manual of Mental Disorders - fourth edition - Text Revision (DSM-IV-TR) (American Psychiatric Association [APA], 2000). There is a plethora of research suggesting that depression often begins in adolescence and can recur in adulthood (Birmaher, Ryan, Williamson, Brent, & Kaufman, 1996a; Costello, Foley, & Angold, 2006; Kim-Cohen et al., 2003; Lewinsohn, Rohde, Seeley, Klein, & Gotlib, 2000a; Pine, Cohen, Gurley, Brook, & Ma, 1998). Research examining the prevalence of mental disorders in community-based samples in the US and Canada indicates that about one in five adolescents in the community exhibits mental disorders (Health Canada, 2002; Michalak, Goldner, Jones, Oetter, & Lam, 2002; Schonert-Reichl & Offer, 1992). Goldberg and Goodyer (2005) report that depressive disorders markedly increase in prevalence between the ages of 13 and 16 years, and continue to rise during adolescence into young adult life. Depression is increasing in prevalence in developed countries (World Health Organization [WHO], 2001) and poses a significant societal concern (Waddell, Offord, Shepherd, Hua, & McEwan, 2002; Waddell & Shepherd, 2002; WHO, 2003). Young people who experience depression demonstrate an increased risk of suicide and injury and are more likely to engage in harmful behaviors, such as disordered eating, the use of tobacco, alcohol, and drugs or controlled substances (Birmaher et al., 2007; Patton et al., 1998; Rutter, 2000). Moreover, the risk of depression also increases in the absence of social support (Brown & Harris, 1978; Kandel & Davies, 1986; Wong & Wiest, 1999), and individuals with depression exhibit social isolation (Kandel & Davies, 1986). Taken together, researchers are in accord in suggesting that depression in adolescence has both short-and long-term consequences for the adolescent and society as a whole. 1 Chapter I: Introduction Immense disparities exist between the modest use of support services and the epidemiological findings of adolescent depression and other mental disorders in the community. Studies in the US and Australia indicate that about two out of three adolescents with mental disorders do not receive services from a health practitioner (Cauce et al., 2002; Patton, Hetrick, & McGorry, 2007). Canadian data on access to support services show similar figures to those found in Australian and US studies. Specifically, about 60 percent of people with mental health problems do not receive services from a health practitioner (Stephens & Joubert, 2001). Treatment services to identify those most in need of intervention are fragmented, limited in their capacity, and in general do not facilitate access to care. Takanishi (2000) indicates that contemporary social interventions rarely aim to prepare adolescents for the rapid and often unpredicted social change they encounter in modern society. The accepted approach to assessing who merits mental health care in the community has been to provide prevalence estimates based on accepted categorical or threshold diagnostic criteria for mental disorders as identified via the DSM-IV-TR (APA, 2000). Threshold is representative of a division between "cases" and "non-cases," which presently clinicians are obliged to use, as they must decide who merits treatment (Goldberg, 2000). Subthreshold depression exists between threshold depression and an asymptomatic state (Fergusson, Horwood, Ridder, & Beautrais, 2005), and is representative of dimensions of depressive symptoms below number and/or duration criteria for depression according to the DSM-IV-TR (APA, 2000; Judd & Akiskal, 2000). Making decisions about who "needs" mental health care based solely on the DSM-IV-TR (APA, 2000) diagnoses does not seem to be optimal (Sareen, Stein, Campbell, Hassard, & Menec, 2005a) because research indicates that depression may best be conceptualized on a continuum consisting of several dimensions (Akiskal, Judd, Gillin, & Lemmi, 1997; Goldberg, 2000; Judd & Akiskal, 2000; Lewinsohn, Solomon, Seeley, & Zeiss, 2000b; Slade & Andrews, 2005). The costs of depression to the individual and to society are high, not only in terms of the impact on affected individuals and their families, but also in terms of disability, diagnostic costs, treatment, and the resultant loss of productivity. In a Canadian report, Stephens and Joubert (2001) specified that the direct and indirect financial costs related to mental health problems in 1998 was $14.4 billion annually, almost double the amount since 1993, and the figures are 2 Chapter I: Introduction expected to continue to rise. Better means of communication with young people, early detection, early intervention, and treatment would greatly reduce overall costs. Adolescents' Understanding of Depression Despite the plethora of research examining the emergence, course, and treatment of adolescent depression, one area in which there exists a paucity of research is in regards to obtaining information from the adolescents themselves with respect to how they understand and define depression. Studies that focus on adolescents' perceptions of their health concerns show that mental health is prominent among them (Goodman et al., 1997; Millstein & Litt, 1990; Tonkin, 2001). Virtually all research and practice on adolescent depression is bound by the DSM-IV-TR (APA, 2000) or the International Classification of Diseases (ICD-10) (WHO, 1992) criteria for depression, yet research on adolescents' understanding of depression is unexplored. There exist three ways in which our understanding of adolescent depression could be furthered. First, insight into the roots of adolescent depression may be found by examining adolescents' conceptions of depression — that is, how they define "depression." It is well documented that the stage of adolescence is marked by rapid physical and social changes, increased cognitive capacity, and the development of differentiated and abstract cognitive abilities (Chapman, 1988; Inhelder & Piaget, 1958). The present study provides a first step in our efforts to understand how adolescents "define" depression and how their conceptions vary by age and gender. A second important factor that could expand our understanding of depression in adolescents may be found by assessing whether or not adolescents can recognize depression in themselves — that is, their self-diagnosis. Adolescents' conceptions of depression can be intertwined with their understanding of their vulnerability. The ability to recognize depression may be influenced by the adolescents' interpretations and processing of information by capacities and competencies in their emotional growth, in their social relationships, in their conceptions and reasoning regarding their social world, and changes in their environment. Adolescents' tendency to initiate or experiment with different forms of behavior that are harmful to their health may be linked to adolescents' understanding of their own vulnerability to depression intertwined with their emotional, cognitive, social capacities, and experiences. The immense disparity between the epidemiological findings of adolescent depression and adolescents' help-seeking and access to 3 Chapter I: Introduction support may be moderated by their conceptions and self-recognition of depression. The present investigation aims to examine adolescents' self-recognition of depression and how they seek help for depression in association to depressive symptomatology. The goal of this investigation is to examine the relevance of self-recognition to depressive symptomatology, and to adolescents' approach-oriented coping specifically via thinking about own need to talk to someone and talking to someone when feeling depressed. The purpose of using these variables is to examine their relationship to depressive symptomatology which is used to screen depression in adolescents in order to offer access to support services for depression. A third way to gain insight into the roots of adolescent depression may be found in examining adolescents' social and emotional understanding in relation to the depressed condition. An individual's capacity to differentiate and integrate the viewpoints of the self and others comes about through an understanding of the relation between the thoughts, emotions, and wishes of each person (Schultz & Selman, 2000a). Further, emotions are critical in organizing the development of social relationships and physiological experiences, as well as cognitive processes, and have been associated with social competence and physical health (Saarni, Mumme, & Campos, 1997; Zeman, Shipman, & Penza-Clyve, 2001). Difficulties in successfully handling relationships, emotions, and issues at this phase of their lives may contribute to greater vulnerability to depression. If we know more about the underlying social and emotional processes and mechanisms of depression in adolescence, we would have more information that may inform early detection, early intervention, and treatment for adolescents. Researchers have only recently begun to examine depressive symptomatology in relationship to adolescents' social or emotional development (Zeman, Cassano, Perry-Parrish, & Stegall, 2006). Hence, a third aim of this research is to explore the relative influence of both social perspective coordination and emotion understanding in relation to depressive symptomatology in adolescents in order to determine the potential influence of these constructs in predicting depressive symptomatology. Significance of the Study The true magnitude of depression during adolescence and its deleterious effect on adolescents themselves, their families, school environments, and the community, is now generally recognized. The incidence of depression is expected to rise during the next 15 years and is predicted to be the leading disability in both men and women (Michaud, Murray, & 4 Chapter I: Introduction Bloom, 2001; National Institute of Mental Health [NIMH] [U.S.], 2002; WHO, 2001). As well, there is a rising trend in depression emerging in adolescence, (Kessler et al., 2003). Effective early strategies for identifying adolescents with depression, and implementing timely interventions, may be hindered, due in part, to ambiguities around the definition and etiology of depression in adolescence. Depression in adolescents might be a barometer used to gauge future patterns of population health because depression can have a profound impact on social relationships, educational attainment, or subsequent employment, as well as the health risk behaviors they elicit. Epidemiological findings highlight the importance of differentiating between risk indicators and risk mechanisms (Achenbach, Howell, McConaughy, & Stanger, 1998; Birmaher et al., 1996b; Lewinsohn et al., 1994) but do not reflect the processes that mediate that risk; namely, the individual. Asking adolescents to provide their own conceptions of depression and how they might recognize depression in themselves can bridge the gap that reflects the processes that mediate that risk. A more complete understanding of how adolescents conceptualize and recognize depression can pave the way to greater understanding of depression and other forms of mental illness, improve communication with them, and uncover possible insights regarding their pathways to seeking support when feeling depressed. Moreover, better communication with adolescents could promote health, improve early detection and intervention, and point to preventative measures. Identifying pathways adolescents follow in talking to someone could inform researchers about adolescents' understanding and coping styles in dealing with depression. Examining emotional and social development in relation to depressive symptomatology can deepen our understanding of mechanisms and processes that underlie depression in adolescence, which, in turn, can inform programming in health promotion, early detection, early intervention, and treatment of depression in young people. 5 Chapter II: Literature Review CHAPTER II: LITERATURE REVIEW Organization of Literature Review The literature review is organized into five main sections. First, an overview of the research on depression in general is presented. In the second section, an overview of the literature on depression in adolescence is reported. Included in this section are: the history of adolescent depression, factors associated with depression in adolescence, and literature that may link with adolescents' conceptions of depression. Prevalence, assessment, treatment, and burden of depression in adolescence are also discussed in this section. Next, an overview of the literature on the relation of adolescents' self-recognition of depression to social cognitive theory is put forth. This section continues with the review of the literature on the pathways to seek support for mental disorders, which include approach-oriented coping via thinking about own need to talk and the action of seeking support. In the fourth section, theories of social perspective coordination and emotions in relation to depressive symptomatology are discussed. Finally, I present the cumulative summary of the literature and this study's research questions. Depression Defining Depression Studies on the concept of adolescent depression abound (APA, 2000; Brockless, 1997; Reynolds, 1994a; Rutter & Sroufe, 2000; Street, Sheeran, & Orbell, 1999), yet the understanding of the condition has been impaired by definitional tensions, and other uncertainties. Researchers and practitioners actively debate the definitions of depression (Good & Kleinman, 1985; Kangas, 2001; Rounsaville et al., 2002). One major challenge is how to operationally define and measure depression (Compas & Oppedisano, 2000). Defining depression in adolescents involves assessing or distinguishing its specific features. Measuring depression involves grouping cases according to those features (Compas & Oppedisano, 2000). Both definitions and measurements are susceptible to conceptual bias (Rutter, 2000). Researchers and practitioners use different labels to define depression, including depressed mood, depressive syndrome, and depressive disorder. Depressed mood refers to depression viewed as affect (Angold, 1988a) -- unhappiness, sadness, or blue feelings -- for an 6 Chapter II: Literature Review unspecified period of time. It may be measured by self-report (Duffy, 2000; Garland, 1994, 1997). Depressive syndrome (Petersen et al., 1993) is defined as a set of depressive and other symptoms occurring simultaneously (Cicchetti & Toth, 1998). Those symptoms may be measured using adolescents,' parents,' or teachers' ratings. The boundaries of depressive disorder are delineated according to specific, clinically-derived standards of diagnostic criteria and involve a constellation of disturbances in depressed mood, behavior, somatic, and cognitive functioning (APA, 2000). Depressed mood seems to be the core component being included in both depressive syndrome and in clinical depression (Wichstrom, 1999). Hence, depressed mood may precede a depressive episode and carry the risk of depressive disorder (Hankin & Abramson, 2001; Harrington & Wood, 1995). Further debate exists among researchers in distinguishing whether depression is "threshold" (a case versus not a case) depression according to the criteria of the DSM-IV-TR (APA, 2000) or depression exists on a continuum of increasing levels of depressive symptomatology (Judd & Akiskal, 2000; Lewinsohn et al., 2000a; Lewinsohn et al., 2000b; Ruscio & Ruscio, 2000; Skodol, Schwartz, Dohrenwend, Levav, & Shrout, 1994). Research findings indicate that dimensional rather than "threshold" or categorical criteria seem to be the more appropriate approach for a diagnosis (Goldberg, 2000; Slade & Andrews, 2005). A number of researchers have shown that depressive symptoms exist on a continuum (Judd & Akiskal, 2000; Lewinsohn et al., 2000a; Lewinsohn et al., 2000b; Ruscio & Ruscio, 2000; Skodol et al., 1994). Having examined the range of subthreshold depressive symptomatology, they have found that it could be as low as two depressive symptoms lasting two or more weeks (Fergusson et al., 2005; Judd & Akiskal, 2000; Judd, Rapoport, Paulus, & Brown, 1994), which is below the required number to warrant a diagnosis of Major Depressive Episode (MDE) or Major Depressive Disorder (MDD) (APA, 2000). The increasing levels of subthreshold depressive symptoms, such as depressed mood, are associated with the increasing levels of psychosocial impairment and the emergence of threshold or major depression (Angst & Merikangas, 1997; Georgiades, Lewinsohn, Monroe, & Seeley, 2006; Lewinsohn et al., 2000b). Classification of Depression Prior to the development of the standardized DSM criteria, which allow researchers and practitioners to identify, diagnose, and treat depression, there was an absence of a generally 7 Chapter II: Literature Review agreed-upon set of guidelines clarifying adolescent symptoms of depression (Compas & Oppedisano, 2000). Currently, two major diagnostic systems exist, the International Classification of Diseases and Related Health Problems -- 10 th revision (ICD-10) (WHO, 1992) and the DSM-IV-TR (APA, 2000). The DSM-IV-TR (APA, 2000) is used predominantly in North America and has similarities with ICD-10 (WHO, 1992), which is most commonly used outside North America. Since the introduction of the third edition of the DSM (DSM-III) (APA, 1980),' the criteria for making a diagnosis of depressive disorders, in terms of symptomatology and their duration in adolescents, have been more explicitly and clearly described (Essau, Conradt, & Petermann, 1999). Although not systematically tested in research studies, the formal and informal classification systems implemented by counsellors, social workers, psychiatrists, psychologists, and educators have played a central role in defining the field (US Department of Health and Human Services, 1999). In the last three decades, definitional problems with the diagnostic criteria have been addressed to develop a reliable and meaningful classification system (US Department of Health and Human Services, 1999). Prevailing practices relating to intervention, professional training and certification, and funding decisions are shaped by the ways in which individuals are described and studied (AACAP, 1998). One concern with the current classification and diagnostic systems is that there is still a large unexplored territory in the developmental, contextual, and relational parameters that characterize depression in adolescents (Cicchetti & Toth, 1998). Another concern related to the contemporary nature of the DSM-IV-TR (APA, 2000) criteria for depression is that subthreshold presentations may go unrecognized and therefore be denied access to support services. Depression in Adolescence Introduction In a paper on adolescent research priorities, Zaslow and Takanishi (1993) emphasize two research needs: "(a) the need to collect data that permit adolescents to describe their own perceptions of their own experiences and (b) the need for interdisciplinary research" (p. 190). ' The ICD included a separate category for mood disorders since 1980, and before then, depression and manic states were included under neurotic or psychotic conditions (Klerman & Weissman, 1988). 8 Chapter II: Literature Review This approach highlights the philosophy of this investigation to move from reliant research and practice assumptions about adolescent depression towards greater collaboration with the adolescents themselves from adolescents' understanding of depression. An extensive review of the extant literature reveals that adolescents' definitions or conceptions of depression are unexplored. This section includes previous research on the history, definitions, classification, epidemiology, assessment, treatment, developmental findings, and other factors associated with adolescent depression, which may inform adolescents' understanding of the condition. History of Adolescent Depression In the mid-twentieth century, the definition of depression was mostly posited by psychoanalytic theorists who maintained that depression was the product of a persecutory superego (Rie, 1966). 2 According to this theory, the superego lacked maturity in childhood and emerged during adolescence as a discrete and uncommon phenomenon (Angold, 1988a; Cicchetti & Toth, 1998), hence adolescent depression was considered rare. Adolescence was seen as a period of emotional upheaval, but significant, persisting mental health problems were thought to be uncommon in this age group (Offer, 1969; Petersen et al., 1993). It was also in the twentieth century that certain observed, unusual mental states began to be recognized as depression in general and treated on a more consistent basis (Jackson, 1986). For example, Adolf Meyer (1957) redefined the constituents and terminology of melancholia, nominating the terminology of depression instead (as cited in Jackson, 1986). In the seventeenth century and before, with the exception of its occasional use, when depression was applied to signify the lowering of spirits or dejection, melancholia was the term most often used to describe depression, in fact, descriptions by Hippocrates date back about 2500 years (Jackson, 1986). A watershed in the evolution of depression as a diagnostic classification occurred in 1971. The theme of the Fourth Congress of Paedopsychiatrists was "Depressive States in Childhood and Adolescence," where it was concluded that depression was a significant and relatively large-scale disorder in this population (Weller, Weller, & Fristad, 1984). A subsequent conference on childhood depression sponsored by the Center for Studies of Child and Family Mental Health at the National Institutes of Mental Health (NIMH) expanded the awareness of 2 Psychoanalytic theorists stressed the importance of adolescent's unconscious thoughts. 9 Chapter II: Literature Review depression in children and adolescence (Weller et al., 1984). Consequently, there was an increase in research in the diagnosis and treatment of depression in young people. In the 1980s, as the research findings on adolescent depression were accumulating (Angold, 1988a, 1988b), more appropriate diagnostic categories and criteria evolved in the Diagnostic and Statistical Manual of Mental Disorders (DSM-III; DSM-III-R; DSM-IV) (US Department of Health and Human Services, 1999). The ICD (WHO, 1992) also included a separate category for affective (mood) disorders (Klerman & Weissman, 1988). The trend began to change in diagnosing depression as a result of refined categories and criteria appropriate for children and adolescents (US Department of Health and Human Services, 1999). In the past two decades, the areas of research, treatment, and self-help groups have grown considerably. Why Examine Depression During Adolescence? Adolescence is a distinct and dynamic period within which increased complex developmental changes and processes take place, including changes in biology, cognitive, social and emotional understanding, school, peer, and family relations (Compas & Oppedisano, 2000; Selman, 2003; Spear, 2003). Adolescence is a critical transitional period in the life course, in which young people make strides by becoming self-aware (Eccles, 1999), testing and rehearsing their social roles and skills beyond their families on the path to adulthood. Transitions from primary to secondary school can bring losses of continuity in relationships, along with changes in curriculum and school culture (Hargreaves, Earl, & Ryan, 1996). Debates continue on the relative importance of the home environment versus schooling and peer relationships, but all three are significant (Rutter, 2000). A personal sense of self is not only critical in self-understanding but also necessary for self-preservation in times of social change (Ball & Chandler, 1989; Chandler, 1994; Chandler & Lalonde, 1998; Noam, Chandler, & Lalonde, 1995). Changes in social life present adaptive challenges that may impact or overburden the coping capacities of some adolescents (Crockett & Silbereisen, 2000). Influences interfering with the acquisition of new roles and skills can negatively influence the adolescents' thinking, inducing vulnerabilities to depression, with wider effects on interpersonal functioning (Spear, 2003). The age of adolescence is commonly considered to be from 12 to 18 years, with less agreement on the margins outside this age range (Spear, 2003). Hall (1904) suggests that the 10 Chapter II: Literature Review construct of adolescence emerges from a confluence of social circumstances. Fabrega and Miller (1995) report that the notion of adolescence, as a distinct stage in the life cycle, evolved out of the historical processes during the late 19th century, including population growth, industrialization, and urbanization in developed countries. With the increased demand for technical skills, education became more widespread, making young people more valued. This distinct period was furthered by separating young people from families and removing them from work responsibilities, in turn, creating a specific group characterized by distinctive outlooks, expectations, conflicts, social adjustments, and maladjustments. Cicchetti and Toth (1998) maintain that to comprehend human development, it is critical to understand the integration of developmental processes at multiple levels. Adolescence is a key transitional period in which adult roles are developed and tested. Factors interfering with the acquisition of these roles and skills may have negative effects on vulnerabilities to depression, occupational, interpersonal, and emotional functioning. How do Adolescents' Define Depression? Despite the vast and evolving research in adolescent depression, little is known from adolescents themselves about depression. This author has found no previous research that explored adolescents' conceptions of depression, which might help identify the structure of what depression means to them and detect factors associated with it.' Researchers tend to concentrate on differentiating between risk indicators and risk mechanisms, yet the processes that mediate that risk, the individual, have been largely ignored. One study, which examined understanding and self-identification of depression in adults with brain damage, showed that discrepancies in clients' and practitioners' identifications of depression existed in 44 percent of the cases (Kalpakjian, Lam, & Leahy, 2002). Kalpakjian's et al. (2002) research was based on 37 individuals receiving brain injury treatment rehabilitation services for at least three weeks and 18 clinicians in various disciplines other than psychology, who provided the treatment in the same programs. The most common disagreement was that in 3 This author has conducted searches of various databases such as Psychinfo, Medline, and Embase, mapping to keywords including 'adolescents or adolescence,"depression,"conceptions,"conceptualizations,' `depressive disorder,' `mood disorders,' `depressed mood,' `depressive syndrome,' `mental illness,' and 'mental health.' Internet searches were also conducted using Google Scholar. 11 Chapter 11: Literature Review which a clinician identified an individual as depressed, whereas the person did not self-identify as such. Further, in an open-ended format, clients and clinicians were asked to provide reasons for their identification of themselves or their client as depressed. Their comments were grouped into three categories paralleling the subscales of the instrument utilized to measure their depressive symptomatology (e.g., Cognition/mood, Anxiety/arousal, Vegetative signs). The fourth category contained "Other" comments that did not fit the three categories. There were differences noted between the clients' and clinicians' comments across all categories, however almost half (47 percent) of the clients generated comments clustered in the "Other" category, and no significant details were reported on these differences. Similar to the example above, asking adolescents to define depression gives voice to the young person's vulnerability to depression that may be distinct from research on adolescent depression. Subjective elements may also influence depressive thinking in adolescents (Cicchetti & Toth, 1998). Understanding these elements may be intertwined with one's vulnerability and experiences before the onset of depression. Exploring the Meaning of Concepts Because adolescents' conceptions of depression are being explored, I include a review of the literature on conceptions, as well as a review of cognitive development at the adolescent stage. In the mid 1750s, Hume defined conceptions as being derived from perceptions (objects, events, or other people we perceive). The term perception referred to a category he used to distinguish between impressions and ideas (Hume, 1998). Impressions are the materials first presented to the mind. Impressions are defined as original perceptions (original sensations, passions, and emotions). Ideas are faint copies or faint images of impressions. Ideas are derived from impressions and become the instruments for imagining, thinking, and symbolizing (Hume, 1998). In "Studies in Cognitive Development," Elkind (1969) described four facets of "a conception of a concept" (p.186): 1) the nature of environmental variability or variations between and within things (e.g., dogs, cats); 2) the essence of the concept or the commonality among things; 3) the major function of the concept, which includes the discrimination of "between-and within-things types of variability" (p. 187); and 4) the content of the concept. 12 Chapter II: Literature Review Rutter and Sroufe (2000) report that "...it makes no sense to ask what a word means. Rather, we have to ask what is the concept or idea to which we wish to attach the descriptive term" (p. 265). Piaget (1929) wrote "a word is always associated with its context until it comes to be regarded as implying the whole context" (p. 84). In order to identify the processes that underlie adolescents' conceptions of depression, it is necessary to gain a better understanding of cognitive development. Conceptualizing is an ongoing process that is being continually refined and upgraded by development, experience, and education. Conceptualizing cannot go on without memory and learning. From early adolescence onwards, cognition 4 tends to involve abstract, self-reflective, self-aware, and multidimensional interpretations (Keating, 1990). Piaget's theory of cognitive development (Inhelder & Piaget, 1958) proposed that adolescence is a time when the individual is able to generate hypotheses and possible solutions in terms of past and potential experiences. Cognitive processes are used for interpreting, understanding, and evaluating one's environment, observations, and interactions with others (Keating, 1990; Noam et al., 1995; Santrock, 1998). Hence, all developmental changes and all learning come about through the modification of structures already present in the individual as a result of one's interaction with the environment. Exploring Piaget's Stages of Cognitive Development Adolescents' conceptions of depression serve to codify certain formulated ideas (or thoughts of depression). Those ideas, according to Piaget's stages of cognitive development (Inhelder & Piaget, 1958), tend to be constructed during the formal-operational stage. For the purposes of this discussion, the formal-operational stage of cognitive development is Piaget's fourth and final level of thinking. In this stage, from age 11 to adulthood, as posited by Piaget, adolescents are able to think and reason about their own thoughts as well as the thoughts of other people. The formal operational thinkers can also conceptualize the real and the possible, be more flexible and more abstract than individuals in the preceding stages (Inhelder & Piaget, 1958; Keating, 1990). 4 Webster defines cognition as knowledge, the act or process of knowing, perception, the product of such a process, something thus known, perceived (Random House, 1996). 13 Chapter II: Literature Review Piaget's previous phase to the formal-operational stage of cognitive development involves concrete operational thought.' In this stage, usually occurring around age 7 to 11, children have intuitive thought as long as the reasoning can be applied to specific or concrete examples, and there is cooperation of rules. However those children give contradictory accounts when questioned separately (Broughton, 1979; Chapman, 1988; Santrock, 1998). At the concrete operational stage, the child can perform mentally what was done before physically (Santrock, 1998). According to Piaget, knowledge is a biological structuring process or equilibration where each individual experiences the different stages of development that correspond to their ages only as temporary resting points (Broughton, 1979). Each equilibration is partial. For every achieved equilibration, there is a higher, more complex form of equilibrium toward which the individual strives to evolve (Chapman, 1988) to maintain equilibrium with the environment. Chapman (1988), in his discussion of Piaget's work (Piaget, 1959), maintains that the individual "while thinking that he has knowledge of the people and things as they are in reality attributes to them not only their objective characteristics but also qualities which come from the particular aspects of things of which he is aware at the time" (Chapman, 1988; Piaget, 1959, pp. 162-163). Piaget's (1896-1980) life work involved asking children what they thought of things and how they defined things from a constructivist point of view. In essence, Piaget was trying to find out what is knowledge and how we acquire it, a cognition that must be considered a process of subjective construction (Glasersfeld, 1978). Piaget postulates that rather than a static hierarchy of structural processes, the stress is on continuity between stages, where the stages are only temporary resting points (Broughton, 1979; Inhelder & Piaget, 1958). Cited in Chapman (1988), Piaget believed that thought precedes action and that the logic of action exists prior to and in addition to the logic of thought. The cognitive operations within each level of development (stage) are organized in a structure (Fodor, 1988; Keating, 1990). The structures bring about constructions that raise new possibilities and lead to the next stage and the new structures (Fodor, 1988). In other words, changing stages involves shifts in the underlying structure, brought about 5 P iaget ' s first stage includes sensorimotor intelligence (before age 3; e.g., children manipulate their marbles according to their own individual desires and motor habits). The second stage incorporates symbolic, intuitive, pre-logical thinking (approximate age 3-7; e.g., children become aware of codified rules but play to his or her own rules (Broughton, 1979; Chapman, 1988). 14 Chapter II: Literature Review through the constructive interaction of the individual with the physical and social world (Keating, 1990). For every achieved equilibration, there follows a movement toward the higher level of complexity (Chapman, 1988). The newly achieved equilibrated stage encompasses the ability to coordinate more perspectives (Noam et al., 1995). Different views pervade Piaget's stages of thinking, particularly in relation to formal operational thought. Some researchers posit that there is an early and a late phase to formal operational thought (Flavell, 1992a; Santrock, 1998). In the early phase, adolescents gain increased ability to think hypothetically. In the late phase, there is a restoration of intellectual balance, where adolescents can test out the products of their reasoning against experience (Broughton, 1983; Santrock, 1998). Other researchers indicate various shortcomings in Piaget's theory: it does not account for unique differences (Overton & Byrnes, 1991); most individuals in early adolescence have not reached formal operational thought (Strahan, 1983); some cognitive abilities emerge earlier than expected (Flavell, 1992b); or some adults reason at concrete operational thinking (Siegler, 1996). Some theorists suggest that the formal operations thinking can be understood better as different ways of "knowing" (Byrnes, 1988; Keating, 1988). However, according to Keating (1990), from early adolescence on, "thinking tends to involve abstract rather than merely concrete representation; to become multidimensional rather than limited to a single issue; to become relative rather than absolute in the conception of knowledge; and to become self-reflective and self-aware" (p. 62). Hence, seeking out adolescents' voices in defining depression can help to advance our knowledge of depression in adolescence. Before we can develop any concept of depression or assign anything to categories, first we have to perceive the existence of such a concept, and then we try to classify it within our existing concepts. If it is an entirely new perception, we may have to form a new concept. The process of coding is independent from identifying the physical or mechanical characteristics of the conceived items (Glasersfeld, 1978). Given the way experiences have been segmented, the frequency with which a construct is repeated and the greater number of larger concepts in which it exists, the more indispensable it becomes in sorting out experiences that can be known (Glasersfeld, 1978). 15 Chapter II: Literature Review Burden of Depression in Adolescents Depression that begins in adolescence often recurs in adulthood (Birmaher et al., 1996b; Costello, Foley, & Angold, 2006; Federal/Provincial/Territorial Advisory Committee on Population Health [ACPH], 2000; Feehan, McGee, & Williams, 1993; Fombonne, Wostear, Cooper, Harrington, & Rutter, 2001; Kessler, Avenevoli, & Merikangas, 2001; Kim-Cohen et al., 2003; Lewinsohn, Allen, Gotlib, & Seeley, 1999; Lewinsohn et al., 2000a; Lewinsohn et al., 2000b; Millstein, Petersen, & Nightingale, 1993; Parker & Roy, 2001; Rutter, Kim-Cohen, & Maughan, 2006). As the new millennium unfolds, we are confronted with the dramatic forecasts of the global increase in depression in the next 15 years. It is predicted to be the leading cause of disability in both men and women, next to heart disease (Michaud et al., 2001; NIMH (U.S.), 2002; WHO, 2001). The conference of the Canadian Economic Roundtable on Depression (2001) identified depression as a major cause of worker disability and corresponding productivity loss (Ministry of Health Services, 2002). Depression and stress disorders at work account for 30 percent of all disabilities recorded at major corporations (Ministry of Health Services, 2002). Based on the National Comorbidity. Survey, Kessler et al. (2003) report that depression is a growing public health problem: 50 percent of the respondents born between 1965 and 1974 had their first Major Depressive Episode (MDE) by the age 18 in contrast to 20 percent of participants born before 1965. In 1993, the economic costs of mental disorders in Canada was estimated at just over $7 billion (Health Canada, 2002; Stephens & Joubert, 2001), and it has about doubled in 1998 (Stephens & Joubert, 2001). The figures may be even higher if one is considering the cost of production loss. The fact that the rate of increase is expected to continue suggests that depression in adolescence is part of a heavy burden extending beyond the affected individuals and their families, to a diminished quality of life. Defining Adolescent Depression During the last four decades, the defining characteristics of adolescent depression have not originated in research studies of adolescents; instead, the criteria for diagnosis have been developed and validated in adults, with age-specific modifications (APA, 2000; Birmaher et al., 1996b; Gotlib & Sommerfeld, 1999). Further, research has shown that adolescents exhibit diagnosable depression, but many manifest more irritability than sadness (APA, 2000; Birmaher 16 Chapter II: Literature Review et al., 1996b; Pataki & Carlson, 1995; Rutter, 2000). However, several studies reveal that some depressive symptomatology in adolescents do reflect diagnostic criteria according to the DSM or the ICD-10 (Patton, Coffey, Posterino, Carlin, & Wolfe, 2000; Roberts, Lewinsohn, & Seeley, 1995). Several central concepts and guiding assumptions underpin our current understanding of adolescent depression. These assumptions are based on the premise that depression arises from complex, multi-layered interactions within the adolescent (including biological and genetic factors), the adolescents' environment (including family members, peers, neighbors, school, community, and the larger social-cultural context), and the inter-relationship of these factors (Frankish, Veenstra, & Moulton, 1999; US Department of Health and Human Services, 1999). Developmental changes within the adolescent and in the environment are also important to discern as both undergo changes. Factors Associated with Depression in Adolescents During adolescence, there is a rise in mortality due to suicides, increased tobacco, alcohol, and drug use, and a dramatic increase in the incidence of eating disorders (Angold, Costello, Farmer, Burns, & Erkanli, 1999; Canadian Council on Social Development, 2001; Hankin & Abramson, 2001; Kessler et al., 2001; Kessler & Walters, 1998; Lewinsohn et al., 2000a; Newman & Bland, 1998; Parker & Roy, 2001; Patton, Coffey, & Sawyer, 2003; Rutter, 2000; Shaffer, Gould, Fisher, & Trautman, 1996; Stewart, Manion, & Davidson, 2002; WHO, 2003; Windle & Davies, 1999). Stressful life events and psychosocial influences are associated with depression (Goldberg & Huxley, 1993; Rutter, 2000; Rutter & Sroufe, 2000). In a study with more than 1,500 participants in grades 9 to 12, Lewinsohn et al. (1999) found that dysphoric mood and symptoms, dysfunctional thinking, and psychosocial stress appeared to be the predictors of first onset of depression in adolescence. Life events, especially their number and timing, may provoke the onset of depression in adolescence (Rutter, 2000) and may be mediated by the individual's response and interpretation of depression. Further, research indicates that depression in general often presents itself along with another disorder (Kreuger, 1999; Reynolds 17 Chapter II: Literature Review & Johnston, 1994; Rutter & Sroufe, 2000), most commonly with dysthymia, 6 anxiety disorder (Compas & Oppedisano, 2000; Compas et al., 1997), or disruptive or antisocial disorder (Angold et al., 1999; Kessler & Walters, 1998; Newman & Bland, 1998). Generally, comorbid diagnoses may increase the risk of recurrent depression and the duration of the depressive episode, as well as impact the response to treatment and help-seeking behaviors (Birmaher et al., 1996a; Ciarrochi, Deane, Wilson, & Rickwood, 2002; Klein, Lewinsohn, & Seeley, 1997). The state sometimes progresses from substance abuse to major depression, possibly because of either the psychopharmacological effects of substance abuse or its interference with psychosocial functioning (Rutter et al., 2006). However, the roots of factors associated with depression lie in the adolescents' understanding of what depression means to them in the context of their world. Adolescents' conceptions of depression may be linked both to overt and covert factors. A wide range of depressive symptoms is linked with adolescent depression. Some of these symptoms overlap with the DSM-IV-TR (APA, 2000) criteria for depression. An overview of the types of depressive symptoms associated with depressive symptomatology that may also be linked with adolescents' conceptions of depression follows. It involves gender, biological parameters, cognitions, emotions, social functioning, and contextual factors. Gender Rates of depression are similar for boys and girls during childhood. After the age of 13, young adolescent girls exhibit depressed affect and symptoms about twice more often than the boys of that age (Cairney, 1998; Hankin et al., 1998; Hankin & Abramson, 1999; Hankin, Abramson, & Siler, 2001; Nolen-Hoeksema & Girgus, 1994; Parker & Roy, 2001). These results have been found in cross-sectional (Angold, Costello, & Worthman, 1998; Lewinsohn, Hops, Roberts, Seeley, & Andrews, 1993; Silberg et al., 1999), prospective longitudinal (Cohen, Cohen, Kasen, & Velez, 1993; Hankin et al., 1998; Weissman, Warner, Wickramaratne, Morgan, & Olfson, 1997), and clinical studies (Cole, Martin, Peeke, Seroczynski, & Fier, 1999; Hankin et al., 1998). Some studies report that depressed adolescent males tend to exhibit more 6 The essential feature of dysthymia is a chronically depressed mood that occurs for most of the day more days than not for at least one year in adolescents and 2 years in adults (DSM-IV-TR, APA, 2000). 18 Chapter II: Literature Review externalizing symptoms, whereas depressed adolescent females show more internalizing symptoms (Frydenberg, 1997; Gjerde, 1995; Schonert-Reichl, 1994). Girls who derive their self-worth from physical appearance versus those who do not, have a higher level of depressive affect (Harter, 1999) or depressed mood at this age (Wichstrom, 1999). Dissatisfaction with body image in high school adolescents appears to be tied to gender (Allgood-Merten, Lewinsohn, & Hops, 1990; Hankin & Abramson, 2001; Hankin et al., 2001). Transition from elementary to junior high school, together with pubertal changes, has been found to account for poor self-esteem and depressed mood among girls rather than boys (Wichstrom, 1999). Evidence supports the thesis of interplay between gender role and negative environmental events (Hankin & Abramson, 1999). Girls are more likely to experience adversity within the family and within the stereotypical female gender role (Wichstrom, 1999) and become depressed (Hankin & Abramson, 1999). In a prospective study of 168 at-risk youth, Duggal, Carlson, Sroufe, and Egeland (2001) found that maternal depression was correlated with depression in adolescent females but not in males. In another study of 1,208 students in grades 9, 10, and 11, Gore, Aseltine, and Colten (1993) found that within the context of family stress, girls who were involved in their mother's problems or who had a strong interpersonal caring orientation, had elevated depressed mood (Gore et al., 1993). In contrast, Robins and Robertson (1998) found that the likelihood of people experiencing adverse life events was affected by their own psychological dysfunction. These findings indicate that there may be an interrelationship between the environment and adolescents' vulnerability to depression, particularly in girls. Biology and Environment Although parental depression increases the chance of mood and medical disorders in the offspring (Kramer et al., 1998), its specificity to depression is not established. Twin and adoption evidence shows that genetic factors account for 50 percent of the transmission of mood disorders, and additional impact comes from the child's environment (Parker & Roy, 2001). Although genetic factors play an important role in individual differences, environmental effects occupy an equally essential component (Rice, Harold & Thapar, 2002; Rutter, 2000). Goodman (2003) reports that offspring may inherit a psychosocial or environmental vulnerability to depression rather than a biological risk for the disorder. 19 Chapter II: Literature Review Cognitions, Emotions, Context, and Social Factors Research findings indicate that many different cognitive and emotional factors are associated with depressive symptomatology, yet their meaning is less defined. Often, labels of a depressed mood (e.g., sadness, blue) are assigned to depression; however, what these readings mean specifically is unclear. Other factors can include negative thinking (Garber, Weiss, & Shanley, 1993), hopelessness (Kreuger, 2002; Stanard, 2000), negative self-perceptions (Kreuger, 2002), negative subjectivity (Lykken & Tellegen, 1996), expressions of negative emotion (Plomin, 1997). Negative emotionality, such as a dysfunctional attributional style, a hostile attributional style, and maladaptive coping patterns, have been linked with affective disorders (Watson, Clark, & Carey, 1988). Irritability and sadness can also be a depressed mood (APA, 2000; Compas & Oppedisano, 2000; Farmer, 2002; Mueller & Orvaschel, 1997), yet little is known about this linkage. Bishop, Dalgleish, and Yule (2004) suggest that depressogenic cognitive biases, or a tendency toward negative statements, can occur as early as five years of age, and point to vulnerability to depression. Emotions and cognitions may be interwoven with adolescents' conceptions of depression. Numerous research findings concentrate on difficulties of social relationships and their link to adolescent depression. Loneliness is one factor that has a direct effect on adolescent depression (Brage & Meredith, 1994). Joiner, Catanzaro, and Laurent (2002) found that depressed individuals seek excessive reassurance, which leads to rejections by others. The list of social problems associated with depression includes alienation (Gjerde, Block, & Block, 1988), lack of self-belonging (Stanard, 2000), maladaptive coping patterns (Kreuger, 2002), behavioral inhibition and shyness (Cherny, Fulker, Corley, Plomin, & DeFries, 1994), and an inability to process social information (Crick & Dodge, 1994; Goodman, 2003). Loehlin (1992) cited overall psychosocial dysfunctioning (Gotlib, Lewinsohn, & Seeley, 1995; Hankin & Abramson, 2001), a temperament dimension of sociability (Goldsmith, Buss, & Lemery, 1997; Plomin, 1997), and a hostile attributional style (Kreuger, 2002). Further, the risk of depression increases in the absence of social support (Brown & Harris, 1978; Hankin & Abramson, 2001; Wong & Wiest, 1999). Moderate depressive symptoms are associated with academic and peer difficulties (Hankin & Abramson, 2001; Nolen-Hoeksema, Girgus, & Seligman, 1992; Petersen, Sarigiani, & Kennedy, 1991). These studies indicate that there is a link between relationship difficulties and depressive symptomatology, however what these relationship difficulties mean specifically is unclear. 20 Chapter II: Literature Review Perhaps adolescents' conceptions of depression may provide clearer distinctions in connection with social relationships. Rutter (2000) suggests that altered patterns of interpersonal interactions bring about further negative experiences that persist over time. Individuals' emotional and cognitive processing of their interpersonal experiences may effect depressive symptomatology. Although research points to contextual characteristics associated with adolescent depression, these factors are not found in the DSM-IV-TR (APA, 2000) criteria for depression. Studies indicate that the following factors can be linked with adolescent depression: the impact of life events and self-esteem (Ge, Conger, & Elder, 1996; Goodyer, 1990; Lewinsohn et al., 1994; Steinhausen & Winkler Metzke, 2000), rejection by parents (Rapee, 1997), and stressful events and strains that reduce social support (Mitchell & Moos,1984). In one study, Larson and Ham (1993) followed recent major events and daily emotional states in 480 children and adolescents in grades five to nine. The researchers found that negative events were significant predictors of negative affect for older students. Post et al. (2003) report that psychosocial stressors often precipitate sensitization to the affective disorders and "the cause of the malady must be sought in permanent internal changes" (Kreapelin, 1921, p. 180). Boyce and Essau (2005) suggest that adverse early experiences can have a lasting effect on the offspring's reactions to future stresses and vulnerability to depression. However, studies also indicate that when there is a strong relationship with a confidante or a social support network, life events may not lead to depression (Brown & Harris, 1978; Wong & Wiest, 1999). Contextual factors include social changes that may have implications for adolescent development and adjustment and link with depressive symptomatology in adolescence. Shifts that alter the organization of social life may present major adaptive challenges and overburden the coping capacities of some adolescents. For instance, Aneshensel and Sucoff (1996) suggest that as the neighborhood becomes more threatening, symptoms of depression increase in some individuals. Further, the role of schools and how they influence individuals, families, and communities can represent a point of convergence of several of these complexities. The diverse ways in which social change affects adolescents and how they negotiate in such times are only now being explored (Petersen, 2000; Takanishi, 2000). 21 Chapter II: Literature Review Emergence, Prevalence, and Recurrence of Depression in Adolescence A marked increase in the prevalence of depressive disorders continues during adolescence into young adult life with a female-to-male ratio about 2 to 1 (ACPH, 2000; Angold & Costello, 2001; Angold, Costello, & Worthman, 1998; Birmaher et al., 1996a; Costello et al., 2006; Feehan et al., 1993; Fombonne et al., 2001; Goldberg & Goodyer, 2005; Kessler et al., 2001; Kim-Cohen et al., 2003; Kreuger, 1999; Lewinsohn et al., 1999; Lewinsohn et al., 2000a; Parker & Roy, 2001; Petersen et al., 1993; Pine et al., 1998; Rutter, 2000; Rutter et al., 2006; Weller & Weller, 2000). Some studies indicate that increases in depression are linked with puberty (Biro, 2002; Williamson et al., 1995), particularly in prevalence rates for depression in girls who transition through puberty (Angold et al., 1998; Williamson et al., 1995). Angold et al. (1998) indicate that pubertal development measured by Tanner Stages predicted the emergence of gender difference in depression, where girls reported increased rates of depressive disorder after Tanner Stage 3. Puberty equates to approximately 13 years for girls and 14 years for boys living in the US, and there are five pubertal Tanner Stages for both boys and girls (Kreipe & Kodjo, 2002; Styne & Glaser, 2002). Garber, Robinson, and Valentiner (1997) report that there was a higher prevalence rate of depression among girls who passed through puberty early (30 percent rate) or late (34 percent), compared to girls passing through puberty on time (22 percent). Further, Ge et al. (1996) report that early-maturing girls experience more symptoms of anxiety and depression and seem to be more vulnerable to continuing these emotional problems compared with the girls maturing on-time or later. The rise in prevalence of depression during adolescence may also be the result of issues related to the impact of technology, globalization, conflicts, and changes in the environment, as they affect mental health. The dimensions of meaningful community participation for adolescents in economic, social, cultural, environmental, and political life can be disrupted. One in five adolescents in the community experience mental health problems (Birmaher et al., 1996b; Health Canada, 2002; Michalak et al., 2002; Lewinsohn, Rohde, & Seeley, 1998; Schonert-Reichl & Offer, 1992; Waddell et al., 2002). The estimated prevalence of major depression diagnosed among young people aged 9 to 17 ranges from 5 percent (Shaffer et al., 1996) to 8.3 percent (Kessler & Walters, 1998). In comparison, in population studies of depression in childhood, the rates of prevalence range between .4 percent and 2.5 percent (Birmaher et al., 1996b), while the adult prevalence rate is 5.3 percent (US Department of Health 22 Chapter II: Literature Review and Human Services, 1999). Longitudinal data from clinical (Fombonne et al., 2001) and community samples (Lewinsohn et al., 1999) show that adolescent-onset depression is strongly associated with the risk of recurrence in adulthood, with rates of 40 to 70 percent for both boys and girls (Fergusson & Woodward, 2002; Garrison et al., 1997; Parker & Roy, 2001; Pine et al., 1998). Twenty to forty percent of depressed young people relapse into depression within two years and 70 percent do so by adulthood (Garrison et al., 1997; Parker & Roy, 2001; Pine et al., 1998). In addition, 20 to 40 percent of young people with depression eventually develop bipolar depression (US Department of Health and Human Services, 1999). Data from the Dunedin birth cohort study have revealed that children, first diagnosed with Major Depressive Disorder (MDD) before the age of 15 years, had a significantly different psychosocial risk profile compared with those first diagnosed in adulthood (Jaffe et al., 2002). Findings from the community (Kessler et al., 2005; Pine et al., 1998) and high-risk samples (Weissman et al., 2005) reveal that anxiety typically first begins in childhood before the onset of adolescent MDD (Rutter et al., 2006), adolescent overanxious disorder is most likely to precede adult MDD (Pine et al., 1998), and adolescent MDD predicts adult anxiety disorders (Fergusson & Woodward, 2002). Stressful life events may serve as a developmental link between anxiety and later depression (Rutter et al., 2006). More recent findings reveal that adolescents with depression levels below the threshold of diagnostic criteria for depression may be similar to adolescents diagnosed with MDD in terms of their risk for adult depression (Fergusson et al., 2005). However, patterns of recurrence do not appear to hold for depression with prepubertal onset (Rutter et al., 2006). Childhood-onset and adolescent-onset depression may represent qualitatively distinct disorders (Weissman et al., 1999). Studies on adolescents have reported that the average length of a MDD episode is about seven to nine months (Birmaher et al., 1996a; Birmaher et al., 1996b) and that about 90 percent of these episodes recur within two years after onset (Strober, Lampert, Schmidt, & Morrel, 1993). Rather than gathering prevalence figures in adolescent mental health problems, present studies are looking into developmental epidemiology to examine timing patterns in the emergence of different mental disorders and their link to the timing of others (Costello et al., 2006). 23 Chapter II: Literature Review Tools Used in Assessing, Measuring, and Diagnosing Depression in Adolescents Assessing depression is an important process because the outcome may potentially have a very profound impact on the young person being assessed (e.g., access to services, is hospitalized, or denied access to services [Reynolds & Johnston, 1994]). To determine the presence or absence of depression in adolescents, virtually all researchers and practitioners use assessment procedures (Reynolds, 1994b). Still, there does not seem to be a gold standard-type instrument for assessing adolescent depression. Measures of psychopathology, either in the form of interviews or questionnaires, can generate a wide range of prevalence estimates, depending on the precision of the scoring criteria. However, those estimates are only as good as the classifications systems used (Costello et al., 2005). Research is needed to establish the extent to which measures and operational definitions accurately classify young individuals (Kazdin, 2001) as having or not having a change in their level of depression. Available tools still lack precise elements of time, core symptoms of adolescent depression, and good validity, reliability, internal consistency and sensitivity to change in adolescent depression. In the following section, I discuss an interview and a self-report tool that have been used extensively in assessing depressive symptomatology in adolescents. Costello, Egger, and Angold (2005) indicate that the highly structured psychiatric interviews, first established decades ago for adults, are integral for diagnosing psychopathology. One of the semi-structured diagnostic interviews presently used for children and adolescents in the age range from 6 to 18 years is the Schedule for Affective Disorders and Schizophrenia, School-Aged Children ( Kiddie-SADS; K-SADS; Puig-Antich & Chambers, 1978) (Brooks & Kutcher, 2001; Merrell, 1999). The entire interview takes two to three hours (Kaufman et al., 1997). It is used to elicit information on a wide range of emotional and behavior problems (Merrell, 1999) and designed to assess and diagnose episodes of psychopathology (Brooks & Kutcher, 2001), with a separate section for depression. The K-SADS classification criteria follow the DSM standards (APA, 1994). Trained mental health professionals and researchers, qualified to make psychiatric diagnoses, often use this measure to assess an individual's condition and provide a diagnosis. A revised version of the K-SADS is referred to as the Present and Lifetime 24 Chapter II: Literature Review version (K-SADS-PL) (Kaufman, Birmaher, Brent, Rao, & Ryan, 1996). 7 It was designed for use in assessing current or present (within the past year) episodes of psychopathology. All versions of the K-SADS begin with a parent interview, and then follow with an interview of the adolescent. Tools used to assess depressive symptomatology in adolescents enable a researcher or practitioner to classify different types of problems, to communicate via a common language with other researchers, practitioners, and affected adolescents, and to discover possible implications for treatment. The Reynolds Adolescent Depression Scale (RADS; Reynolds, 1987; RADS-2; Reynolds, 2002), the measure used to assess depressive symptomatology in this study, is a self- report tool designed to evaluate the severity of depressive symptomatology in adolescents. It has been extensively tested throughout the world, and has well established psychometric properties, including validity and reliability. Its assessment criteria coincides with depression criteria stemming from the DSM-III (APA, 1980) for the RADS (Reynolds, 1987), and the DSM-IV-TR (APA, 2000) for the RADS-2 (Reynolds, 2002). This scale, can be administered individually or in groups by researchers, clinicians, or trained lay interviewers. Rationale for Choice of Measure for this Study Although there does not seem to be an ideal robust core of instruments to assess adolescent depression, there is a need for an easily administered, valid, and reliable tool to measure depressive symptomatology in adolescents to provide accountability in clinical practice and intervention programs. Most instruments are used to assess MDD according to DSM or ICD criteria. The strength of such interviews is that they are typically coded and rule-bound, and emphasize direct face-to-face interpersonal communication. However, there are a number of weaknesses in the structured interviews (e.g., K-SADS). They may be too rigidly reliant on the DSM criteria which are often insensitive to developmental and relationship issues. Besides they are time consuming (two to three hours), are intended for use by researchers and clinicians highly trained in their use, and have not been shown to function as an index of severity with sensitivity to change. Further, most commonly used diagnostic instruments (e.g., K-SADS) show 7 The K-SADS-PL was one of the sources that served as a guide in creating categories and subcategories of adolescent depression generated by adolescents' definitions of depression in this study. 25 Chapter 11: Literature Review weakness in their ability to adequately diagnose MDD in adolescents (Brooks & Kutcher, 2001). There are also a number of weaknesses associated with the self-report instruments (e.g., RADS- 2). These weaknesses include the fact that the cut-off score, which is useful for identifying individuals for further clinical evaluation, can affect the variability; and they do not have sufficient psychometric information to allow best decisions (Myers & Winters, 2002). However, the self-report measure has a number of advantages that are important to the focus of this study. They include adolescents' perspectives and are efficient in terms of cost, time, and privacy considerations, because data are collected from one source. Besides, adolescents are reliable reporters of their depression (Weiner, Reich, Herjanic, Jung, & Amado, 1987; Yule, 1993). Hence, the self-report measure RADS-2 was utilized in this study because adolescents' perspectives were integral to the present investigation. Treatment for Depression in Adolescents Despite considerable research over the past half-century, there is no overriding agreement regarding treatment approaches for adolescent depression (AACAP, 1998). Research findings "cannot confidently provide an evidence-based explanation for how or why even our most effective interventions produce change" (Kazdin & Nock, 2003, p. 1129) and "for whom a particular treatment is well suited" (Kazdin, 2002, p. 58). Theoretical perspectives aside, the choice of initial therapy depends on the expertise of the clinician, on the individual with the presenting problem, and on the complexity of impacting psychosocial factors (Birmaher et al., 1996b). Most treatments fall under two general categories, psychotherapy and pharmacotherapy (AACAP, 1998; Birmaher et al., 1996b; Compton, Burns, Egger, & Robertson, 2002; Michael & Crowley, 2002). Psychotherapy is a learning process, done largely by verbal communication, in which mental health practitioners strive to help individuals with mental health problems and disorders (US Department of Health and Human Services, 1999). Psychotherapy has multiple specific treatment parameters depending on the theoretical approach (AACAP, 1998; Kazdin & Nock, 2003). For example, cognitive behavioral therapy, often used these days for treating depression is based on Beck's coherent cognitive model of emotional disorder (Beck, Rush, Shaw, & Emery, 1979; Fennel, 1989). Cognitive behavior therapy involves a directive, time- limited, and structured approach, based on an underlying theoretical rationale that an individual's 26 Chapter II: Literature Review affect and behavior are largely determined by the way in which he or she structures the world (Beck et al., 1979). Pharmacological interventions involve medications such as tricyclic antidepressants (TCAs), serotonin reuptake inhibitors (SSRIs), and monoamine oxidase inhibitors (MAOIs). The combination of the two therapies, psychotherapy and pharmacotherapy, known as multimodal, has been found to be the most beneficial for the treatment of acute depressive disorders in young people (AACAP, 1998). Opinions differ regarding the most beneficial order of the treatments and which is the most efficient modality (Jacobson, Dobson, & Truax, 1996). In fact, the combination of the two therapies is seldom studied (AACAP, 1998). In North America, the use of pharmacological therapies has increased over the past decade for both children and adolescents. There are few studies, however, of the safety and efficacy of most of the medication prescribed (Emslie et al., 1997; Kurt & Crowley, 2002). Controversy also exists about possible serious side effects when antidepressants are used to treat depression in adolescents (Garland, 2004a, 2004b). Treatment of adolescent depression mostly consists of psychotherapy and pharmacotherapy (Brent et al., 1997; Clarke, Rohde, Lewinsohn, Hops, & Seeley, 1999; Clarke et al., 2001; Mufson, Weissman, Moreau, & Garfinkel, 1999); however, there is emerging evidence that early intervention with at-risk offspring adolescents of adults treated for depression may prevent depression (Clarke et al., 2001). Furthermore, protective factors may even hedge depressive symptomatology. Studies indicate that influences of family and school attachments can protect adolescents against harmful behaviors associated with mental health problems (Ma & Zhang, 2002; Resnick et al., 1997). For example, the B.C. government initiated planning and providing for improved early intervention, working with community organizations and institutions (e.g., Child and Youth Mental Health Plan for BC; Ministry of Children and Family Development, 2004). In the Adolescent Health Survey of over 36,000 students in grades 7 to 12 in the US, Resnick, Harris, and Blum (1993) report that the three protective factors for "quietly disturbed behaviors" (e.g., poor body image, emotional stress, suicidal involvement, p. 4) were family connectedness, school connectedness, and low family stress. 27 Chapter 11: Literature Review Summary of Findings The evidence indicates that researchers and practitioners have trouble with definitions, assessment and treatment tensions, ambiguities, and limitations involved in the concept of depression. Depression is a growing burden in terms of indirect and direct costs, and the incidence of mental health problems in developed nations is expected to increase. Virtually all research and practice in North America are guided by the DSM-IV-TR (APA, 2000) criteria for depression to assess, intervene, and treat depression in adolescents. However, it is not clear that adolescents conceptualize depression within the same guidelines. The more important question is: Which concepts of depression are meaningful to adolescents? Adolescents' conceptions of depression can help bridge the present gap in knowledge. According to Piaget's theory of cognitive development (Inhelder & Piaget, 1958), adolescents can interpret, understand, and evaluate their own environment on a broad level and arrive at possible solutions in terms of past and potential experiences. Adolescence is a time of testing: marked by puberty, high rates of mental health concerns, and the emergence of harmful behaviors. It is the time when family dynamics, social relationships, educational achievements, and employment choices are the life events that are thrust upon the young people. For these reasons, this study seeks to expand the understanding of adolescent depression by investigating adolescents' conceptions of depression and differentiating those conceptions according to the meanings adolescents attach to them. Research findings indicate that a multitude of factors, including other disorders, are associated with depression (Armstrong & Costello, 2002; Kreuger, 1999; Reynolds & Johnston, 1994a; Rutter & Sroufe, 2000). These factors include harmful behaviors, gender, mood alterations, problems with emotional or social functioning, and lack of support. The risk that the initiation of harmful behaviors (e.g., smoking) may be linked with depression in adolescents is high (Patton et al., 1998). Moreover, studies indicate that comorbid diagnoses may increase the risk of recurrent depression and the need for treatment (Birmaher et al., 1996a; Ciarrochi et al., 2002; Klein et al., 1997). Gender is also associated with depression in adolescents; girls exhibit depressed affect and symptoms at about twice the rate of boys (Hankin et al., 1998; Hankin & Abramson, 1999, 2001; Nolen-Hoeksema & Girgus, 1994; Parker & Roy, 2001). The risk of depression increases in the absence of social support (Brown & Harris, 1978; Wong & Wiest, 1999). Irritability, sadness, loneliness, alienation, (Gjerde et al., 1988) and lack of self-belonging 28 Chapter II: Literature Review can increase the risks. The characteristics associated with depression in adolescents (e.g., depressed appearance, cognitions, emotions, social impairment or contextual influences) are matched by the literature on comorbidity in both clinical and community population samples, although the findings are limited in examining which factors are meaningful to adolescents' understanding of depression. In essence, what is the effect of loneliness on depression? Is loneliness more meaningful than another factor such as irritability? Are these factors meaningful in adolescents' conceptions of depression? What is their significance in their access to support for depression? Mental health is prominent in adolescents' concerns, yet little is known about what these concerns really mean to them. Adolescents tend to shy away from existing mental health services. Those who try them often drop out early in treatment, perhaps because of the lack of meaning for the factors they associate with depression. This present investigation examines adolescents' conceptions of depression in association with age, gender, depressive symptomatology, and pathways to talking to someone when feeling depressed, to determine which conceptions can be useful to researchers and adolescents alike. Adolescents' Self-Recognition of Depression and Pathways to Choosing to Talk to Someone When Feeling Depressed Introduction The study of factors associated with adolescent depression, including the presence of adolescents' perceived mental and physical health concerns, their understanding of what depression actually means, and their ability to recognize its presence, is a new area of research. One Australian study (Fuks Geddes, 1997) examining self-recognition of depression in a sample of 340 adolescents, showed that adolescents' self-recognition of depression was associated with more severe depression. However, questions remained about what depression meant to the adolescents. Understanding whether an individual thinks he or she has depression can be the most critical factor in detecting depression, ease in communicating with them, and in planning meaningful interventions. Most adolescents with mental health problems do not seek access to available services; and most adolescents who do enter treatment often stop the treatment within the first seven visits 29 Chapter II: Literature Review (Birmaher et al., 2007). In this section, I discuss the sources that can be linked to self-recognition of depression and pathways to talking to someone when feeling depressed. Adolescents' Self-Recognition of Depression Individuals vary enormously in their vulnerability to depression. Diagnostic criteria continue to be revised, based on the observed presence, duration, and severity of symptoms the individual presents; however, the researchers do not consider an individual's vulnerability to depression. Neither has it been established when the individual destabilizes and develops symptoms of depression (Goldberg & Goodyer, 2005; Goldberg & Huxley, 1993), nor do diagnostic criteria describe when possible subthreshold depression is imminent or what categories offer premonitions of the condition (Petersen et al., 1993). Goldberg and Huxley (1993) suggest that a common process causes people to experience symptoms or to destabilize. Once symptoms develop, another set of factors plays a part in how long these symptoms will persist (Goldberg & Huxley, 1993). Hankin and Abramson (2001) introduce vulnerabilities to depression, such as interpersonal aspects, cognitive vulnerability, or genetic risk. The degree of stability of cognitive vulnerability from childhood to adulthood is still unknown (Rutter et al., 2006). Whether the adolescents recognize their vulnerability to depression or recognize the condition itself has not been explored in research. Self-recognition of depression is a self-reported, self-diagnosis, where a person detects the presence or absence of depression and identifies what depression means to him or her, based on his or her conceptions of depression. Self-recognition involves a perceptual and conscious awareness and attentiveness to the condition. Self-recognition of depression may involve identifying cognitions, emotions, symptoms, and behaviors that the individual connects with depression, as it relates to social relationships. Webster's dictionary (Random House, 1996) defines "recognize" as "to identify from knowledge of appearance or characteristics...to perceive as existing or true; realize" (pp. 1611-1612). The understanding of self-recognition of depression in adolescents is a new territory that needs to be explored. In a theoretical paper, Cauce et al. (2002) propose that help-seeking cannot begin in earnest until a problem or mental health need is recognized. The way adolescents perceive depression, their conceptions of depression, and their self-diagnosis or assessment of the condition in themselves, may be important considerations in identifying individuals who meet 30 Chapter II: Literature Review subthreshold or clinical criteria of depression. How adolescents define depression may interconnect with their understanding and assessment of having the condition themselves. Self- recognized depression may help construct conceptions of depression that feed upon and may propagate depressive thinking. This thinking may be prompted by the adolescents' experiences and relationships with others and influence their coping strategies and access to support. Self- recognition can be a reality check, as the individual asks "What does this mean to me?" in relation to the concept of "what depression means to someone else." What adolescents think of as important can influence communication with them and service delivery. In order to detect depression in the self, the individual needs to have established conceptions of depression that apply to others and to self. The process of developing conceptions that culminate in self-recognition of depression presupposes that the individual has an understanding of others in comparison to the self This recognition of depression in the self may precede or parallel the onset of depressive symptoms and observable associated behaviors. Self- recognition of depression may have connections with social cognition and understanding of emotions, because for an adolescent, others are social mirrors to compare with himself or herself Adolescents identifying themselves as depressed may be aware of their loss of connectedness in terms of emotional and social factors. Self-recognition of depression may involve conscious attentiveness to conceptions of depression and vulnerability or sensitivity to depression. This vulnerability or sensitivity may be connected to Piaget's construct of equilibration versus disequilibration. Equilibration, a condition towards which the body strives, is the state of balance between assimilation and accommodation. In assimilation, dramatic experiences are integrated into the existing developmental schemas or incorporated within an object in the existing structure. In accommodation, each new life situation creates the opportunity for a developmental spurt, a reorganization of a cognitive and affective schema. In essence, accommodation is the organism's ability to adapt to the world by changing its internal structure or schema (Noam et al., 1995). Disequilibration is the state of imbalance between assimilation and accommodation (Noam et al., 1995). Adolescents may have the abilities to conceptualize conflicts and incongruities between their inner thoughts and actions (Selman, 1980) and their interactions. Damon and Hart (1988) suggest that self-understanding is part of a conceptual system integrating thoughts and attitudes about oneself They maintain that self-understanding is an 31 Chapter II: Literature Review evaluative insight that can extend to the past, present, and future life directions, providing the cognitive base for personal identity. Through self-understanding, one incorporates self-interests, connects and differentiates from others, including persons, places, objects, or events. Social understanding and self-understanding constantly inform each other. Self-awareness (or equivalent measures like self-directed attention or ruminative responses) has links to depression in adolescents (Hansell & Mechanic, 1985; Mechanic & Hansell, 1987). In their research, Mechanic and Hansell (1987) found that a higher score of self- awareness was more likely to be associated with adolescent depression. In a longitudinal study of adolescents followed in four Waves in grades 7, 8, 9, and post-high school (where the mean age was 20.7 years), Chen, Mechanic, and Hansell (1998) suggest that there seemed to be a mutual influence between self-awareness and depressed mood throughout adolescence, most apparent during middle adolescence with no gender difference. In another two-Wave study of 802 and 686 students in grades 9 to 12, Allgood-Merten et al. (1990) found that girls were more aware of inner states, more self-reflective, more publicly self-conscious, and had greater social anxiety than boys. Do Adolescents Think About Own Need to Talk to Someone When Feeling Depressed? Adolescents' own appraisal or thinking about own need to talk to someone when depressed can be one form of approach-oriented coping, and may be related to their conceptions of depression and self-recognition of depression. Their thinking may also be informed by their social relationships and their environment. An adolescent's decision that he or she needs to talk to someone when depressed may either facilitate or discourage the action to talk to someone at that time. Research in the understanding of approach-oriented coping specifically via thinking about own need to talk to someone when depressed, its association to private mental state processes such as self-recognition of depression, and its relationship to talking to someone when feeling depressed is unexplored. However, approach-oriented coping has been explored in association with depressive symptoms (Folkman & Lazarus, 1986; Herman-Stahl, Stemmler, Petersen, 1995; Murberg & Bru, 2005; Seiffge-Krenke & Klessinger, 2000). In a longitudinal study of the different types of coping styles on adolescents' depressive symptoms, approachers reported the fewest depressive symptoms (Seiffge-Krenke & Klessinger, 2000). Adolescents' thinking about own need to talk to someone when depressed is one form of 32 Chapter II: Literature Review approach-oriented coping that can be reflected in internal coping (e.g., "I analyze the problem and think of various possible solutions" [Shulman, Seiffge-Krenke, & Samet, 1987, p. 371], "I think about the problem and try to find different solutions" [Herman-Stahl et al., 1995, p. 657]), or reflecting on possible solution (Seiffge-Krenke, 1995). Lazarus (1993) reports that "coping thoughts and actions under stress must be measured separately from their outcomes in order to examine, independently, their adaptiveness or maladaptiveness" (p. 235). Hence, in this study, adolescents' own thinking about needing to talk to someone when depressed was distinguished as a separate factor in approach-oriented coping. Compas, Connor-Smith, Saltzman, Harding Thomsen, and Wadsworth (2001) report that "coping responses refer to intentional physical or mental actions in reaction to a stressor and directed towards the environment or an internal state. Coping goals are the objectives or intents of coping responses and reflect the motivational nature of coping" (p. 88-89). These researchers further elaborate that approach-oriented coping is part of engagement coping (e.g., problem- focused or problem solving, or seeking social support) which has been found to be associated with better adjustment (Compas et al., 2001). Coping in terms of problem-solving (e.g., "think hard about what to do") has also been used to examine children's coping with pain (Walker, Smith, Garber, & Van Slyke, 1997, p. 398). Researchers tend to examine access and barriers to mental health services (Samargia, Saewyc, & Elliot, 2006; Saunders, Resnick, Hoberman, & Blum, 1994). Currently, the accepted approach to assess who "needs" mental health care in the community has been to provide prevalence estimates, based on accepted diagnostic criteria for mental disorders. There is emerging literature on the value of self-perceived need for mental health care (Mojtabai, Olfson, & Mechanic, 2002; Rabinowitz, Gross, & Feldman, 1999; Sareen, Stein, Campbell, Hassard, & Menec, 2002; Sareen et al., 2005a; Sareen, Cox, Afifi, Clara, & Yu, 2005b; Sareen, Cox, Afifi, Yu, & Stein, 2005c) which may connect to internal approach-oriented coping. Asking the individual to identify and assess his or her own need to talk to someone when depressed as a way of approach-oriented coping, with or without using DSM criteria, may provide insight into adolescents' one type of problem-solving for own depressive symptoms. Deciding whether to seek help is the perception of need (Mojtabai et al., 2002). In a study examining self-perceived need in 1,792 participants aged 15 to 54 years, who met the criteria of diagnosis (12-month DSM-III-R mood, anxiety, or substance disorder), 33 Chapter II: Literature Review Mojtabai et al. (2002) found that participants without diagnoses were less likely to perceive a need for professional help than participants with a diagnosis, 10% versus 32%, respectively. Mojtabai et al. (2002) also found that mood disorders and the potential for suicide were strong predictors of perceived need, but that the decision to seek professional help was not. In a general Canadian population survey of participants aged 15 to 64, Sareen et al. (2002) examined the self- perceived need for mental health treatment and found that the need, independent of the DSM diagnoses assessed in their survey, was associated with suicidal thoughts. Sareen et al. (2005a) concluded that making decisions about who needs mental health care based on DSM diagnoses may not be optimal. Sareen et al. (2002) suggest that self-perceived need for formal (professional) support is a better indicator of who needs care for emotional symptoms in the community than selections based on DSM criteria. The study findings show that there is under-recognition and under-treatment of depression in adolescents both in clinical referrals (Cooper-Patrick et al., 1997; Goldberg & Huxley, 1993) and in young people who decide not to seek help for reasons that are not well understood. Even when an adolescent has sought help, there are discrepancies and differing interpretations between the help-seeker, the practitioner, and the parent. After the individual has sought help, a filter process in accessing services, controlled by practitioners at different levels of care, sets in (Goldberg & Huxley, 1993). In a study where support services were accessed, Wildman, Kinsman, and Smicker (2000) reported that children's self-report of difficulties in daily functioning related to their psychosocial problems differed from the information provided by their mothers. Moreover, physicians were able to identify less than one in five children who self-reported problems. Perhaps going further back before access to support services is considered, and examining adolescents' approach-oriented coping via thinking about own need to talk to someone when depressed, may be tied to early detection and intervention of depression in adolescents that may communicate with young people. Do Adolescents Talk to Someone When Feeling Depressed? Talking to someone when feeling depressed is a form of approach-oriented coping behavior or action undertaken by the individual to actively seek social support. Seiffge-Krenke (1995) defines this distinction as an effort to manage a problem by actively undertaking concrete actions to solve the problem. This behavioral form of approach-oriented coping by seeking social 34 Chapter II: Literature Review support can include "talked to someone who could do something concrete about the problem (Folkman and Lazarus, 1986, p. 109; Lazarus, 1993, p. 237); talking "about problems-when they appear and do not worry about them later" (Herman-Stahl et al., 1995, p. 657); or "talk to someone to find out what to do... talk to someone who will understand how you feel... talk to someone so that you'll feel better" (Walker et al., 1997, p. 398). Schonert-Reichl (2003) indicates that seeking help is a problem-focused coping strategy associated with better adjustment. Compas et al. (2001) report that correlates of health are like correlates of coping (Seiffge- Krenke, 1990) that may point to the ability to seek help. Association exists between engagement coping (e.g., seeking social support) and lowered internalizing symptomatology (Compas et al., 2001). In contrast, avoidant coping is often accompanied by the more severe characteristics of depression (Seiffge-Krenke, 1990). In a longitudinal study of 194 adolescents (composed of four annual assessments beginning with a mean age of 13.9 during the first survey) different types of adolescents coping styles with depressive symptoms were explored (Seiffge-Krenke & Klessinger, 2000). These researchers (Seiffge-Krenke & Klessinger, 2000) found that adolescents with an approach-oriented coping style reported the fewest depressive symptoms at Time 3 and Time 4, whereas adolescents with avoidant coping reported the most depressive symptoms at both Times. Two years after the study, Seiffge-Krenke and Klessinger (2000) report that all forms of avoidant coping were linked with high levels of depressive symptomatology. Other studies that examined approach-oriented coping and depressive symptomatology have had similar findings in such that approach-oriented coping was reflected with fewer depressive symptoms (Herman-Stahl et al., 1995; Murberg & Bru, 2005). It appears that the type of coping may be critical in detecting adolescents' early behaviors in seeking out social support for depression. There has been a wealth of research into help-seeking behaviors for mental health problems, however not for an approach-oriented coping action such as specifically in talking to someone when feeling depressed. In general, about half of the adolescents who identify themselves as having a mental disorder seek help (Garland & Zigler, 1994; Hodgson, Feldman, Corber, & Quinn, 1986). Mechanic (1986) reports that half of all people entering mental health care either have symptoms and complaints that do not fit the criteria for a diagnosis or seek help for problems other than those with which they present. 35 Chapter II: Literature Review Early intervention has not occurred for adolescents with mental disorders (Patel, Flisher, Hetrick, & McGorry, 2006; Patton et al., 2007; Viner & Barker, 2005). Studies show that about two out of three adolescents with mental disorders do not receive services (Burns et al., 1995; Cauce et al., 2002; Keller, Lavori, Beardslee, Wunder, & Ryan, 1991; Offer et al., 1991; Patton et al., 2007; Quine et al., 2003; Romans-Clarkson, Walton, Dons, & Mullen, 1990; Samargia et al., 2006; Sareen et al., 2005c). In a paper about the Great Smoky Mountains Study, based on a cohort of 1,420 adolescents, Costello, Copeland, Cowell, and Keeler (2007) report that one out of three young adolescents in need of mental health support received some form of intervention but that most of the support existed outside traditional mental health services. An explanation for the high drop-out rate or failure to keep the first appointment may be that adolescents rarely make the referrals themselves. Another factor may be that their adaptive social capacity and coping styles are affected by depressive symptomatology. A study differentiating help-seeking for mental health services in adolescents looked at two groups, a group with minor depression (mDEP) and a group with major depressive disorder (MDD). The study found that the mDEP group used more mental health services than the MDD group (Gonzalez-Tejera et al., 2005). The immense disparity between adolescents' modest use of support services and epidemiological findings of adolescent depression in the community confirms that we need to pay more attention to what adolescents are saying about depression and how their conceptions of depression fit with their approach-oriented coping, specifically via their own need to talk to someone and actually talking to someone when feeling depressed. Social Cognition The theoretical roots of social-cognitive development stem from the cognitive theory of Piaget (Inhelder & Piaget, 1958) and can be differentiated in their relationships to affectivity, individual differences, and personal meanings (Noam et al., 1995). Social cognition is the merging of the cognitive and social development theories. Social cognition is a complex system, consisting of related, interwoven processes enabling the individual to understand others and the personal relationships with them (Schultz & Selman, 2000b). It has only been in the past 15 years that the relation between the social-cognitive developmental perspective and depression received any consideration (Noam et al., 1995). Noam et al. (1995) report that social-cognitive theory is not only intrapsychic but interactive and there 36 Chapter II: Literature Review is a constant exchange between internalizations and actions in the world. A personal sense of self in relation to social context is vital to self-understanding. Brown and Harris (1978) recognized that people attach meaning to their experiences in ongoing interpersonal difficulties and that an individual can attribute certain sensitized vulnerabilities and meanings to earlier adverse experiences within the immediate social context. Depression and Links to Social and Cognitive Development Social and cognitive development may be deeply embedded in adolescent's conceptions of depression and their perception of self in relationship to others. Adolescent's particular understanding of how they fit in his or her environment can give rise to depression when they process and integrate depressive qualities in that space (Carpenter, 2000; Cicchetti & Toth, 1998). Joiner, Coyne, and Blalock (1999) and Joiner et al. (2002) suggest that the interpersonal context affects whether a person becomes depressed, affects the individual's subjective experience while depressed, and affects how the condition is manifested in his or her behavior as well as the resolution of the disorder. An adolescent's self-recognition of depression may be informed by just such an understanding and awareness of self in relationship to others. Hence, adolescents' approach-oriented coping, specifically thinking about own need to talk to someone and the action of talking to someone when feeling depressed, may be associated with an individual's awareness and attentiveness to factors in self in relationship to others in his or her environment. Summary of Findings Adolescents' self-recognition of depression, linked with thinking about own need to talk to someone and talking to someone when feeling depressed, may be interwoven with one's own understanding of their own conceptions and vulnerabilities to depression. The literature is in accord in suggesting that there is under-recognition and under-treatment of adolescent depression both in clinically-referred and non-referred populations (Cooper-Patrick et al., 1997). The non- referred population likely includes those young people who decide not to seek help for reasons that are not well understood. Part of the explanation may be that individuals differ in their conceptions of depression, their self-recognition of depression, and in their thinking about own need to talk to someone when depressed. 37 Chapter II: Literature Review Adolescents' approach-oriented coping such as thinking about own need to talk to someone when depressed in relation to talking to someone when feeling depressed can play a critical role in early detection of depression and may deepen our understanding of their coping styles. Adolescents' self-recognition of depression, thinking about own need to talk to someone when depressed, and choosing to talk to someone can be inextricably linked with their social- cognitive development, their experiences of social interactions, and their understanding of their own vulnerabilities that predispose them to the early onset of depression. Exploring these issues may prove critical in paving the way to greater insight into their coping capacities at an early phase of depressive symptomatology. Precursors to the Development of Depressive Symptomatology Introduction Competence in communication depends upon the capacity and ability to decipher emotion signals (Schultz, Barr, & Selman, 2001; Terwogt & Stegge, 2001) which are susceptible to emotional reactions. Presently, we know little about the mechanisms and differentiations in social and emotional development that underlie depressive symptomatology in adolescents, and in the differentiations in their social and emotional development. This may be due to discrepancies in emotional understanding, or perhaps, social and emotional development are precursors to the development of depressive symptomatology. This section reviews the literature on social and emotional mechanisms that may be important to the development of depressive symptomatology in adolescents. The Development of Social Perspective Coordination: Its Link to Adaptive Social Functioning and to Depressive Symptomatology The DSM-IV-TR (APA, 2000) uses general impairment in functioning, including deficits in social functioning, as a criterion in assessing and diagnosing depression. Studies on social adaptive patterns indicate that deficits in social functioning can predict psychopathology in adolescence and adulthood (Cowan, Pederson, Babigian, Isso, & Tost, 1973; Schultz & Selman, 2004). Mature and competent interpersonal understanding is attained by developing a coordinated social perspective that stems from the ability to differentiate and integrate the 38 Chapter II: Literature Review perspectives of self and others. Previous research, which used the Relationship Questionnaire (4 + Relationship Questionnaire), showed how social perspective coordination develops in age- related stages (see Table 1). However, social perspective coordination also evolves in interaction with its three related but separate capacities: interpersonal understanding (Schultz & Selman, 2004; Selman, 1980), interpersonal negotiation strategies (Schultz, Selman, & LaRusso, 2003), and personal meaning concerning an awareness of relationships (Levitt & Selman, 1996; Schultz et al., 2003). Interpersonal understanding involves a theoretical understanding of the nature of relationships. Interpersonal negotiation strategies include both the independent and interpersonal strategies needed to maintain good relationships. Personal meaning includes awareness of the values, attitudes, and beliefs an individual holds about particular relationships (Schultz & Selman, 2000b; Schultz et al., 2003). In essence, social perspective coordination, developed within the social-cognitive theory, assumes that there is a structure in the way individuals think about social relations, and these conceptions grow more complex with age (Schultz et al., 2003). Table 1 Developmental Levels of Psychosocial Maturity in Social Perspective Coordination, as also Operationalized in the Relationship Questionnaire Theoretical Foundation ^ 4 + Relationship Questionnaire Level^Social Perspective ^ Agea Interpersonal ^ Interpersonal ^ Personal Coordination Understanding Negotiation Meaning 0^Egocentric Recognizing that others have preferences different from the self's. No strategies in figuring out these differences. 3-5^Undifferentiated,^Impulsive^Dismissive None No differentiation of physical and psychological characteristics of others, takes no social perspective. The self's own viewpoint is the same as that of the other. 39 Chapter II: Literature Review Theoretical Foundation ^ 4 + Relationship Questionnaire Level^Social Perspective ^ Agea Interpersonal ^ Interpersonal ^ Personal Coordination Understanding Negotiation Meaning 1 Unilateral or Subjective Recognizing that self and others have differing perspectives. No judgment of own behavior from the perspective of the other. 6-7^Differentiated, 1st^Unilateral Person Understanding of physical and psychosocial characteristics, takes first-person social perspective; recognizes internal states exist yet not observable by others. Impersonal, rule-based. 2^Reciprocal^7-8 Able to take a second-person perspective in how own behavior may appear to another and perceive that others can do the same. Reflective, 2nd Person Understanding that self and others are capable of doing things they may not want to do themselves, takes second-person social perspective. Cooperative Rule-based, personal 3^Mutual Coordinating perspectives of self and others where reciprocal perspectives must be mutually coordinated to achieve a particular social outcome. 12- 14 Third-person Understanding where one can step outside the interpersonal situation and view it objectively (seeing self as both actor and object), or able to take third-person social perspective. Mutual or third-person perspective, characterized by the ability to engage in recursive perspective taking ("You know that I know that you know..."). Compromising^Need-based, isolated Note. Adapted from Keating (1990), Schultz, Selman, & LaRusso (2003), and Selman (2003). a Age when capacity norms emerge. 40 Chapter II: Literature Review The fifth, higher Level of social perspective coordination in each psychosocial competency was not included in Table 1. It is not part of the 4 + Relationship Questionnaire because the complexity of the competencies at this developmental level are difficult to capture in multiple-choice answers and this level develops in response to life tasks faced in later adolescence and adulthood (Schultz et al., 2003). The age range of the fifth level includes 15 to 18 year olds and it follows the same headings as Table 1 to determine interpersonal understanding (societal-symbolic agreement between subjects), interpersonal negotiation (showing collaborative strategies), and personal meaning (need-based or integrated awareness of relationships) (Schultz et al., 2003). In sum, persons achieve psychosocial competence as their capacity (social cognition) to coordinate and gain perspective on the relation between their own and others' points of view grows (Schultz & Selman, 2000b). These conceptions, central to the individuals' social functioning, shape human behavior (Schultz et al., 2001). Selman (2003) reports that "the degree of awareness of personal meaning serves as the linchpin between the individual's understanding of factual knowledge and his or her actual behavior -- and between adolescents' perspectives on risk and relationships" (p. 59). Studies that concentrate on difficulties in social relationships linked with adolescent depression often use general and vague terms to define the awareness of interpersonal relationships (e.g., lack of self-belonging, maladaptive coping patterns, behavioral inhibition, social information processing deficit). Perhaps, using social perspective coordination to differentiate individuals' capacity for social-cognitive development can introduce a distinction that applies more specifically to depressive symptomatology. Understanding Emotions and Links to Depressive Symptomatology Emotional disorder (Angold & Costello, 2001) is one of the terms used in the current literature to define depression in adolescents, yet little is known of emotional development and even less of its relationship to depressive symptomatology. Studies support the premise that competence in social relationships depends upon the ability to read emotions in others (Schultz, Izard, & Ackerman, 2001; Terwogt & Stegge, 2001) and to understand our own emotional reactions. The ability to accurately discriminate emotions in facial expressions (Davidson, Ekman, Saron, Senulis, and Friesen, 1990) is a critical component of successful social behavior 41 Chapter II: Literature Review (Baird et al., 1999). Depressed individuals tend to exhibit social isolation (Kandel & Davies, 1986) and the risk of depression increases in the absence of social support (Brown & Harris, 1978; Kandel & Davies, 1986; Wong & Wiest, 1999). It may be necessary to learn more about distinctions in emotions to gain insight into interpersonal competence and discern the role of emotions in the development of depressive symptomatology. The word "depression" is defined in various ways in the literature. Often, depression is used interchangeably with the terms that connote emotion (e.g., emotional disorder), mood (e.g., mood disorder), or affect (e.g., affective disorder) (Gohm & Clore, 2000). However, it is frequently a challenge to comprehend what these terms mean, separately or in relation to each other. Depression is used in other definitions as well (e.g., emotional disturbance, depressed mood, mood disorder), although it remains unclear what distinctions are being drawn. Vague definitions connecting emotional understanding to depressive symptomatology, such as the expression of negative emotion or negative emotionality, are also used in literature, yet the real meaning of those definitions is difficult to decipher. Defining Emotions Webster's dictionary (Random House, 1996) defines emotion as feeling, an affective state of consciousness in which joy, sorrow, fear, hate, or the like are experienced. The dictionary distinguishes emotion from cognitive and volitional states of consciousness. Chen and Dornbusch (1998) define emotion as an active, on-going process that is intuitive, motivational, organizing, and adaptive and occurs in response to internal and external events. Izard and Schwartz (1986) suggest that the images, symbols, thoughts, and memories associated with emotional feelings derive from cognitive processes; hence, the cognitive structure where emotions are housed can be used as a rationale for cognitive behavioral therapy in the treatment of depression. Clore and Martin in their book Theories of Mood and Cognition (2001) further differentiate emotion, as well as affect and mood. This paragraph defines their interpretations of these three terms. Affect refers to the representations of knowledge (cognitions) and the term affective refers to the representations of value (e.g., goodness) that have positive or negative personal value. Mood is an affective state without objects of feeling; mood has less salient causes than emotion and usually lasts longer than emotion (Clore & Martin, 2001). Finally, emotion is 42 Chapter II: Literature Review an affective state with objects; emotion involves simultaneous representations of appraisals in multiple systems (e.g., cognitive, physiological, experiential, expressive) that often include feelings, physiology, facial expressions, thoughts. Emotion represents appraisal or perception of something as good or bad: emotion is not just a state of feeling but a feeling about something. Emotions can act like moods, when their objects are not focalized. In other words, a prerequisite for emotion is that a situation should be perceived as positive or negative (Clore, 1994). Development of Emotions LeDoux (1994) suggests that emotions are conscious states, produced by unconscious processes. For example, emotional experiences, such as fear or anger, reflect the representation in conscious awareness of the information processes and behavioral responses characteristic of each emotion. Emotion is not a subset of cognitive theories; emotion should rather be distinguished on its own (LeDoux, 1994). Individual's emotions deliver information to other people through distinctive facial and vocal expressions. Within the individual, the information delivery happens through distinctive thoughts and feelings (Clore, 1994). The mechanisms underlying the diverse changes associated with the lack of regulation and expression of emotions are largely unknown. The relationship of these mechanisms to depressive symptomatology is just beginning to be explored. Recent research supports the notion that there are three levels of emotion: awareness of emotion, social emotion, and regulation of emotion (Penza-Clyve & Zeman, 2002; Terwogt & Stegge, 2001; Zeman, Cassano, Perry-Parrish, & Stegall, 2006). Although awareness of emotion seems to occur early in childhood, the first level, awareness of emotion, appears around the age of 11, when adolescents begin to understand that emotional reactions can be influenced by previous emotional states (Harris & Olthof, 1982). This level of understanding emotion includes an inner mental state (Harris & Olthof, 1982); however, each person may develop and function at different rates. Piaget's formal-operational stage of cognitive development, from age 11 to adulthood, is influenced by prior cognitive states. A coordinated social perspective also assumes similar stages that define the way individuals think about social relations. This age range of cognitive development and social-cognitive development reflects developmental timing related to emotion development. Hence, in adolescence as opposed to childhood, young people become aware of the broad variety of emotions available to help them function in a complex world 43 Chapter 11: Literature Review (Harris & Olthof, 1982; Terwogt & Stegge, 2001). Little is known about differentiating emotions in relation to depressive symptomatology. At the second level, social emotions become mechanisms of spontaneous self-control, requiring the fulfillment of several prerequisites. The latter include awareness of the standards of conduct, an obligation to regulate behavior with respect to these standards, and an ability to recognize discrepancies between one's own behavior and these internalized standards (Ausubel, 1955). Guilt and shame are among the so-called social emotions, because social norms and values play a critical role in their experience (Terwogt & Stegge, 2001). Guilt and shame are also characteristics experienced in adolescent depression; they are used as diagnostic criteria for depression in the DSM-IV-TR (APA, 2000). The third level of emotion, the regulation of emotion, is self-management of emotional expression. The disposition to regulate one's emotions (emotional intelligence) has been associated with lower levels of depression (Salovey, Mayer, Goldman, Turvey, & Palfai, 1995). Emotionally competent or intelligent behavior calls for regulating emotions: making emotional adjustments in the service of personal or social norms (Terwogt & Stegge, 2001). Clore et al. (2001) posit that affect tends to be experienced as relevant to whatever is currently in consciousness. The informational value of feelings depends on the focus or attention of the individual and his or her goals in the situation (Clore & Martin, 2001). The affect-as-information hypothesis states that feelings may directly affect both judgment and information processing. Whatever the origins of their problems, adolescents must consider the feelings of others when showing their own emotional reactions; otherwise, they will most likely harm the quality of their social interactions (Terwogt & Stegge, 2001). Research is just beginning to explore how adolescents manage sadness or anger. Anger can be linked to irritability, a state that is used as diagnostic criteria for depression in DSM-IV- TR (APA, 2000). Understanding and finding ways to manage sadness and anger has important implications for helping adolescents who exhibit maladaptive forms of emotional expression. Because regulation of emotion is intertwined with cognitive processes in private mental states, the interaction of these processes may affect individual conceptions of depression. 44 Chapter II: Literature Review Emotional Competence One of the most important developmental tasks adolescents face is gaining competence in emotional functioning (Saarni, 1999; Shipman, Zeman, Fitzgerald, & Swisher, 2003). Difficulties in managing emotions have been associated with poor social functioning (Shipman et al., 2003). Ciarrochi et al. (2002) suggest that adolescents who have strong social support but are low in emotional competence are still less likely to use that support. Although emotion is considered a cornerstone of human experience (Southam-Gerow & Kendall, 2002), research into the role of emotion in contributing to depression is scarce. Emotional and social competence requires the skills of emotion awareness (Penza-Clyve & Zeman, 2002; Zeman et al., 2006). Willingness to express emotions is important in social relationships; however, it is not as essential as awareness of emotion in building emotional competence (Saarni, 1999). The reluctance to express emotions is critical in disrupting social relationships and can lead to psychopathology (Lane & Schwartz, 1987; Saarni, 1999). Zeman et al. (2001) report that participants of their study, who reported using effective strategies for coping with sadness, tended to experience lower levels of depressive symptomatology. This report echoes other literature showing that children who do not attain developmentally appropriate levels of emotional competence are at risk for developing psychological disturbances (Eisenberg et al., 1995; Zeman et al., 2001). The Affect Regulation Interview (ARI; Zeman & Garber, 1996; Zeman & Shipman, 1997) and the Sadness Inhibition subscale (CSMS; Zeman et al., 2001) were significantly and positively correlated with the self-reports of sadness, indicating that the Sadness Inhibition subscale can distinguish between regulating specific emotions rather than considering it a general indicator of negative emotional control. Analyses of the Emotion Awareness Scale (EAS; Penza- Clyve & Zeman, 2002) also showed convergent validity, indicating that the Sadness Inhibition subscale is related to poor emotion awareness and emotion management. The Sadness Inhibition subscale was positively associated with higher scores on the Child Depression Inventory (CDI; Kovacs, 1985; Kovacs, 1992; Zeman et al., 2001). 45 Chapter II: Literature Review Summary of Findings Awareness and regulation of emotions and self-management of emotional expressions like sadness and irritability can affect adolescents' conceptions and self-recognition of depression. Maturity and competence in social perspective coordination can also add to the adolescents' self-recognition of depression. Poor awareness of emotion may lead to emotions being expressed in a non-constructive way, like anger (Penza-Clyve & Zeman, 2002), which may also affect adolescents' understanding of depression. Noam et al. (1995) maintain that "the ways people represent themselves...and their social world are essential in understanding the individual's psychological world. These representations... are continuously updated based on maturation and social interaction, providing an important key to a developmental understanding of health and dysfunction" (p. 430). Cumulative Summary of the Literature In summary, the review of the literature reveals that we do not have comprehensive knowledge of adolescents' understanding of depression yet. Research into adolescents' definitions of depression and their ability to recognize depression in themselves seems to be unavailable. Gaining knowledge of adolescents' conceptions of depression is relevant for theoretical and practical reasons. In theory, adolescents' conceptions and self-recognition of depression can inform their cognitive, social-cognitive, and emotional development, and in turn their development can influence their conceptions and self-recognition of depression. From a developmental perspective, adolescents' conceptions of depression would vary as a function of the developmental milestones and social goals salient for this particular age group. In practice, adolescents' conceptions of depression can provide detailed distinguishing characteristics of depression. Knowing those characteristics can have implications for the adolescents themselves and for those with whom they interact, and in detecting particular characteristics associated with depression in adolescents. Adolescents' interpretations of depression may be linked in influencing the adolescents' willingness to acknowledge the associated characteristics and to engage in support before more severe complications arise. This consideration may determine whether past research on adolescent depression is of concern in the same light to adolescents themselves. 46 Chapter II: Literature Review Research is scarce about adolescents' thinking about own need to talk to someone and propensity to talk to someone when feeling depressed before experiencing some of the known complications and sequelae of untreated depression. We know that most young people do not access available services. The field of depressive symptomatology in association with adolescents' self-recognition of depression and their pathways to talking to someone when feeling depressed is unexplored, although it is clear that self-recognition of depression can serve as early detection of depressive symptomatology. Despite the abundance of research in epidemiological findings and factors associated with adolescent depression, social and emotional developmental links to adolescent depression is a new area of research. Individuals' construction of thinking and their emotional reactions in relation to interpersonal experiences can have a possible depressive effect. Understanding the way that social and emotional development relates to depressive symptomatology in adolescents can help promote health in this population and lead to early intervention programs that communicate to young people. Research Questions and Design This research involves a cross-sectional, mixed-method design that addresses the following research questions: Research Questions 1. What are adolescents' definitions or conceptions of depression? This question is designed to investigate the ways in which adolescents understand and define depression. This is to determine whether past research on adolescent depression is considered by and of concern in the same light to adolescents themselves. A more complete understanding of how adolescents define or conceptualize depression can pave the way to advancing our understanding of depression and better communication with them that can assist us in the design and implementation of successful interventions. Adolescents' conceptions of depression will be classified into categories guided mainly by the DSM-IV-TR (APA, 2000). 2. Do adolescents' conceptions of depression vary by grade, gender, depressive symptomatology, self-recognition of depression, and approach-oriented coping (specifically via thinking about own need to talk and talking to someone) when feeling depressed? 47 Chapter II: Literature Review The aim of this question is to differentiate adolescents' conceptions of depression that may be associated with the various factors presented in this question. Distinguishing adolescents' conceptions of depression in relation to these variables can provide a map of the unique conceptions of depression that may be important in relation to these various factors. 3. What is the association of self-recognition of depression to depressive symptomatology, and approach-oriented coping (specifically via thinking about own need to talk and talking to someone) when feeling depressed? The purpose of this question is to examine whether adolescents' self-recognition or self- diagnosis of depression is associated with depressive symptomatology, and pathways to talking to someone when feeling depressed. The nature of the problem like depression, and its interpretation, can influence adolescents' willingness to acknowledge associated cognitions, emotions, and behaviors, and engage in support before more severe complication arise. Adolescents' self-recognition of depression may precede some of the known complications and sequelae of untreated depression. Adolescents' self-recognition of depression can play a critical role in their early coping capacities that include their pathways to talking to someone when feeling depressed. Coping strategies can only begin when the affected individual recognizes that he or she has a problem with depression. 4. What are the relations of social perspective coordination and emotional understanding to depressive symptomatology in adolescents? This question is designed to place dimensions of depressive symptomatology within a theoretical framework by examining adolescents' levels of social and emotional understanding in relation to severity of depressive symptomatology. The contributions of adolescents' social and emotional capacities to severity of depressive symptomatology are investigated to help provide information to assist in the design of interventions that communicate to adolescents. 48 Chapter III: Methodology CHAPTER III: METHODOLOGY Participants Participants in this study were 332 8 th and 11 th grade students.' Distribution of participants by grade, gender, and mean age with standard deviations, are presented in Table 2. Students were recruited from four public high schools in two school districts (one urban and one suburban) located in a large Western Canadian city. Two schools were drawn from each participating school district. Adolescents were recruited from classrooms representing a wide range of abilities, with the majority (75%) of participants recruited from a general required course (i.e., Social Studies). The remaining students were recruited from elective courses (i.e., Psychology, French, Art). All of the students in the participating classrooms were invited to take part, but only those adolescents who had informed written parental or guardian consent (Appendix A), and provided their own signed assent (Appendix B), participated in the present investigation. The overall participation rate for this study was 63%. Table 2 Distribution of Participants by Grade, Gender, and Mean Age, N = 332 Grade^Number of Girls^Number of Boys^Total^Mean Age (% in Grade)^(% in Grade) (N = 332) (SD) Grade 8 66 (57.9%) 48 (42.1%) 114 13.96 (.41) Grade 11 118 (54.1%) 100 (45.9%) 218 16.91 (.37) As can be seen in Table 2, boys and girls were represented relatively equally in the grade levels, and there were about twice as many grade 11 students as grade 8 students who participated. 8 The initial sample included 356 students who were recruited from 8th and 11 th grade classrooms. However, 24 students who were recruited from these classrooms were actually enrolled in either 9 th (n = 5), 10 th (n = 4), or 12 th (n = 15) grade. To eliminate potential confounds regarding the sample selection (Blaikie, 2003) of 8 th and 11 th graders, these 24 students were removed from the sample. 49 Chapter III: Methodology In order to obtain a general picture of the socioeconomic status (SES) of the participating students, SES was established by cross-referencing each participating school's postal code in the Canada's 2001 Census Postal Code Table and retrieving the Census tract name (CTNAME) or the community in which the school was located from the Canadian Census Analyzer (Refs: http://datacentre.chass.utoronto.ca). Each CTNAME was then entered separately in the 2001 Profile for Census Metropolitan Areas, Tracted Census Agglomerations and Census Tracts, 2001 Census to retrieve the SES information for every neighbourhood that each participating school was situated in (http://www12.statcan.ca/english/census01; http://data.library.ubc.ca/ (data product: Census 2001). Participants attended schools in communities that contained on average middle to higher income families. 9 According to Canadian census track data for average family income (http://datacentre.chass.utoronto.ca, 2006; http://www12.statca.ca, 2006), the range of the socioeconomic status (SES) 19 for the communities in which each school was located spanned from $76,000 to $107,000. Table 3 presents the distribution of participants by ethnicity. Table 3 Sample Ethnicity (N = 332) Ethnicity White ^ 35.3% Non-White 64.7% Note. Sample size varies as a function of missing information, Ethnicity, N = 320. 9 Average market income and median earnings by economic family types were based on the 2000 Canadian Census (http://www40.statcan.ca/101cst01/famil22b.htm Retrieved May 13, 2008; Stats Canada — Catalogue no.97-S63, [p.26]; J. Dobson, personal communication, May 9, 2008). to Average family income was based on 20% of data of all families in the 2000 Canadian Census (http://www12.statca.calenglishIcensus01/Products/Standard/profiles/RetrieveProfile.cfm...,  2006). 50 Chapter III: Methodology As can be seen in Table 3, "White"" participants comprised the minority of students taking part in this study and the remaining majority of adolescents consisted of "Non-White" 12 adolescents. Table 4, presents the distribution of participants by first language spoken at home and birth country. Table 4 Sample Characteristics, N = 332 First Language Learned at Home English 65.7% Chinese 7.5% Korean 5.7% European 5.7% Mandarin 5.4% Cantonese 5.4% Other (e.g., Japanese, Vietnamese, Hindi) 3.0% Fa rsi/Persia n 1.5% Country Born Canada 65.7% East and Southeast Asia 23.1% British Isles/Europe 4.3% Other (e.g., Columbia, Israel, South Africa) 2.7% South and West Asia 2.4% U.S. 1.8% Note. Sample sizes vary as a function of missing information, First language learned at home, N= 332; Country born, N= 329. Caucasian and White participants comprised the "White" group of adolescents. 12 "Non-White" group of participants included East and Southeast Asian (25.6%), Mixed (14.4%), British Isles and European (8.8%), Other (e.g., First Nation, Black, Human) (4.1%), West and South Asian (3.4%), and Canadian (8.4%). Although "Canadian" is not an ethnicity per se, it was included here because of its inclusion in the Canadian Census data. Specifically, in the 1996 Canada Census, the format of the ethnic origin question changed. Although the wording of the ethnic origin question remained as in previous censi, the respondents were asked to provide their ethnic origin(s) in four write-in spaces in contrast to previous check-off categories provided to respondents from 1971 to 1991. In the 1986 Canada Census, 0.5% of the total population reported Canadian as a single or multiple ethnic response, 3.8% reported Canadian in 1991, 31% reported Canadian in 1996, and 39% reported Canadian in 2001 (http://www12.statcan.ca/english/census01/Products/Referece/tech_rep/ethnic.cfm,  April 30, 2007; J. Dobson, Stat Canada Senior Analyst, Ethno-cultural stats, personal communication, May 10, 11, 2007). 51 Chapter III: Methodology As shown in Table 4, English was the predominant language first learned at home (65.7%) and the majority of participants were born in Canada (65.7%). Measures Demographic Information Participants' demographic characteristics were obtained from self-reports collected via closed and open-ended questions in the survey (Appendix C). Students were queried via forced- choice questions as to their gender ("Are you male or female") and birth date ("What is your birth date"). Their grade ("Grade"), ethnicity/race ("Ethnicity/race"), the country of birth ("What country were you born in"), and first language learned at home ("What is the first language you learned at home"), were obtained via open-ended questions. Social Perspective Coordination and Emotion Understanding Social Perspective Coordination The Relationship Questionnaire (4+ Rel-Q) (Schultz & Selman, 2000a; Schultz & Selman, 2000b; Appendix D), developed by the Group for the Study of Interpersonal Development (GSID; L. H. Schultz, personal communication, February 15, 2006), is a multiple- choice measure that assesses developmental levels of interpersonal competency and maturity of social perspective coordination. This instrument requires third to fourth grade reading skills (Schultz & Selman, 2000b) and is suitable for administration to students in fourth grade through twelfth grade (Schultz, Barr, et al., 2001; Schultz, Selman, et al., 2003). Social perspective coordination is at the core of the development of social cognition, the growth of the capacity to analyze and coordinate self and others' points of view (Schultz & Selman, 2000a). Social perspective coordination has its origins in research - based practice / practice — based research in the mid 1970s when it was first developed from one-on-one interviews to assess interpersonal understanding of children aged 6 to 16 both in the clinical and community settings. Descriptions of these interviews were grounded on the theoretical principles generated by Piaget's and Kohlberg's research (Selman, 2003). In the 1980s to the beginning of 1990's, the GSID expanded from social reasoning or what goes on inside an individual to the arena of 52 Chapter III: Methodology interactions or action choices of young people based on clinical pair therapy work in young people (Selman, 2003). In the early 1990s, the work was extended into public schools. The 4+ Rel-Q consists of 24 questions comprising five subscales that focus on relationships with both peers and adults: understanding of interpersonal relationships (6 questions); hypothetical interpersonal negotiation (4 questions); social perspective-taking (4 questions); real-life interpersonal negotiation (4 questions); and awareness of personal meaning of relationships (6 questions) (Schultz & Selman, 2000a; Schultz et al., 2003). Two 4+Rel-Q subscales represent interpersonal understanding (understanding and social perspective-taking), two subscales designate interpersonal skills (hypothetical and real-life interpersonal negotiation), and the fifth subscale represents awareness of (inter) personal meaning (Schultz et al., 2003). The 4+Rel-Q also provides an overall psychosocial maturity scale that is computed by averaging the subscales. In the 4+Rel-Q, there are four multiple-choice responses to each question that pose dilemmas or common social situations with peers or adults. Each response represents a point in the continuum of four theoretical developmental levels of psychosocial maturity in the coordination of social perspectives ranging from egocentric (Level 0) to mutual (Level 3) (Schultz et al., 2003). Level 0 designates the lowest point in the continuum (no social perspective-taking) to Level 3 (third-person social perspective-taking; see Table 1 and the Literature Review for more detailed discussion). Respondents rate each statement on a 4-point Liked scale ("poor or bad," "average or OK," "good," "excellent"). In addition to rating each item for each question, the participant is also asked to choose one of the given responses as the best answer. Each best answer has an assigned developmental value. Higher scores indicate a more mature level of social perspective coordination. Previous research has found support for the reliability of the 4+Rel-Q. Schultz and Selman (2000b) and Schultz et al. (2003) reported that Cronbach's alpha reliabilities on the 4+Rel-Q for the average item-rating score and average best-response score were .82 and .71, respectively. In the present study, Cronbach's alpha reliabilities for the average rating score and average best response score were .73 and .62, respectively. Based on previous research where best-response scores were eliminated from analyses due to low reliabilities (Pedersen, 2004), only the average rating score was utilized in the present investigation. 53 Chapter III: Methodology Self-Awareness of Emotion and Expression of Emotion To assess participants' self-reports of two aspects of emotion, self-awareness of emotion and expression of emotion to others, the Emotion Expression Scale for Children (EESC) (Penza- Clyve & Zeman, 2002; see Appendix L1-1) was used. The EESC is a 16-item questionnaire developed to assess two aspects of emotion expression, Poor Emotion Awareness (eight items), and Expressive Reluctance (eight items). The Poor Emotion Awareness subscale assesses an individual's difficulty in identifying internal emotional experiences, and the Expressive Reluctance subscale assesses individual's reluctance to express emotion to others (Sim & Zeman, 2006). Items on EESC are rated on a 5-point Likert-style scale with scores of / = "not at all like me" or "not at all true," 2 = "a little bit like me" or "a little true," 3 = "kind of like me" or "somewhat true," 4 = "a lot like me" or "very true," and 5 = "always like me" or "extremely true." Higher scores represent poorer emotion awareness and greater reluctance to express emotion. These two subscales of the EESC have been implemented in a number of studies to address its specific emotion awareness or expression (J. L. Zeman, personal communication, March, 6, 2007; Sim & Zeman, 2004; Suveg, Kendall, Comer, & Robin, 2006; Zeman, Shipman, & Suveg, 2002). For the present study, the two subscales, Poor Emotion Awareness and Expressive Reluctance were utilized separately to address the participants' specific emotion expression. Research conducted assessing the psychometric properties of the EESC provide support regarding its reliability to measure poor emotion awareness and reluctance to express negative emotion states (Penza-Clyve & Zeman, 2002). On research using a community sample of 208 fourth and fifth grade children, Cronbach's alphas for the EESC subscales for Poor Awareness and Expressive Reluctance were .83 and .81, respectively (Penza-Clyve & Zeman, 2002). In the present study, the Cronbach's alphas for subscales Poor Awareness and Expressive Reluctance were .84 and .85, respectively. Research supporting the validity of EESC indicates that it is positively correlated with sadness and anger inhibition and dysregulated expression, and negatively correlated with emotion regulation coping (Sim & Zeman, 2004; Zeman et al., 2002) in the Children's Anger Management Scale (CAMS) (Zeman, Shipman, & Penza-Clyve, 2001) and the Children's Sadness Management Scale (CSMS) (Zeman et al., 2001). 54 Chapter 111: Methodology Emotion Regulation: Sadness Adolescents' self-reports of specific behaviorally oriented emotion regulation of sadness were assessed via the Children's Sadness Management Scale (CSMS) (Zeman et al., 2001; see Appendix L1-2). The CSMS consists of three subscales: Sadness Inhibition, Sadness Dysregulated Expression, and Sadness Regulation Coping (Penza-Clyve & Zeman, 2002; Zeman et al.,, 2001; Zeman et al., 2002). Sadness Inhibition (four items) assesses sadness being turned inwards, or masking or suppressing emotional expression of sadness (e.g., "I hide my sadness"). Sadness Dysregulated Expression (three items) reflects the outward expression of sadness in culturally dysregulated manner (e.g., "I whine/fuss about what's making me sad"). Sadness Regulation Coping (five items) includes strategies for coping with sadness in a constructive manner (e.g., "I try to calmly deal with what is making me feel sad") (Penza-Clyve & Zeman, 2002; Zeman, Cassano, et al., 2006; Zeman, Shipman, & Penza-Clyve, 2001; Zeman, Shipman, & Suveg, 2002). Responses are rated on a 3-point scale (1 = "hardly ever true," 2 = "sometimes true," 3 = "often true"). Although, the CSMS was developed for children aged 9 to 12, the measure has been used successfully with adolescents 14 years and older (Sim, Zeman, & Nesin, 1999; Zeman et al., 2001). Higher scores on these subscales reflect higher levels of sadness inhibition, sadness dysregulated expression, and sadness regulation coping. These three subscales of the CSMS have been separately implemented in a number of studies to address their specific regulation of sadness (J. L. Zeman, personal communication, March, 6, 2007; Suveg & Zeman, 2004; Zeman et al., 2002). For the present study, the three subscales were utilized separately to address the participants' specific emotion regulation of sadness. Coefficient alphas for the three sadness regulation subscales have ranged from .62 to .77 and test-retest reliability spanned from .61 to .80 (Suveg & Zeman, 2004; Zeman et al., 2001). Cronbach's alphas for subscales in this study were .29 for Sadness Dysregulated Expression, .80 for Sadness Inhibition, and .53 for Sadness Regulation Coping. Subscales with Cronbach's alphas below .50 were not used in further analyses in the present study (J. L. Zeman, personal communication, March 6, 2007). Previous research indicates convergent validity with the Affect Regulation Interview (ARI) (Zeman & Garber, 1996; Zeman & Shipman, 1997) and the Sadness Inhibition subscale are significantly and positively correlated with self-reports of sadness, indicating that the Sadness Inhibition subscale can distinguish somewhat between regulation of specific emotions 55 Chapter III: Methodology rather than being a general indicator of negative emotional control. Analyses of the Emotion Awareness Scale (EAS) (Penza-Clyve & Zeman, 2002) also showed convergent validity indicating that the Sadness Inhibition subscale is related to poor emotion awareness and emotion management. The Sadness Inhibition subscale was positively associated with higher scores on the Child Depression Inventory (CDI) (Kovacs, 1985, 1992; Zeman et al., 2001). Emotion Regulation: Anger To assess adolescents' self-reports of specific behaviorally oriented emotion regulation of anger, the Children's Anger Management Scale (CAMS) (Zeman et al., 2001; see Appendix L1-3) was utilized. The CAMS is composed of three subscales: Anger Inhibition, Anger Dysregulated Expression, and Anger Regulation Coping (Penza-Clyve & Zeman, 2002; Zeman Shipman, & Penza-Clyve, 2001; Zeman, Shipman, & Suveg, 2002). Anger Inhibition (four items) assesses the suppression of anger expression (e.g., "I hold my anger in"). Anger Dysregulated Expression (three items) reflects the outward expression of anger in a culturally dysregulated manner or culturally inappropriate expression of anger (e.g., "I say mean things to others when I am mad"). Anger Regulation Coping (four items) includes strategies for coping with anger by using adaptive methods of anger management (e.g., "I try to calmly deal with what is making me feel mad") (Penza-Clyve & Zeman, 2002; Zeman, Shipman, & Penza-Clyve, 2001; Zeman, Shipman, & Suveg, 2002). The construction of the 11 items parallels the CSMS (Penza- Clyve & Zeman, 2002) where responses are rated on a 3-point scale (1 = "hardly ever true," 2 = "sometimes true," 3 = "often true"). Higher scores produced on these subscales reflect higher levels of Anger Inhibition, Anger Dysregulated Expression, and Anger Regulation Coping, respectively. These three subscales of the CAMS have been used in a number of studies to address its specific anger regulation (J. L. Zeman, personal communication, March, 6, 2007; Sim & Zeman, 2006; Suveg & Zeman, 2004; Zeman et al., 2002). For the present study, the three subscales were utilized separately to address the participants' specific emotion regulation of anger. In previous research conducted assessing reliability of the CAMS with a community sample of preadolescents, the internal consistency (Cronbach's alpha) ranged from .62 to .77 with test retest reliability ranging from .61 to .80 (Penza-Clyve & Zeman, 2002; Suveg & Zeman, 2004; Zeman et al., 2001). Cronbach's alphas for the present study were as follows: .51 56 Chapter III: Methodology for Anger Dysregulated Expression, .68 for Anger Regulation Coping, and .79 for Anger Inhibition. Depression: Symptomatology, Conceptions, Self-Recognition, Thinking About Own Need to Talk to Someone, and Talking to Someone when Feeling Depressed Depressive Symptomatology Depressive symptomatology was assessed via the 2nd Edition Reynolds Adolescent Depression Scale (RADS-2) (Reynolds, 2002), a 30-item screening instrument designed to evaluate depressive symptoms in adolescents ages 11 to 20 years. The RADS-2 does not provide a formal diagnosis of depression; rather, it facilitates the screening of self-reported depression in adolescents by producing an empirically derived cutoff score, indicating clinical severity of depression by assessing the frequency of occurrence of item-specific symptoms, and yielding a depression total scale score. The RADS-2 was developed for school-based, clinical, and research populations; takes about five to ten minutes to complete; and has a second/third grade reading level (Reynolds, 2002). Adolescents respond to the items that best describe their CURRENT (up to 2 weeks) symptom status. Responses are rated on a 4-point, Likert-type scale (1 = "almost never," 2 = "hardly ever," 3 = "sometimes," or 4 = "most of the time"). The Depression Total raw score can range from 30 to 120, in which higher scores indicate greater levels of depressive symptomatology. The cutoff score provides a metric identification of adolescents likely to manifest clinical levels of depressive symptoms. A Depression Total raw score between 30 and 75 is viewed within a normal range of symptom endorsement. A Depression Total raw score of 76 and above represents mild to severe clinical level of depression. Following the recommendations of Reynolds (2002), the sum was created totaling the 30 items on a 4-point scale. If a participant left one or more of the RADS-2 items blank, but had to respond to at least 80% of the full 30-item scale with a maximum of six items missing information on the RADS-2, Depression Total score was obtained by prorating the scores (full details can be obtained from Psychological Assessment Resources [PAR]; Reynolds, 2002). Prorating an incomplete total score for the RADS-2 was achieved by multiplying the obtained sum of the Depression score by 30 and dividing that total by the number of items completed (Reynolds, 2002). 57 Chapter III: Methodology To obtain validity support for the RADS-2 clinical cutoff scores in identifying adolescents with MDD, Reynolds (2002) conducted a study with a sample of 214 adolescents that included a sample of 107 adolescents with a primary diagnosis of MDD and an age- and gender-matched sample of adolescents drawn from a school based restandardization sample (N = 3,300). A series of 2 X 2 contingency tables were utilized using the adolescents' diagnostic status to define a clinical level of depression and the RADS-2 Depression Total cutoff scores as the selection criteria. A cutoff score of T score of 61 corresponds to 76, which is equivalent to a percentile rank of 86 based on the restandardization sample of 3,300 (Reynolds, 2002). The cutoff score was developed for use with the Depression Total scale and not with its subscales, hence only the Depression Total scale is recommended for the identification of adolescents at risk for depressive disorders (Reynolds, 2002). The data from the RADS-2 Depression Total scale reflect reliable indicators of depression in adolescents (Reynolds, 2002). The strong psychometric properties of the RADS-2 (Reynolds, 2002) have been supported in several validation studies. The RADS-2 has its roots in the Reynolds Adolescent Depression Scale (RADS) (Reynolds, 1986a; Reynolds, 1987) that was initially developed almost two decades ago. Reliability determined using coefficient Cronbach's alpha (Cronbach, 1951; Reynolds, 2002) of the RADS-2 predecessor, the RADS, (Reynolds, 1986a; Reynolds, 1986b; Reynolds, 1987) was .92 based on research on school-based population standardization sample of 2,240 adolescents (Reynolds, 1987, 2002; Reynolds & Mazza, 1998). Reliability on the RADS-2 Depression Total scores in the school-based samples based on the total school-based restandardization sample of 3,300 adolescents was .93, and for the clinical sample based on 101 adolescents, Cronbach's alpha was .94 (Reynolds, 2002). Cronbach's alpha for the RADS-2 Depression Total scores in the present investigation was .85. Adolescents' Conceptions of Adolescent Depression To assess adolescents' descriptions or definitions of adolescent depression, a measure Conceptions of Adolescent Depression (COAD) was developed for use in the present study. As part of this measure, one prompting, open-ended question, "Depression in someone my age is..." (see Appendix E for complete description of the item), was designed to enable adolescents to freely express their personal ideas regarding depression in someone their age. The wording of the question in the present study was partially adapted from previous studies on personal meanings 58 Chapter III: Methodology of death (Holcomb, Neimeyer, & Moore, 1993; Yang & Chen, 2006), on concepts of emotion (Fehr & Russell, 1984, 1991; Russell & Fehr, 1994) on concepts of pleasure (Dube & Le Bel, 2003), and informed by the cognitive developmental theory of Piaget (e.g., Chapman, 1988). To facilitate as many entries that came readily to mind, the participants were asked (in the oral and written directions) to consider how a depressed person their age would think, or feel, or act, if he or she were depressed, and to include all aspects of this person. An entire page titled, "Understanding of Depression" was allocated to this question and two additional pages in order to ensure ample space to respond. There was no set limit as to the number of definitions the participants had to stop listing responses. To keep from biasing the formulation of respondents' own COAD, and to control the amount of time spent on this item, this measure was presented as the first measure (Holcomb et al., 1993) immediately after the demographic section of the questionnaire. Self-Recognition of Adolescent Depression To evaluate adolescents' self-reports of self-assessed recognition or detection identified as depression in self, a measure Self-Recognition of Adolescent Depression (S-ROAD) was developed for use in the present study (Appendix F). As part of this measure, one forced-choice question, "Thinking of your own definition or description of depression, would you consider yourself to have ever been depressed?" was used to determine participants' self-assessed presence or absence of depression in self. Time recall in those who recognized depression in self was assessed using the question, "When did you feel this way?" This item had five Likert-style values, from "presently or in the past two weeks," to "more than 12 months ago" prior to the survey (Appendix G1). The S-ROAD has its roots in the Recognition of Adolescent Depression (ROAD) (Fuks Geddes, 1997) that was initially developed to examine recognition of depression in self in a sample of 340 adolescents from a cohort sample of 1,947 Victorian secondary students in Australia. The ROAD was administered in Wave Five (of six Waves) of the Victorian Adolescent Health Cohort Study (AHCS) to assess mental health problems, health risk behaviors, and health factors in adolescents. In a theoretical paper, Cauce et al. (2002) distinguish that help-seeking can begin only when a problem or mental health need is recognized as an epidemiologically defined need or as a subjective or perceived need. 59 Chapter III: Methodology Thinking About Own Need to Talk to Someone when Depressed To assess adolescents' thinking about own need to talk to someone when depressed, one question was developed for use in the present investigation. As part of this item, a forced-choice question, "Have you ever been so depressed that you thought you needed to: Talk to someone," was used to determine participants' own thought processes of self-focused attention in detecting or self-assessing depression in self, and identifying a need to talk to someone when depressed (Appendix G). The three values to this item comprised "yes," "no," and "I have never been depressed." This item was developed and utilized in the analyses to bridge self-assessment of self-recognized depression and adolescents' internal approach-oriented coping specifically via thinking about own assessed need to talk to someone when depressed. The thought of needing to talk to someone can be linked to other studies which have examined internal approach-oriented coping (Herman-Stahl et al., 1995; Seiffge-Krenke & Klessinger, 2000; Shulman et al., 1987; Walker et al., 1997). One such example of internal coping is, "I think about the problem and try to find different solutions" (Seiffge-Krenke, 1995; Seiffge-Krenke & Klessinger, 2000, p. 621). Talking to Someone When Feeling Depressed To evaluate one type of approach-oriented coping in seeking social support involving directed voluntary action initiated by the affected individual, such as talking to someone when feeling depressed, one question was developed for use in the present study (Appendix H). As part of this item, a forced-choice question, "At that time of feeling depressed, did you talk to anyone about how you were feeling" was used to determine participants' self-reported presence or absence in talking to someone. Assessment of adolescents' responses was based on two values ("yes" or "no"). The wording of the question in the present study was partially adapted from a previous item in the ROAD (Fuks Geddes, 1997). Seeking social support specifically via talking to someone when feeling depressed can be linked to other studies that have examined behavioral approach-oriented coping (Herman-Stahl et al., 1995; Seiffge-Krenke & Klessinger, 2000; Walker et al., 1997). One such example of active coping is "I talk right away about the problem when it appears and don't worry much" (Seiffge-Krenke & Klessinger, 2000, p. 623). 60 Chapter III: Methodology Content Analysis of Participants ' COAD Content analysis of participants' responses to the COAD item was the method used in this study to construct categories and subcategories of adolescent depression. The following sections describe the processes used sequentially in developing these constructs. Adolescent depression, as defined by the participants, served as the context for their given responses. Coding Schemes The sequential process in developing the constructs of adolescent depression was first undertaken in the Access (Microsoft Version 2003) program (Appendix I). Three columns were allocated for this initial process of data entry, one column for the 332 unique ID numbers, another column for the numbered raw responses provided in the COAD item, and the third column for the broken-down raw responses (from second column) into numbered elements of information called units (see next paragraph for explanation). Initially, data entry of every student's unique identity ID number (first column) and the numbered (unique) raw responses 13 (second column) matching each unique ID number, was completed. Next, in the third column, each recorded response (second column) was separated into units and each unit was numbered and matched to the unique ID number. For example, if ID number 1 had a list of six bulleted responses or a list of responses separated by lines, to the COAD item in the questionnaire, six rows were created in the ACCESS program to contain the COAD information provided by ID number 1. The first column contained the number 1 in each of the six rows. The second column comprised the six rows of raw responses (adjacent to the number 1 in the first column) that were automatically sequentially numbered from 1 to 6 by ACCESS. The third column contained the six rows of units and each unit on each of the six rows represented the distinct COAD information of the raw response. Units were individually numbered and received a different number to each distinct unit of information or writing style. Units of information distinguished themselves as the smallest specific description of communication content (Holsti, 1969; Krippendorff, 2004; Neuendorf, 2002; Tashakkori & Teddlie, 1998) on which participants' data on adolescent depression were analyzed. Units of 13 Each unique numbered raw response in ACCESS corresponded to another response, either a different bullet or on a new line, provided in each student's list of responses to the COAD item. 61 Chapter 111: Methodology information have been coded on which participants' data were analyzed in previous studies on personal meanings of death (Yang & Chen, 2006), concepts of pleasure (Dube & Le Bel, 2003), and in a sample of five preadolescents' and adolescents' descriptions of being ill (Forsner, Jansson, & Soerlie, 2005). A response unit represents one bulleted response or a response separated by a new line by each participant in his or her generated list of definitions of adolescent depression. A COAD unit distinguishes itself by the separation of each response unit into elements or the smallest information of adolescent depression via content analysis. Analyses in this study are based on the COAD unit information. Single words were the smallest units used in content analysis (Holsti, 1969) and were more easily grouped into a category than participants' responses containing phrases. For example, a response "angry" was transformed into an "angry" unit and would not require any separation of raw text. However, a phrase response containing "sad, negative thinking" was divided into two units of meaningful information, "sad" and "negative thinking." Participants' matched responses and their units to their ID's in Access (Version 2003), allowed me to store the text data and maintain continuity in examining the computerized records of the original responses linked to the ID's and the created message units of descriptions of adolescent depression. Classification of Participants' Units of Adolescent Depression To classify every assigned unit within its specific theme or category of adolescent depression, each assigned unit, developed in Access (2003), was then transcribed into Microsoft Word (2003) and clustered within the emerging categories, based on the unit information, in a developing coding manual (Appendix J). In the preliminary phase of this study, participants' units of descriptions of adolescent depression were also printed and cut up so that each piece of cut-up paper contained one unit. Each message unit was then clustered into its relevant categories pasted on long sheets of paper and suspended on the walls of a room to acquire a visual and immediate perspective of all the assigned categories (Appendix K). To better reflect and bridge the meaning of unit definitions of participants' COAD to established criteria for depression, I adapted depression symptomatology criteria from several sources to develop the construct categories to contain the units defining them. To my knowledge, as there were no studies to draw on adolescents' conceptions or understanding of depression 62 Chapter III: Methodology using content analysis, depressive symptomatology criteria predominantly stemmed from the DSM-IV-TR (APA, 2000). 14 Table 5 presents DSM-IV-TR (APA, 2000) criteria for MDE. The DSM-IV-TR (APA, 2000) criteria were used because of the extensive and sustained research in gathering data, in classifying and developing criteria, and recording information about mental illness (in the United States) since its inception in 1840 (APA, 2000). The DSM-IV-TR (APA, 2000) is also used extensively by different health professionals. Other sources that served as guides to developing constructs in this study were the Depression Supplement in the K-SADS-PL (Kaufman et al., 1996), the Clinician's Thesaurus (Zuckerman, 2005), and anger definitions (Russell & Fehr, 1994). For example, subcategory "Irritability" and category "Depressed Mood" are used as constructs or categories, or subcategories for participants' COAD in this study. Although "Irritability" is defined as a symptom of "Depressed Mood" criteria for depressive symptomatology in the DSM-IV-TR (APA, 2000), there is limited information in this source defining "Irritability." Russell and Fehr (1994) provide further definitions of anger that corresponded to the students' COAD units of adolescent depression such as frustration or hate that can be designated to the subcategory "Irritability" in this study. 14 The criteria of depressive symptomatology in the DSM-IV-TR (APA, 2000) stemmed from Major Depressive Episode. 63 Chapter III: Methodology Table 5 DSM-IV-TR Criteria for Major Depressive Episode (MDE) Criteria for Major Depressive Episode A.^Five (or more) of the following symptoms have been present during the same 2-week period and represent a change from previous functioning; at least one of the Symptoms is either (1) depressed mood or (2) loss of interest or pleasure. Note: Do not include symptoms that are clearly due to a general medical condition, or mood- incongruent delusions or hallucinations. (1) Depressed mood most of the day, nearly every day, as indicated by either subjective report (e.g., feels sad or empty) or observation made by others (e.g., appears tearful). Note: In children and adolescents, can be irritable mood. (2) Markedly diminished interest or pleasure in all, or almost all, activities most of the day, nearly every day (as indicated by either subjective account or observation made by others). (3) Significant weight loss when not dieting or weight gain (e.g., a change of more than 5% of body weight in a month), or decrease or increase in appetite nearly every day. Note: In children, consider failure to make expected weight gains. (4) Insomnia or hypersomnia nearly every day. (5) Psychomotor agitation or retardation nearly every day (observable by others, not merely subjective feelings of restlessness or being slowed down). (6) Fatigue or loss of energy nearly every day. (7) Feelings of worthlessness or excessive inappropriate guilt (which may be delusional) nearly every day (not merely self-reproach or guilt about being sick). (8) Diminished ability to think or concentrate, or indecisiveness, nearly every day (either by subjective account or as observed by others). (9) Recurrent thought of death (not just fear of dying), recurrent suicidal ideation without a specific plan, or a suicide attempt or a specific plan for committing suicide. B.^The symptoms do not meet criteria for a Mixed Episode. C. The symptoms cause clinically significant distress or impairment in social, occupational, or other important areas of functioning. D. The symptoms are not due to direct physiological effects of substance (e.g., a drug abuse, a medication) or a general medical condition (e.g., hypothyroidism). E.^The symptoms are not better accounted for by Bereavement, i.e., after the loss of a loved one, the symptoms persist for longer than 2 months or are characterized by marked functional impairment, morbid preoccupation with worthlessness, suicidal ideation, psychotic symptoms, or psychomotor retardation. Note. From Diagnostic and Statistical Manual of Mental Disorders (p. 356), by American Psychiatric Association, 2000, Washington DC: American Psychiatric Association. Copyright © 2000 by the American Psychiatric Association. Reprinted with permission. 64 Chapter III: Methodology Some categories based on the DSM-IV-TR (APA, 2000) criteria for depression were combined or broken down depending on the frequency count and the commonalities of these categories. For example, the two DSM-IV-TR (APA, 2000) criteria for a MDE, "Weight and Eating Changes," and "Sleep Changes," were combined into the construct or category in this study, "Appetite/Weight/Sleep Changes." The DSM-IV-TR (APA, 2000) criterion, "Distress, Social, Occupational, and Functional Impairment" (see Criterion C in Table 5), was segmented into "Social Impairment" and Functional Impairment/Distress" categories in this study. When clusters of similar or repetitive units of information filled ample numbers within a specific category such as "Depressed Mood," a subcategory defining that particular group of units of information was created within that category "Depressed Mood." For example, subcategory "Sadness" was created when the frequency of similar and like words or units presented themselves in sufficient frequencies in the "Depressed Mood" category to create the "Sadness" subcategory (Appendix J). The development of subcategories was mainly guided by the DSM-IV-TR (APA, 2000) criteria for depression and, more infrequently, the K-SADS-PL (Kaufman et al., 1996), the Clinician's Thesaurus (Zuckerman, 2005), and subcategory "Irritability" (Russell & Fehr, 1994). If the content of similar units of information did not fit any selected source for criteria for depression, new categories and subcategories were developed to contain the distinct meaning of the message units (e.g., category "Contextual/Causal" or subcategory "Inner Pain") defined by the participants in their descriptions of depression. Allowance for keeping the emergent extra categories, which simply were not found in the designated resources, was justifiable to avoid a potential for circular reasoning by trying to fit all categories particularly into DSM-IV-TR (APA, 2000) or K-SADS-PL (Kaufman et al., 1996) depressive symptomatology criteria. The coding manual of construct categories and subcategories of COAD, and their respective units of information, emerged as an iterative process refined as content analysis progressed to develop the constructs of adolescent depression. Delineated themes and subthemes of similar or repetitive units were systematically sorted within the emerging categories and subcategories of adolescent depression (Appendix J, Appendix L) until relatively discrete and meaningful categories and subcategories of COAD were identified. Every raw text COAD unit was assigned to the developed categories and subcategories in the codebook. For example, the word unit "confused" was assigned to the DSM-IV-TR (APA, 65 Chapter III: Methodology 2000) category "Indecisiveness/Diminished Ability to Think or Concentrate." Text unit "disconnect from friends" was allocated to the category "Social Impairment" based on the segmented DSM-IV-TR (APA, 2000) criterion "Distress/Social/Occupational/Functional Impairment." In the process of clustering the units of information into the categories, some units maintained their raw text as part of the category (e.g., "Suicidal") or subcategory label (e.g., "Sadness"). However, most units were clustered with similar or repetitive units of meanings and allocated under a distinct category. For example, the units "their marks might drop" and "thinking there is no use in school" were grouped in the category "Functional Impairment/Distress." Microsoft Word (2003) was used to maintain and track the columns of numbered categories, subcategories, and the COAD units of depression categorized and clustered next to the emerging categories in the adjoining columns (Appendix J, Appendix L) to produce the final codebook. The general rule was not to aggregate the categories and the subcategories codes prematurely (0.L. Holsti, personal communication, September 16, 2005). Each numbered unit of adolescent depression information in the codebook in Word (2003) matched the numbered unit in Access (2003). After completion of categories and subcategories, and confirming that all the numbered units matched in Access (2003) and Word (2003), only the numbers were then entered into Statistical Package for the Social Sciences for Windows (SPSS) program (Version 13, 2004) against the ID to reflect each participant's units of information. Access (2003) was used to transcribe the numbers into SPSS (Version 13, 2004) because the numbered units were linked to each ID in this study. If different participants provided the same unit of information (e.g., "sad") to define adolescent depression, the same number "1" was assigned to this particular message unit for any individual who provided the unit "sad" in his or her responses. There are three different versions of the codebook. Codebook A contains the categories and subcategories with every unit provided by the participants to deliver the sum of all the units. Codebook B includes only the numbered units contained within each category and subcategory to produce the sum of units for each category or subcategory. Lastly, Codebook C contains the 18 categories and 17 subcategories containing only the different units assigned to each numbered unit of adolescent depression. These different units also matched with the numbered units in Access (2003) and Word (2003), and the numbers in SPSS (Version 13, 2004) that represented 66 Chapter III: Methodology these units corresponding to their ID's (also the same as found in Access [2003]). For example if a unit "sad" was provided by ID 1 and ID 7, the "sad" unit was assigned number "1" and each unit "sad" from each ID was assigned number 1 in the particular Codebook under the assigned Category or Cluster that it belonged to in "Depressed Mood." This way, when "sad" was transcribed into SPSS, "sad" was always appointed number "1." There were 1,473 numbers representing the different units of information that were entered into SPSS (Version 13, 2004) allocated to their respective ID numbers. In developing the categories and subcategories of COAD, the level of severity or the frequency of the defined categories or subcategories were not identified but just the presence of the category or subcategory. For example, "sadness" and "deep sadness" were each allocated into the subcategory of "Sadness" in the "Depressed Mood" category. Although "depressed" is used in DSM to define depressed mood in MDE as described by the person (DSM-IV-TR; APA, 2000; K-SADS-PL; Kaufman et al., 1996 in Depressive Disorders), the few responses with "depressed" as a definition of adolescent depression, were placed in the "Miscellaneous" category because "depressed" was the same word used to define "depression." No unit of description of adolescent depression was coded as belonging to more than one category. However, units that belonged to different subcategories in a category were included once in each separate subcategory if mentioned by the same individual. Hence, the category did not contain the sum total of its subcategories. The aim was to find out if an individual provided a definition unit belonging to a subcategory or a category (if it did not contain any subcategories) rather than how many times each individual provided a unit definition for each construct. There was an occasion where a cluster of units of descriptions of adolescent depression such as the subcategory "Quiet/Silent" was assigned to the category "Agitation/Retardation" but could have also been included as a subcategory in "Social Impairment" (see Discussion Chapter). Since categorizing the units was guided by the DSM-IV-TR (APA, 2000) criteria for depression, "Quiet/Silent" has been placed in the "Agitation/Retardation" category because "Retardation" in the DSM-IV-TR (APA, 2000) includes "speech that is decreased in volume, inflection, amount, or variety of content, or muteness" (APA, 2000, p. 350). The "Miscellaneous" category that emerged from adolescents' responses made references to other qualities of depression not related to the participants' COAD units. Some of these units of information that defined other dimensions of depression included "hard to deal with without help," "potentially preventable," 67 Chapter III: Methodology "something that needs to be dealt with and resolved," "more rare in Canada," "depressed," "go to do sports," "reborn," or "unfortunate." Intercoder Reliability In order to establish validation of the coding scheme (Neuendorf, 2002) for participants' constructs of COAD, a graduate student in Educational and Counselling Psychology was invited to judge and score a random sample of 326 units of participants' COAD units of descriptions of adolescent depression following a brief training phase by this investigator just prior to coding. Cohen's kappa (Cohen, 1960, 1968), utilized to evaluate intercoder reliability in studies conducting content analysis (Greener & Crick, 1999; Nauta, Hospers, Jansen, & Kok, 2000; Neuendorf, 2002), was used to evaluate intercoder reliability. Training involved instruction about the meaning of all the categories labeled alphabetically, using Access (2003), and coding instructions in how to assign his or her selected category to each randomly selected unit of COAD onto printed Access (2003) sheets of randomly selected units of adolescent depression definitions (Appendix M). The intercoder had a list of all the categories labeled alphabetically printed on a couple of sheets and a set of Access (2003) printed sheets containing the units that required the alphabet letter of the category to match the unit (Appendix N). The intercoder entered his or her choice of the different alphabet letter that was matched to each of the 18 categories against each visible, next to the randomly selected unit. In setting up the selected sample of units in Access (2003) for the intercoder, the sample was first randomly selected in SPSS (Version 13, 2004). The selected sample was then saved in the Excel (2003) program and finally transferred into Access (Version 2003). The intercoder had two columns in view, the listed raw responses in one column, and an empty column to enter his or her choice of code or a category alphabet letter to correspond next to the raw response (Appendix M). The random sample of 326 units of participants' definitions of adolescent depression represented 15% of the total sample (or 15.7% of total units in this study). A sample size between 50 and 300 is the recommended one for intercoder reliability (0.R. Holsti, personal communication, March 16, 2006; Neuendorf, 2002; Krippendorff, 2004; Yang & Chen, 2006). In a previous study on strategies for making friends, Wentzel and Erdley (1993) used 10% of their total of 2,300 strategies for a sample for interrater reliability. 68 Chapter III: Methodology A total of 1.5 hours were taken to train the intercoder. The total time used by the intercoder to code all the presented units against his or her choice of categories was 2 hours and 55 minutes. Intercoder reliability using Cohen's kappa was 0.83. After the initial intercoder reliability was assessed, a time was scheduled for this investigator and the same intercoder to discuss the non-matched items. After discussion, the intercoder reliability for Cohen's kappa was 0.94. Familiarity with the nature of the material and capability to handle the categories in terms of the data language reliably was helpful (Holsti, 1969; Krippendorff, 1980) but not a necessary requirement for the intercoder. Constructs of participants' categories and subcategories, category definitions, and examples are presented in Table 6, Table 7 and Table 8. 69 Chapter III: Methodology Table 6 Categories, Subcategories, and Examples of Participants' Content Analyzed Conceptions of Adolescent Depression (COAD) for Depressed Mood, N = 332 #^Category ^ Category Definition ^ #^Subcategory ^ Examples 1 Depressed Mood Focus of descriptors is^la distinguished mainly by emotions of sadness or irritability, and includes moodiness. Descriptions also include inner pain and depressed appearance. Depressed cognitions are predominantly characterized by hopelessness, no motivation, boredom, and pessimism. ld le Sadness Irritability Moody Somatic Symptoms (bodily aches or pains) Inner Pain Sad, sadness, down, unhappy, low, gloomy. Anger, angry, mad, angry outbursts, frustrated, annoyed disappointment, dissatisfaction, upset. Moody, emotional. Difficulty breathing. Inner pain, hurt, hurting inside, broken heart. lb lc Teary, crying, won't smile or laugh, sad face, stop caring about their appearance, downcast, goth. Hopeless, empty, lack of faith, empty. Unmotivated, doesn't care about what she/he does, an unwillingness to do anything, not pursue goals, apathetic. boredom, bored, indifference if ^ Depressed Appearance (observations of the person's facial expression and demeanor) lg ^ Hopelessness lh ^ Unmotivated (Lacking the interest or ability to initiate action). ii^Boredom/Indifference lj ^ Pessimism ^ Pessimistic, thinks negatively 70 Chapter III: Methodology Table 7 Categories, Subcategories, and Examples of Participants' Content Analyzed Conceptions of Adolescent Depression (COAD) for Social Impairment, N = 332 #^Category ^ Category Definition^#^Subcategory ^ Examples 9 Social Impairment Focus of descriptors is distinguished mainly by perceived disconnection, or behavioral disconnection, or feeling lonely, or aggression or being bullied. 9a^Perceived Disconnection 9b^Behavioral Disconnection Feelings of isolation, unwanted. Unsociable, stay alone, left out. 9c^Lonely ^ Lonely, loneliness, feel lonely, lonesome, aloneness. 9d^Aggression/Bullied She puts down other people, steal/do bad things, bullied, aggressive Table 8 Remaining Categories, Subcategories, and Examples of Participants' Content Analyzed Conceptions of Adolescent Depression (CORD), N = 332 #^Category ^ Category Definition ^ # Subcategory ^ Examples 2^Loss of Interest or Pleasure Focus of descriptors is distinguished mainly by diminished or loss ofinterest or pleasure from previous state of interest or pleasure in function. Lose interest in school, lose interest in hobbies, lose interest in activities, the same things they used to enjoy. 3^Appetite/Weight/Sleep Focus of descriptors is Changes^distinguished mainly by weight loss when not dieting or weight gain, or decrease or increase in appetite, craving of specific food, insomnia or hypersomnia. Sudden increase or decrease in weight, lose appetite, sleepy, not enough sleep, eat a lot, unable to sleep, hungry, overeating, weight, get very skinny, are fat. 71 4^Fatigue/Loss of Energy Focus of descriptors is distinguished mainly by decreased energy, tiredness, fatigue without physical exertion. 5^Low Self-worth/Guilt^Focus of descriptors is distinguished mainly by feelings of worthlessness or excessive or inappropriate guilt, or insecurity. Chapter III: Methodology #^Category ^ Category Definition ^ # Subcategory ^ Examples 6^Indecisiveness/ Diminished Ability to Think or Concentrate 7^Feel Trapped 8^Suicidal Focus of descriptors is distinguished mainly by diminished ability to think or concentrate, or slowed thinking, inattention, indecisiveness, or easily distracted, or memory difficulties. Focus of descriptors is distinguished mainly by an inability to see any way out of undesirable internal and/or external circumstances, or feel imprisoned, or disempowered, or impotent, or constrained as in not able to realize some desired state of being. Focus of descriptors is distinguished mainly by thoughts of death, recurrent suicidal ideation without a specific plan, or a suicide attempt, or a specific plan for committing suicide. Self-harm behaviors can also include hurting themselves, or cutting self, or self- mutilation. Tired, lack of energy, physically exhausted, exhausted, less energy, fatigued, low energy level. Lack of self-esteem, low self-esteem, think you are ugly, self- hatred, feel worthless, lack of confidence, guilt, ashamed, feel insignificant, useless, self-criticism, regret, insecure. Seem lost, aren't with it, confused, distracted, short attention span, don't pay a lot of attention, forgetful, doubtful, unable to organize themselves. When they are under the pressure to do something they don't want to do,  being trapped in your life. Might tell you they are thinking about suicide, possible thoughts of suicide, may resort to self mutilation, thinks of death, attempts suicide. 72 Chapter III: Methodology #^Category ^ Category Definition ^ # Subcategory ^ Examples Their marks might drop, thinking there is no use in school, distressed, never going out and doing things. Focus of descriptors is distinguished mainly by a decline or not taking part in activities, distress, interference in other important areas of functioning including less interest in hobbies, or academic problems like truancy or school failure. Focus of descriptors is distinguished mainly by retardation that includes quiet or silent, speech that is decreased in volume, inflection, amount, or variety of content, or muteness, or slowed speech, thinking, and body movements. Agitation includes an inability to sit still restlessness. 10 Functional Impairment/ Distress 11. Retardation/ Agitation 11a Quiet/Silent Quiet, silent, don't feel like talking. Restlessness, agitated constantly, slow. 12 Anxiety Symptoms Focus of descriptors is distinguished mainly by feeling anxious, or stressed, or worry, or fear. 12a Fear 12b Anxious/ Stressed Afraid, fear, feel scared. Stressed, always worry, anxiety, stressed out. 13 Substance Use 14 Contextual/ Causal Focus of descriptors is distinguished mainly by alcohol or other substance use or abuse. Focus of descriptors is distinguished mainly by environmental or contextual, or causal references. These can include a negative life event, difficulties or deficiencies in their environment, familial or other interpersonal stress, or other problems relating to the context in which a person's difficulties have developed. Might involve drinking alcohol, take drugs, alcohol, cocaine. Stress (family), stress, busy schedule, pressure, stressful life, bogged down, someone who may have experienced great trauma, loss, unsafe family, rough life, problems, family problems, money problems, too much school work. 73 Chapter III: Methodology #^Category ^ Category Definition ^ # Subcategory ^ Examples 17 Escaping Reality 16 Masking/Stigma 15^Biological Focus of descriptors is distinguished mainly by biological influences like genes, infections, hormones, toxins, or chemical factors. Focus of descriptors is distinguished mainly by concealing, veiling intrapersonal emotions or cognitions with a facade or contradictory external appearance from own recognized state of being. Awareness and attention of self's intrapersonal, physical appearance, and interpersonal understanding is central in own perceptions of self's state in relation to one's environment. Place importance of self's understanding and own actions in relation to his or her social world, and often act according to the confines of their perceptions of social expectations rather than one's own. Focus of descriptors is distinguished mainly by avoiding, refusing to face their internal or external, or social/relational circumstances. This escaping reality may produce a resultant fantasy state. Chemical imbalance, hormones out of control, making you unable to fully control yourself. Who is too scared to tell anyone about his/her depression, false outgoingness, something someone would hide from friends. They cut themselves to numb out the pain inside themselves, doing drugs/alcohol to escape reality. 74 Chapter III: Methodology Category ^ Category Definition ^ # Subcategory ^ Examples 18 Miscellaneous Focus of descriptors is distinguished mainly by responses that did not answer the question, not captured by previous categories, or irrelevant, too general, or did not make sense. Harmful, they might develop mental illnesses, needs help, intimidating, rough, common, needs help it really depends on what kind of depression, not understanding the reason, untreated has a problem / problems with himself a state of mind, sickness. Procedures This section describes the process in accessing classrooms, and administering the questionnaire booklets containing the measures and demographic questions utilized in this study, and follow-up protocol for referrals of participants screened as depressed. A one-time, two hour pilot study was conducted to refine the questionnaire and the process for the main study (Appendix 0, Appendix N1). Ethics approval to conduct this project was obtained from the University of British Columbia (UBC; see Appendix M1) and two District School Boards. Principals within the participating school districts were informed of the study's approval by their respective School Boards. I then contacted the principals to provide them with information packages of this study (Appendix P), in order to obtain permission to recruit participants from their schools. Principals then distributed information about this study to school personnel. Following this, meetings with the school counsellors and teachers interested in participating in this study were held in order to schedule classroom presentations and data administration. Data were gathered in classrooms in the late spring during regular class sessions. Teachers remained in their classrooms while students completed the questionnaires. This study was presented to participants as an investigation to learn about adolescents' understanding of depression in someone their age, and to provide their voices as to what they thought depression was and what they did about it. One to two weeks prior to administering the 75 Chapter III: Methodology questionnaires, 15-minute presentations were given to each participating class informing the students about the purpose of the study. Information letters (Appendix Q - Parent or guardian information letter, Appendix R - Student information letter) and Parental consent form (Appendix A-Parental) were distributed to students to take home to their parents. Parents and students were both informed that participation in the study was voluntary, and confidentiality of student responses was emphasized. A $70 gift certificate was presented to each participating classroom teacher to decide with all the students in their respective classes in how to spend the fund. 15 Two research assistants trained by this author, one male and one female, accompanied me for each 15-minute presentation and throughout all the participating classroom data administration to distribute and to collect questionnaires, monitor student contact, and respond to students' questions and concerns. Questionnaires were administered to students in one 60-minute class period during school hours. Before distributing the questionnaires, approximately five minutes were allocated to go over the instructions in how to respond to the COAD item, confidentiality issues, and the post- questionnaire follow-up protocol for referrals (see section Follow-Up Protocol Post Data Administration in this Chapter), and my 20 minute availability to discuss their personal questions in confidence in a defined location in school outside the classroom after data administration. Assent forms were signed by participants just prior to data administration (Appendix B - Student). Questionnaires were identified by number codes rather than student names. Students completed a numbered identification page (Appendix S) attached to the front of the questionnaire which was collected as soon as it was filled out and kept in a separate envelope from the completed questionnaires. This identification page was used to match students' names to their unique ID number in the questionnaires when they were screened as depressed. Throughout data administration, I or the two trained research assistants provided assistance, as needed, to any participant who had questions pertaining to the questionnaires. Each questionnaire was distributed in its own separate envelope which each participant used to insert her or his completed survey. Figure 1 illustrates the sequence of the questionnaire's content administered 15 Some classrooms decided to donate the fund and match it, to a charity, other students and their teacher bought books for the classroom, while another classroom had a pizza lunch. 76 Chapter III: Methodology to participants in the present study. The demographic information was presented at the start of the questionnaire. As previously implemented in a study on personal meanings of death (Holcomb et al., 1993), to keep from biasing the formulation of participants' own personal meanings of adolescent depression, to ensure the potential of an adequate number of responses, and to control the amount of time spent on the survey, the COAD item immediately followed the demographics component. The other instruments followed concluding with the RADS-2. 77 Chapter III: Methodology Figure 1 School Data Administration Demographics (^ Conceptions of Adolescent Depression (COAD) Self-Recognition of Adolescent Depression (S-ROAD) ( Thought Needed to Talk to Someone 7 ^ L ^Talking to Someone When Feeling Depressed  Emotional Awareness and Emotion Expression (EESC) Emotion Regulation (CSMS) and (CAMS) Social Perspective Coordination (4+Rel-Q) Depressive Symptomatology (RADS-2) Students who did not participate in this study worked on independent assignments previously assigned by their teacher, or were given activities (e.g., puzzles, brain-teaser books) that we supplied. Altogether, the administration of the questionnaire was completed within 60 minutes of scheduled class time for both grade levels. Follow-Up Protocol Post Data Administration A resource pamphlet developed for this study (Appendix T) was distributed, immediately after data administration, to all the students in the participating classrooms to provide them with access information to various resources in and outside of their community. Several resource 78 Chapter III: Methodology pamphlets were left with the participating teacher for any students who were away on the day of data administration and the school counsellors of the participating schools. Immediately after the completion of data collection, students were invited to discuss with me any further questions they may have had. I made myself available outside the classroom in a designated spot for 20 minutes. This process afforded confidentiality to any adolescents who wanted to immediately discuss their personal concerns or questions with me outside the classroom. Protocol for Participants Who Were Screened as Depressed Each student participating in the survey was screened for depressive symptomatology using RADS-2, and the referral protocol developed for this study (Appendix U, Appendix V), was implemented for those screened as depressed. Potential depression was assessed after data administration, while still in the school, in an unused room previously allocated by the participating teacher or school personnel for this purpose. For each adolescent who completed the questionnaire that day, I tabulated their total scores twice on the RADS-2. The scores were then again reviewed with one of the research assistants who was present for the scoring of RADS-2 for all participants throughout data administration. Those ID numbers that contained at and above the cutoff score of 76 (screened as depressed) were immediately matched with the same numbers and names on identification pages that were stored separately. These were separated that day in a separate envelope to enable me to follow them up using the follow-up protocol later that day. The students' up-to-date home numbers, work numbers, and cell numbers were provided by the parents or guardians on the returned consent forms that included the students' names (Appendix A). Within 24 hours of data collection, I contacted and communicated directly with each participant who screened with potential depression at their home via telephone. In that same time frame, the students' parents or guardians were sent a resource pamphlet and a note (Appendix T, Appendix W) to inform them that their child was referred to the school counsellor. Referrals to the school counsellors were made within 24 hours of administering the questionnaires for those participants screened with depression. Often, referrals were made to the school counsellors immediately after the scoring of the RADS-2 was completed. If the counsellors were not available that day, I communicated directly with them via the telephone to 79 Chapter III: Methodology provide only the referred names. The process of immediate referrals was made easier because communication was already established, prior to data collection, with the school counsellors via phone or face to face to confirm my presence in their school the same day, and they had been provided with the list of the classrooms that were participating in the present study. Just prior to the day of data administration, the office front desk personnel were also provided with a list of participating classrooms to ease communication among the school personnel and to inform them about our presence in their school. 80 Chapter IV: Results CHAPTER IV: RESULTS The results of this research are presented in four sections corresponding to the findings related to this study's four questions. The first section begins with a presentation of participants' conceptions of adolescent depression and frequencies at different phases of data reduction. The second section follows with a presentation of the findings of the association of students' conceptions of depression to grade, gender, depressive symptomatology, self-recognition of depression, thinking about own need to talk to someone when depressed, and talking to someone when feeling depressed. In the third section, I present the results of the association of self- recognition of depression to depressive symptomatology, and to pathways to choosing to talk to someone at the time of feeling depressed. The results of regression analyses examining the relations of social perspective coordination, emotion awareness and expression of emotion, and emotion regulation, to dimensions of adolescents' depressive symptomatology, are presented in the final fourth section. Categories and Subcategories of Adolescents' Conceptions of Depression and their Frequencies To address the first question in this study in which the students were asked to generate their own definitions or conceptions of adolescent depression, frequency counts of participants' generated definitions of depression were summed across the categories and subcategories. In the initial part of this section, participants' total frequencies of response units, the sum of identified COAD units, students' range of COAD units, and means of COAD units (for unit definition and its operationalization, see Coding Schemes section in Methodology Chapter), are presented. Next, I report the frequency counts and percentages of students' COAD units of all their definitions of adolescent depression for each category and subcategory. Following is a discussion of the process of deriving definition units of adolescent depression to determine the presence of a COAD unit in a category or subcategory generated by each respondent. The final part of this section ends with a presentation of the data reduction strategies employed. 81 Chapter IV: Results To examine the frequencies of participants' response units,' the identified COAD units, range of units, means of COAD units, the frequency counts for adolescents' responses were tabulated in ACCESS (2003); and COAD units, range of units, and means of COAD units were calculated using the SPSS (Version 13.0, 2004; see Table 9). Table 9 Frequency Distribution of Participants' Response Breakdown of Conceptions of Adolescent Depression (COAD), N = 332 Conceptions of Adolescent Depression (COAD) Raw Response Units f COAD Units f Range of Units COAD Units M (SD) Missing Responses n (%) COAD (with Miscellaneous Category) COAD (without Miscellaneous Category) 1,816 1, 813 2,074 1,940 0-29 0-26 6.25 (3.54) 5.84 (3.38) 5 (1.5%) 8 (2.4%) Note. Including the Miscellaneous Category, there were two boys (1.4% of boys) and three girls (1.6% of girls) who did not provide descriptions of adolescent depression. Excluding the Miscellaneous Category, there were four boys (2.7% of the boys) and four girls (2.2% of the girls) who did not provide unit definitions for adolescent depression. As can be seen in Table 9, in total, participants provided 2,074 COAD units of depression definitions. Students submitted a range of 0 to 29 COAD units with an average of six COAD units per individual. To assess the occurrence of participants' COAD units in each category and subcategory, frequency counts and percentages of COAD unit definitions of depression were summed in each 16 The following unit definitions were copied from the Coding Schemes section described in the Methodology Chapter. "A response unit represents one bulleted response or a response separated by a new line by each participant in his or her generated list of definitions of adolescent depression. A COAD unit distinguishes itself by the separation of each response unit into elements or the smallest information of adolescent depression via content analysis. Analyses in this study are based on the COAD unit information." 82 Chapter IV: Results category and subcategory. Table 10 displays the 18 categories, and the 17 subcategories, and the frequency distribution of COAD units that emerged from participants' definitions of depression. The frequencies of COAD units that the students most referred to were in Depressed Mood and Social Impairment. These two categories contained 14 of the 17 subcategories (10 subcategories in Depressed Mood and four subcategories in Social Impairment). Table 10 Frequency Distribution and Percentages of COAD Unit Definitions, Total units = 2,074 (N = 332) ITEM COAD Unit Definitions f COAD Unit Definitions % 1. DEPRESSED MOOD 715 34.5% a. Sadness 263 12.7% b. Irritability 161 7.8% c. Moody 30 1.4% d. Somatic Symptoms 1 .04% e. Inner Pain 41 2.0% f. Depressed Appearance 73 3.5% Depressed Cognitions 146 7.1% g. Hopelessness 52 2.5% h. Unmotivated 53 2.6% i. Bored/Indifferent 14 .7% j. Pessimism 27 1.3% 2. LOSS OF INTEREST OR PLEASURE 13 .60/0 3. APPETITE/WEIGHT/SLEEP CHANGES 42 2.0% 4. FATIGUE OR LOSS OF ENERGY 37 1.8% 5. LOW SELF-WORTH/GUILT 121 5.8% 6. INDECISIVENESS/DIMINISHED ABILITY 63 3.0% TO THINK OR CONCENTRATE 7. FEEL TRAPPED 24 1.2% 8. SUICIDAL 89 4.3% 83 Chapter IV: Results ITEM COAD Unit Definitions f COAD Unit Definitions 0/0 9. SOCIAL IMPAIRMENT 441 21.3% a. Lonely 119 5.7% b. Perceived Disconnection 93 4.5% c. Behavioral Disconnection 194 9.4% d. Aggression/Bullied 35 1.7% 10. FUNCTIONAL IMPAIRMENT/DISTRESS 54 2.6% 11. RETARDATION/AGITATION 46 2.2% a. Quiet/Silent 43 2.1% b. Agitated 2 .09% c. Retardation 1 .05% 12. ANXIETY SYMPTOMS 90 4.3% a. Fear 34 1.6% b. Anxious/Stressed 56 2.7% 13. SUBSTANCE USE 32 1.5% 14. CONTEXTUAL/CAUSAL 128 6.2% 15. BIOLOGICAL 8 .4% 16. MASKING/STIGMA 27 1.3% 17. ESCAPING REALITY 10 18. MISCELLANEOUS 134 6.5% Note. See Methodology Chapter for definitions of categories and subcategories of adolescent depression. Out of a total of 2,074 COAD units of definitions of adolescent depression that were generated by 327 adolescents in the sample (5 did not respond to this item), 1,156 (56%) of the COAD definition units were concentrated in the categories Depressed Mood and Social Impairment. These two categories contained 82% of the identified subcategories. Particular COAD units clustered within the Miscellaneous category (e.g., out of sight, a person who has hallucinations, thinks often of the past, trying to please your parents, overanalyze everything, depression does not really mean much to me, as a way of relief, taken a little too lightly, emotion) did not address the question or were captured by the developed categories or subcategories (see Table 8 for definitions and more examples of the Miscellaneous category in 84 Chapter IV: Results the Methodology Chapter). The 134 Miscellaneous COAD unit definitions were almost evenly distributed between the 8 th graders and the 11 th graders. Eighth graders had 49 COAD unit definitions (2.4% of total COAD unit definitions) of depression and the 11 th graders 85 COAD units (4.1% of the total COAD unit definitions). Recall, there were almost twice as many grade 11 as grade 8 students participating in this study. Boys and girls were almost evenly split in the number of COAD units in the Miscellaneous category. Boys provided 60 COAD units (2.9% of the total COAD unit definitions) and girls 74 COAD units (3.6% of the total COAD unit definitions). This category was excluded from further analyses leaving 1,940 COAD units generated by participants' definitions of depression. Gender and Grade Frequency Distributions of Conceptions of Adolescent Depression Unit Definitions Generated Per Individual Grade 8 participants (n = 114) had a higher overall number of COAD units defining adolescent depression than the grade 11 students (n = 218). Table 11 displays the number of COAD units defined per participant, range of units, and means for differences in gender and grade. Regarding gender differences in the frequencies of COAD units provided, girls (n = 184) had a considerably higher frequency count than boys (n = 148) did. Table 11 Gender and Grade Differences in the Number of COAD Units Generated Per Person, Range of COAD Units, and Means of COAD units Variable COAD Units f Range of Units COAD Units M (SD) Grade 8 744 0 - 21 6.53 (3.91) 11 1,196 0 - 26 5.49 (3.02) Gender Boys 735 0 - 21 4.97 (2.94) Girls 1,205 0 - 26 6.55 (3.56) Note. N= 332, Total COAD units = 1,940, M= 5.84 (3.38). No Miscellaneous category in this analysis. 85 Chapter IV: Results To calculate gender and grade differences in the average number of COAD units generated per individual, t tests were conducted. For the average number of COAD units generated, girls (M= 6.55, SD = 3.56) defined an average of COAD units significantly more often than did boys (M= 4.97, SD = 2.94), t (330) = 4.35, p = .000 (two-tailed). For grade differences in the average number of COAD units generated per person, the 8 th graders (M= 6.53, SD = 3.91) defined an average of COAD units significantly more often than did the 11 ` I' graders (M= 5.49, SD = 3.02), t (330) = 2.69, p = .008 (two-tailed). Frequency Distributions of Conceptions of Adolescent Depression Units Generated At Least Once Per Individual To assess the presence of each category or subcategory generated by each participant's COAD units, each construct was counted as "yes" or coded as "1" (participant provided a definition for the construct), or "no" or "0" (participant did not produce a description for the construct). Hence, students received a score of "1" if a category was presented in their COAD unit definition(s), regardless of the number of times the category was represented (Carlson Jones, 2001). This analysis approach examined the intention of the participants to really mean to list that category and its subcategory definition of adolescent depression. Given that the free responses ranged from 0 to 26 COAD units offered per individual, more than one type of unit descriptions of adolescent depression could be provided by participants that fit the same category or subcategory. Table 12 illustrates the range of COAD definition units as defined by participants in each category and subcategory, and the frequency distribution of categories and subcategories generated at least once by the participants. This process involved collapsing the range of responses to either defined or not defined by each student producing frequency distributions for presenting category and subcategory data at least once by each participant (Carlson Jones, 2001). In Appendix X, Table X-1 illustrates only the frequency distribution of categories and subcategories defined at least once by each participant. 86 Chapter IV: Results Table 12 Frequency Distribution and Percentages of LOAD Units Generated by Participants at Least Once and the Range of LOAD Information Units in Each Category and Subcategory (N = 332) Category/Subcategory Defined Units n = f of units Defined Units % Not Defined Unit n (%) 1. DEPRESSED MOOD 287 = 1 or more units 86.4% 45 (13.6%) 72 = 1 unit 21.7% 97 = 2 units 29.2% 58 = 3 units 7.5% 40 = 4 units 12.0% 12 = 5 units 3.6% 4 = 6 units 1.2% 2 = 7 units .6% 2 = 9 units .6% a. Sadness 209 = 1 or more units 63.0% 123 (37.0%) 162 = 1 unit 48.8% 41 = 2 units 12.3% 5 = 3 units 1.5% 1 = 4 units . 3% b. Irritability 120 = 1 or more units 36.1% 212 (63.9%) 90 = 1 unit 27.1% 22 = 2 units 6.6% 6 = 3 units 1.8% 1 = 4 units . 3% 1 = 5 units . 3% c. Moody 27 = 1 or more units 8.1% 305 (91.9%) 24 = 1 units 7.2% 3 = 2 units . 9% d. Somatic Symptoms 1 = 1 unit . 3% 331 (99.7%) e. Inner Pain 38 = 1 or more units 11.4% 294 (88.6%) 35 = 1 unit 10.5% 3 = 2 units . 9% f. Depressed Appearance 58 = 1 or more units 17.5% 274 (82.5%) 46 = 1 unit 13.9% 10 = 2 units 3.0% 1 = 3 units . 3% 1 = 4 units . 3% 87 Chapter IV: Results Category/Subcategory Defined Units n = f of units Defined Units % Not Defined Unit n (%) Depressed Cognitions 137 = 1 or more units 41.3% 195 (58.7%) g.^Hopelessness 50 = 1 or more units 15.1% 282 (84.9%) 47 = 1 unit 14.2% 3 = 2 units . 9% h.^Unmotivated 48 = 1 or more units 14.5% 284 (85.5%) 43 = 1 unit 13.0% 5 = 2 units 1.5% i. Bored/Indifferent 14 = 1 or more units 4.2% 318 (95.8%) 14 = 1 unit 4.2% j. Pessimism 25 = 1 or more units 7.5% 307 (92.5%) 23 = 1 unit 6.9% 2 = 2 units .6% 2. LOSS OF INTEREST 10 = 1 or more units 3.0% 322 (97.0%) OR PLEASURE 7 = 1 unit 2.1% 3 = 2 units . 9% 3. APPETITE/WEIGHT/ 30 = 1 or more units 9.0% 302 (91.0%) SLEEP CHANGES 20 = 1 unit 6.0% 9 = 2 units 2.7% 1 = 4 units .3% 4. FATIGUE OR LOSS 37 = 1 or more units 11.1% 295 (88.9%) OF ENERGY 37 = 1 unit 11.1% 5. LOW SELF-WORTH/ 87 = 1 or more units 26.2% 245 (73.8%) GUILT 62 = 1 unit 18.7% 19 = 2 units 5.7% 4 = 3 units 1.2% 1 = 4 units . 3% 1 = 5 units . 3% 6. INDECISIVENESS/ 50 = 1 or more units 15.1% 282 (84.9%) IMPAIRED 40 = 1 unit 12.0% CONCENTRATION 8 = 2 units 2.4% AND ATTENTION 1 = 3 units .3% 1 = 4 units .3% 88 Chapter IV: Results Category/Subcategory Defined Units n -= f of units Defined Units % Not Defined Unit n (%) 7. FEEL TRAPPED 19 = 1 or more units 5.7% 313 (94.3%) 15 = 1 unit 4.5% 3 = 2 units . 9% 1 = 3 units . 3% 8. SUICIDAL 69 = 1 or more units 21.8% 263 (79.2%) 51 = 1 unit 15.4% 16 = 2 units 4.8% 2 = 3 units .6% 9. SOCIAL 228 = 1 or more units 68.7% 104 (31.3%) IMPAIRMENT 105 = 1 unit 31.6% 70 = 2 units 21.1% 35 = 3 units 10.5% 10 = 4 units 3.0% 3 = 5 units . 9% 2 = 6 units .6°/0 1 = 7 units . 3% 1 = 8 units . 3% 1 = 9 units . 3% c.^Behavioral 140 = 1 or more units 42.2% 192 (57.8%) Disconnection 102 = 1 unit 30.7% 30 = 2 units 9.0% 3 = 3 units .9% 2 = 4 units .6% 3 = 5 units .9% b.^Perceived 71 = 1 or more units 21.4% 261 (78.6%) Disconnection 53 = 1 unit 16.0% 14 = 2 units 4.2% 4 = 3 units 1.2% a.^Lonely 117 = 1 or more units 35.2% 215 (64.8%) 115 = 1 unit 34.6% 2 = 2 units .6% d.^Aggression/Bullied 27 = 1 or more units 8.1% 305 (91.9%) 20 = 1 unit 6.0% 6 = 2 units 1.8% 1 = 3 units . 3% 89 Chapter IV: Results Category/Subcategory Defined Units n = f of units Defined Units % Not Defined Unit n (%) 10. FUNCTIONAL 44 = 1 or more units 13.3% 288 (86.7%) IMPAIRMENT/ 36 = 1 unit 10.8% DISTRESS 7 = 2 units 2.1% 1 = 3 units . 3% 11. AGITATION/ 44 = 1 or more units 13.3% 288 (86.7%) RETARDATION 42 = 1 unit 12.7% 2 = 2 units .6% a.^Quiet/Silent 41 = 1 or more units 12.3% 291 (87.7%) 39 = 1 unit 11.7% 2 = 2 units .6°/0 12. ANXIETY 65 = 1 or more units 19.6% 267 (80.4%) SYMPTOMS 48 = 1 unit 14.5% 13 = 2 units 3.9% 2 = 3 units .6% 1 = 4 units .3% 1 = 6 units .3% a. Fear 30 = 1 or more units 9.0% 302 (91.0%) 26 = 1 unit 7.8% 4 = 2 units 1.2% b. Anxious/Stressed 43 = 1 or more units 13.0% 289 (87.0%) 34 = 1 unit 10.2% 8 = 2 units 2.4% 1 = 6 units . 3% 13. SUBSTANCE USE 20 = 1 or more units 6.0% 312 (94.0% 10 = 1 unit 3.0% 9 = 2 units 2.7% 1 = 4 units . 3% 14. CONTEXTUAL/CAUSAL 73 = 1 or more units 22.0% 259 (78.0%) 40 = 1 unit 12.0% 20 = 2 units 6.0% 9 = 3 units 2.7% 2 = 4 units .6% 1 = 6 units . 3% 1 = 7 units . 3% 90 Chapter IV: Results Category/Subcategory Defined Units n = f of units Defined Units % Not Defined Unit n (%) 15. BIOLOGICAL 8 = 1 or more units 2.4% 324 (97.6%) 8 = 1 unit 2.4% 16. MASKING/STIGMA 24 = 1 or more units 7.2% 308 (92.8%) 21 = 1 unit 6.3% 3 = 2 units . 9% 17. ESCAPING REALITY 10 = 1 or more units 3.0% 322 (97.0%) 10 = 1 unit 3.0% Note. The total number of COAD units provided at least once per individual was 1,194 with the Miscellaneous category, and after removal of the Miscellaneous category (89), the remaining total of COAD units was 1,105. As can be seen in Table 12, Depressed Mood and Social Impairment contained the highest range of units. The remaining total number of COAD units provided at least once per individual was 1,105 excluding the Miscellaneous category. Percentages of COAD Units Generated by Boys and Girls At Least Once in Each Construct of Adolescent Depression To assess the occurrence of boys' and girls' COAD units in each category and subcategory, frequency counts of COAD units of adolescent depression were summed and calculated into percentages in each construct of adolescent depression. Table 13 displays the 17 categories and 17 subcategories, and the percentages of COAD units that emerged from boys' and girls' definitions of depression at least once. 91 Chapter IV: Results Table 13 Percentages of COAD Units Generated by Boys and Girls at Least Once in Each Category and Subcategory of Adolescent Depression, N = 332 Category Subcategory Boys n = 148 % Girls n = 184 % Depressed Mood 84.5% 88.0% Sadness 62.2% 63.6% Irritability 32.4% 39.1% Moody 6.1% 9.8% Somatic Symptoms 0% .5% Inner Pain 11.5% 11.4% Depressed Appearance 10.8% 22.8% Depressed Cognitions Hopelessness 18.9% 12.0% Unmotivated 14.2% 14.7% Bored 6.1% 2.7% Pessimism 7.4% 7.6% Loss of Interest or Pleasure 1.4% 4.3% Appetite/Wt/Sleep Changes 7.4% 10.3% Fatigue or Loss of Energy 10.8% 11.4% Low Self-Worth/Guilt 18.9% 32.1% Indecisiveness/Impaired 13.5% 16.3% Concentration and Attention Feel Trapped 4.7% 6.5% Suicidal 14.9% 25.5% Social Impairment 60.1% 75.5% Behavioral Disconnection 37.8% 45.7% Perceived Disconnection 14.2% 27.2% Lonely 32.4% 37.5% Aggression/Bullied 4.7% 10.9% Functional Impairment/Distress 15.5% 11.4% Retardation/Agitation 8.1% 17.4% Quiet 7.4% 16.3% Anxiety Symptoms 15.5% 22.8% Fear 6.1% 11.4% Anxious/Stressed 10.8% 14.7% Substance Use 4.1% 7.6% Contextual/Causal 22.3% 21.7% Biological 1.4% 3.3% Masking/Stigma 3.4% 10.3% Escaping Reality 2.0% 3.8% As can be seen in Table 13, the majority of boys generated their COAD units in the categories Depressed Mood, Social Impairment, and the subcategory Sadness. Further, about 20% or more boys provided COAD units in the constructs Irritability, Behavioral Disconnection, Lonely, Hopelessness, Low Self-Worth/Guilt, and Contextual/Causal. Although boys generated 92 Chapter IV: Results lower percentages of COAD units than girls in most categories and subcategories of depression, the constructs in which boys had somewhat similar or slightly higher percentages of COAD units than girls included Hopelessness, Inner Pain, Bored, Functional Impairment/Distress, and Contextual/Causal. Similar to boys, the majority of the girls generated COAD units in the constructs Depressed Mood, Social Impairment, and Sadness. However, girls had a greater number of different constructs than boys in which they generated at least 20% of their COAD units. These constructs included Irritability, Depressed Appearance, Low Self-Worth/Guilt, Suicidal, Behavioral Disconnection, Perceived Disconnection, Lonely, Anxiety Symptoms, and Contextual/Causal. To assess the occurrence of boys' and girls' COAD units for each construct of adolescent depression in grade 8 and in grade 11, frequency counts of COAD units of adolescent depression were summed and calculated into percentages (see Table 14). 93 Chapter IV: Results Table 14 Percentages of Adolescent Depression Constructs by Grade Level in Boys, and in Girls, N = 332 Boys n = 148 Girls n = 184 Category Subcategory Grade 8 n = 48 Grade 11 n = 100 Grade 8 n = 66 % Grade 11 n = 118 Depressed Mood 93.8% 80.0% 87.9% 88.1% Sadness 66.7% 60.0% 63.6% 63.6% Irritability 37.5% 30.0% 39.4% 39.0% Moody 6.3% 6.0% 7.6% 11.0% Somatic Symptoms 0% 0% 1.5% 0% Inner Pain 22.9% 6.0% 12.1% 11.0% Depressed Appearance 18.8% 7.0% 33.3% 16.9% Depressed Cognitions Hopelessness 16.7% 20.0% 6.1% 15.3% Unmotivated 20.8% 11.0% 10.6% 16.9% Bored 6.3% 6.0% 1.5% 3.4% Pessimism 12.5% 5.0% 10.6% 5.9% Loss of Interest or Pleasure 4.2% 0% 0% 6.8% Appetite/Wt/Sleep Changes 6.3% 8.0% 7.6% 11.9% Fatigue or Loss of Energy 16.7% 8.0% 6.1% 14.4% Low Self-Worth/Guilt 12.5% 22.0% 39.4% 28.0% Indecisiveness/Impaired 20.8% 10.0% 18.2% 15.3% Concentration and Attention Feel Trapped 6.3% 4.0% 4.5% 7.6% Suicidal 20.8% 12.0% 33.3% 21.2% Social Impairment 60.4% 60.0% 71.2% 78.0% Behavioral Disconnection 35.4% 39.0% 48.5% 44.1% Perceived Disconnection 14.6% 14.0% 34.8% 22.9% Lonely 37.5% 30.0% 33.3% 39.8% Aggression/Bullied 2.1% 6.0% 16.7% 7.6% Functional Impairment/Distress 16.7% 15.0% 12.1% 11.0% Retardation/Agitation 12.5% 6.0% 13.6% 19.5% Quiet 12.5% 5.0% 13.6% 17.8% Anxiety Symptoms 14.6% 16.0% 16.7% 26.3% Fear 4.2% 7.0% 9.1% 12.7% Anxious/Stressed 10.4% 11.0% 9.1% 17.8% Substance Use 6.3% 3.0% 10.6% 5.9% Contextual/Causal 22.9% 22.0% 30.3% 16.9% Biological 0% 2.0% 0% 5.1% Masking/Stigma 2.1% 4.0% 10.6% 10.2% Escaping Reality 0% 3.0% 7.6% 1.7% As can be seen in Table 14, the majority of boys, in both grade 8 and in grade 11, generated definitions of adolescent depression in the constructs Depressed Mood, Social Impairment, and Sadness with greater percentages of COAD units contributed by grade 8 boys in each of these constructs. Regarding the subcategory Inner Pain, 23% of grade 8 boys provided a 94 Chapter IV: Results COAD unit in this construct in contrast to 6% of the boys in grade 11. However, the subcategory Hopelessness was one construct where the boys in grade 11 generated more COADs (20%) than the boys in grade 8 (17%), and had higher percentages of COADs than the girls in both grade levels. Similar to boys, the majority of the girls, in grade 8 and in grade 11, contributed their COADs in Depressed Mood, Social Impairment, and Sadness constructs. In contrast to boys, regarding the subcategory Inner Pain, girls' percentages of COADs remained similar in both grades 8 and 11, 12% and 11% respectively. Similar to boys, girls in grade 11 had higher percentages of COADs than girls in grade 8 in the subcategory Hopelessness. Just over 10% of the girls in both grade levels generated their COADs in the category Masking/Stigma. In contrast, less than 5% of the boys provided their COADs in this construct. Gender and Grade Frequency Distributions of COAD Unit Definitions Generated At Least Once Per Individual Table 15 displays the number of COAD units defined at least once per participant, range of units, and means for the differences in gender and grade. Using descriptive analyses (SPSS, Version 13.0, 2004), grade 8 participants (n = 114) had a higher overall number of COAD units than the grade 11 students (n = 218). Regarding gender differences in the frequencies of COAD unit descriptions provided, girls (n = 184) had a higher frequency count than boys (n = 148) did. 95 Chapter IV: Results Table 15 Gender and Grade Differences in the Number of COAD Units Defined at Least Once per Participant, Range of Units, and Means of COAD units Variable COAD Units I Range of Units COAD Units M (SD) Grade 8 396 0 - 8 3.47 (1.64) 11 709 0-10 3.25 (1.53) Gender Boys 427 0 - 7 2.89 (1.38) Girls 678 0-10 3.68 (1.63) Note. N= 332, Total COAD units = 1,105, M= 3.33 (1.57). No Miscellaneous category in this analysis and further analyses. To calculate gender and grade differences in the average number of units generated at least once per individual, t tests were conducted (see Table 15). For the average number of COAD units generated at least once, girls (M= 3.68, SD = 1.63) defined COAD units significantly more often than did boys (M= 2.89, SD = 1.38), t (330) = 4.76, p = .000 (two- tailed). Girls also had a broader range of constructs than did boys. For grade differences in the average number of COAD units generated at least once per person, the 8 th graders (M = 3.47, SD = 1.64) defined an average of COAD units slightly more often than did 11 th graders (M= 3.25, SD = 1.53), t (330) = 1.22, p = .223 (two-tailed) (not significant). However, the range of constructs was wider for the 11 th graders than it was for the 8 th graders. Construct Reduction According to criteria outlined by Bigelow (1977), categories or subcategories containing less than five percent (< 5%) of the participants were omitted from further analyses. Data reduction (Blaikie, 2003) was employed to reduce the number of constructs because participants' low percentages in these constructs could obscure the findings by not providing adequate numbers in cells in subsequent analyses. Three categories and two subcategories were removed leaving 14 categories and 15 subcategories (see the removed < 5% in Table 12 and remaining > 5% constructs in Table Y-2 in Appendix Y). Categories and subcategories elicited by fewer 96 Chapter IV: Results than 5% of participants were Loss of Interest or Pleasure, Biological, Escaping Reality, Somatic Symptoms, and Bored/Indifferent. These categories and subcategories were omitted from further analyses rather than merged with other remaining larger categories because these constructs depicted participants' specific descriptors of depression, and reliabilities using Cohen's kappas were already established for the categories. Subsequent analyses were carried out on the 14 categories and 15 subcategories that were generated by five or more percent (> 5%) of the participants. A total of 1,077 COAD units were generated by five or more percent of adolescents at least once for the remaining 14 categories. Summary Girls significantly more often than boys, and 8 th graders more often than 11 th graders, defined an average number of COAD units generated at least once. Girls generated a wider range of different constructs than boys. However, although the 8th graders generated on average more COAD units than the 11 th graders, the grade 11 students had a wider range of COAD units than the grade 8 adolescents. Depressed Mood (n = 287, 86.4%) and Social Impairment (n = 228, 68.7%) were the two categories most referred to in adolescents' definitions of depression (see Table 12 or Table Y-2 in Appendix Y). The next four categories lagging behind the first two categories were: Low Self- Worth (n = 87, 26.2%); Contextual/Causal (n = 73, 22.0%); Suicidal (n = 69, 20.8%); and Anxiety Symptoms (n = 65, 19.6%). The six least referred to categories (including the categories mentioned by < 5% of participants) were: Biological (n = 8, 2.4%); Loss of Interest or Pleasure (n = 10, 3.0%); Escaping Reality (n = 10, 3.0%); Feel Trapped (n = 19, 5.7%); Substance Use (n = 20, 6.0%); and Masking/Stigma (n = 24, 7.2%). 1 7 In keeping with the high percentage of participants providing COAD units that referred to Depressed Mood and Social Impairment, 12 out of the 15 subcategories were clustered within these two categories. The six subcategories most referred to were: Sadness (n = 209, 63.0%); Behavioral Disconnection (n = 140, 42.2%); Irritability (n = 120, 36.1%); Lonely (n = 117, 35.2%); Perceived Disconnection (n = 71, 21.4%); and Depressed Appearance (n = 58, 17.5%). In contrast, the subcategories that participants defined least commonly (including the 17 See Methodology Chapter for definitions of these categories. 97 Chapter IV: Results subcategories described by < 5% of participants) were: Somatic Symptoms (n = 1, .3%); Bored/Indifferent (n = 14, 4.2%); Pessimism (n= 25, 7.5%); Moody (n = 27, 8.1%); Aggression/Bullied (n = 27, 8.1%); and Fear (n = 30, 9.0%). Depressed Mood and Social Impairment categories, as defined by the high percentage of participants' descriptions of adolescent depression units, are found as criteria used to diagnose MDE in the DSM-IV-TR (APA, 2000; see Table 5 for criteria for MDE in the Classification of Participants' Units of Adolescent Depression section in Methodology Chapter) and the K-SADS- PL (Kaufman et al., 1996). However, some descriptions or subcategories that constitute Depressed Mood as described by the participants in this study, differ or add to the depression criteria found in the DSM-IV-TR (APA, 2000) or the K-SADS-PL (Kaufman et al., 1996). Further, adolescents highlighted Social Impairment, generated by their COAD units, as the key impairment. This finding contrasts with the DSM-IV-TR (APA, 2000) criterion for MDE which integrates social impairment within a general functioning impairment (see Criterion C in Table 5). Association of Adolescent Depression Constructs to Grade, Gender, Depressive Symptomatology, Self-Recognition of Depression, and a Pathway to Talking to Someone at the Time of Feeling Depressed The second question in this study was, "Do adolescents' conceptions of depression vary by grade, gender, depressive symptomatology, self-recognition of depression, thinking about own need to talk to someone, and talking to someone when feeling depressed?" To address this question, percentages were calculated to present the data (Blaikie, 2003) and chi-square (x 2) tests of significance at .05 level were used to compute the associations between these dichotomous variables (Blaikie, 2003; Carlson Jones, 2001; Clinical Epidemiology & Biostatistics Unit, 1995; Tabachnik & Fidell, 2001; Yang & Chen, 2006). In this second section, a presentation is put forth of the analyses used to examine the associations of the generated categories and subcategories of adolescent depression to grade level; gender; depressive symptomatology; self-recognition of depression; thinking about own need to talk to someone; and talking to someone at the time of feeling depressed. Given the vast amount of data produced by these analyses, only significant x2 and near significant results are 98 Chapter IV: Results presented in partial tables (p > .00 and < .10), and the complete tables can be viewed in the Appendices. Near significant results are presented in the body of the text tables rather than in the Appendices to provide an overview of the differences in these findings. Association between Adolescent Depression Constructs and Grade Table 16 lists the percentage distribution and x2 analyses results (significant and near significant) of adolescent depression categories and subcategories by grade. Some significant differences were found between the 8 th and 1 1 th graders in association with three subcategories (all contained in the Depressed Mood category)• Inner Pain, Depressed Appearance, and Pessimism; and one Suicidal category. Chi-square tests revealed that grade 8 participants were significantly more likely than grade 11 students to provide description units in each of these four significant associations: Inner Pain, x2 (1, N= 332) = 4.67,p = .03; for Depressed Appearance, x2 (1, N= 332) = 11.38, p = .00; for Pessimism, x2 (1, N= 332) = 3.74, p = .05; and for Suicidal, x2 (1, N = 332) = 5.60,p = .02. More than 80% of the students in both grades 8 and 11 mentioned COAD units in the category Depressed Mood, and more than 60% in the category Social Impairment and in the subcategory Sadness. Subcategories Irritability, Behavioral Disconnection, and Lonely, contained more than 30% of the students' COAD units in both grades (see Table Z-3 in Appendix Z for complete results). 99 Chapter IV: Results Table 16 Percentages, Significant and Near Significant (< .10) Chi-Square Tests in Adolescent Depression Categories and Subcategories by Grade, N = 332 Category / Subcategory Grade 8 (%) n = 114 Grade 11 (%) n = 218 x 2 (1, N= 332) Depressed Mood Inner Pain 16.7% 8.7% 4.67, p = .03 Depressed Appearance 27.2% 12.4% 11.38, p = .00 Hopelessness 10.5% 17.4% 2.79, p = .10 Pessimism 11.40/0 5.5% 3.74, p = .05 Suicidal 28.1% 17.0% 5.60, p = .02 Contextual/Causal 27.2% 19.3% 2.74, p = .10 Note. Categories and Subcategories deleted if N < 5% for this Table and the remaining Tables in this chapter. Results are shown only where p> .00 and < .10 for this Table and the remaining Tables in this chapter. Complete results for this Table and the remaining Tables in this chapter are shown in the Appendices. As can be seen in Table 16, significantly higher percentages of grade 8 adolescents than grade 11 participants provided COAD units in the constructs Inner Pain, Depressed Appearance, Pessimism, and Suicidal. Association between Adolescent Depression Constructs and Gender Table 17 lists the percentage distribution and x 2 analysis results of adolescent depression categories and subcategories by gender. Nine depression constructs differed significantly between boys and girls in the percentages of adolescents who provided COAD units in these areas. A significantly higher percentage of girls than boys generated COAD units in the categories: Low Self-Worth, Suicidal, Social Impairment, Retardation/Agitation, Masking/Stigma; and in the subcategories: Depressed Appearance, Perceived Disconnection, Aggression/Bullied, and Quiet. More than 80% of the boys and girls mentioned COAD units in the category of Depressed Mood, and more than 60% in the category of Social Impairment and in the subcategory of Sadness. Subcategories Irritability, Behavioral Disconnection, and Lonely, contained more than 30% of the students' COAD units for both boys and girls (see Table A1-4 in 100 Chapter IV: Results Appendix Al for complete results). These high percentages of COAD units in these constructs by both boys and girls suggest that they are representative of the adolescent depression concept as understood by the participants. Table 17 Percentages, Significant and Near Significant (< .10) Chi-Square Tests in Adolescent Depression Constructs by Gender, N = 332 Category / Subcategory Boys °/0 n = 148 Girls % n = 184 X2 (1, N= 332) Depressed Mood Depressed Appearance 10.8% 22.8% 8.21, p = .00 Hopelessness 18.9% 12.0% 3.11, p = .08 Low Self-Worth/Guilt 18.9% 32.1% 7.33, p = .01 Suicidal 14.9% 25.5% 5.68, p = .02 Social Impairment 60.1% 75.5% 9.05, p = .00 Perceived Disconnection 14.2% 27.2% 8.23, p = .00 Aggression/Bullied 4.7% 10.9% 4.14, p = .04 Retardation/Agitation 8.1% 17.4% 6.15, p = .01 Quiet 7.4% 16.3% 5.97, p = .02 Anxiety Symptoms 15.5% 22.8% 2.77, p = .10 Fear 6.1% 11.4% 2.84, p = .09 Masking/Stigma 3.4% 10.3% 5.90, p = .02 As can be seen in Table 17, girls had significantly higher percentages of COAD units than boys in the constructs Depressed Appearance, Low Self-Worth, Suicidal, Social Impairment, Perceived Disconnection, Aggression/Bullied, Retardation/Agitation, Quiet, and Masking/Stigma. The only construct where boys had a higher percentage of COAD units in than girls was Hopelessness (not significant). Association between Adolescent Depression Constructs and Gender in Grade 8, and in Grade 11 Chi-square tests revealed that adolescent depression constructs differed significantly between boys and girls in grade 8, and boys and girls in grade 11. Table 18 lists the percentage 101 Chapter IV: Results distribution and x2 tests (< .10) findings of adolescent depression constructs by gender in grade 8, and in grade 11. In grade 8, significant differences between boys and girls were found in association with one category, Low Self-Worth, and two subcategories, Perceived Disconnection and Aggression/Bullied. In grade 8, a higher percentage of girls than boys provided COAD units in each of these three constructs. In grade 11, x2 tests demonstrated significant gender differences in two categories, Social Impairment and Retardation/Agitation; and in two subcategories, Depressed Appearance and Quiet. Again, the percentage of girls mentioning COAD units in each of these constructs was higher than for boys (see Table 18). For grades 8 and 11, 80% (or more) of the boys and girls mentioned COAD units in the category Depressed Mood, and 60% (or more) in the category of Social Impairment and subcategory Sadness. Subcategories Irritability, Behavioral Disconnection, and Lonely, contained 30% (or more) of both boys' and girls' COAD units (see Table B1-5 in Appendix B1 for complete results). This high percentage of these constructs suggests that the understanding of these constructs of adolescent depression as common definitions does not differ between boys and girls at the different grade levels. 102 Chapter IV: Results Table 18 Percentages, Significant and Near Significant (< .10) Chi-Square Tests of Frequency in Adolescent Depression Constructs by Gender in Grade 8, and in Grade 11, N = 332 Category / Subcategory Grade 8 n = 114 Grade 11 n = 218 BOYS % GIRLS % X2 (1, N = 114) BOYS oh GIRLS oh X 2 (1, N = 218) Depressed Mood 80.0% 88.1% 2.72, p = .10 Depressed Appearance 18.8% 33.3% 2.99, p = .08 7.0% 16.9% 4.94, p = .03 Hopelessness 16.7% 6.1% 3.32, p =. 07 Fatigue or Loss of Energy 16.7% 6.1% 3.32, p = .07 Low Self-Worth/Guilt 12.5% 39.4% 9.96, p = .00 Suicidal 12.0% 21.2% 3.24, p = .07 Social Impairment 60.0% 78.0% 8.28, p = .00 Perceived Disconnection 14.6% 34.8% 5.89, p = .02 14.0% 22.9% 2.80, p = .09 Aggression/Bullied 2.1% 16.7% 6.28, p = .01 Retardation/Agitation 6.0% 19.5% 8.54, p = .00 Quiet 5.0% 17.8% 8.44, p = .00 Anxiety Symptoms 16.0% 26.3% 3.38, p = .07 Masking/Stigma 2.1% 10.6% 3.09, p = .08 4.0% 10.2% 3.03, p = .08 Chapter IV: Results As can be seen in Table 18, girls had higher percentages of COAD units than boys in all the significant constructs in each grade; however, the significant constructs differed between grades. Association between Adolescent Depression Constructs and Grade Level in Boys, and in Girls Chi-square tests revealed that a significant difference in the percentages of COAD units was found between boys in grade 8 and boys in grade 11 in the category Depressed Mood, x2 (1, N= 332) = 4.67,p = .03; and in its two subcategories, Inner Pain, X2 (1, N= 332) = 9.13,p = .00; and Depressed Appearance, x2 (1, N= 332) = 4.64, p = .03 (see Table 19) A higher percentage of boys in grade 8 than grade 11 boys provided COAD units in each of these significant findings. Significant differences were found between girls in grade 8 and girls in grade 11 in the category Contextual/Causal, x2 (1, N= 332) = 4.44, p = .04; and the subcategory Depressed Appearance, x2 (1, N= 332) = 6.45,p = .01. A greater percentage of grade 8 girls than grade 11 girls provided COAD units in these constructs (for complete Table results, see Table C 1-6 in Appendix Cl). For both boys and girls in grades 8 and 11, 80% (or more) mentioned COAD units in the category Depressed Mood, and 60% (or more) in the category Social Impairment and subcategory Sadness. Subcategories Irritability, Behavioral Disconnection, and Lonely, contained 30% (or more) of the boys' and girls' COAD units. This high percentage of these constructs suggests that the understanding of these constructs of adolescent depression as common definitions does not differ between grades for both boys and girls. 104 Chapter IV: Results Table 19 Percentages, Significant and Near Significant (< .10) Chi-Square Tests in Adolescent Depression Constructs by Grade Level in Boys, and in Girls, N = 332 Category / Subcategory Boys n = 148 Girls n = 184 Grade 8 n = 48 % Grade 11 n = 100 % X2 (1, N = 148) Grade 8 n = 66 Oh Grade 11 n = 118 0/0 x 2 (1, N = 184) Depressed Mood 93.8% 80.0% 4.67, p = .03 Inner Pain 22.9% 6.0% 9.13, p = .00 Depressed Appearance 18.8% 7.0% 4.64, p = .03 33.3% 16.9% 6.45, p = .01 Hopelessness 6.1% 15.3% 3.40, p = .07 Pessimism 12.5% 5.0% 2.65, p = .10 Fatigue or Loss of Energy 6.1% 14.4% 2.92, p = .09 Indecisiveness/Impaired Concentration and Attention 20.8% 10.0% 3.26, p = .07 Suicidal 33.3% 21.2% 3.28, p = .07 Social Impairment Perceived 34.8% 22.9% 3.06, p = .08 Disconnection Aggression/Bullied 16.7% 7.6% 3.57, p = .06 Quiet 12.5% 5.0% 2.65, p = .10 Contextual/Causal 30.3% 16.9% 4.44, p = .04 Chapter IV: Results As can be seen in Table 19, the only construct that was significant for both boys and girls where the higher percentages were found in grade 8, was in Depressed Appearance. Depressed Mood and its subcategory Inner Pain, were significant findings for boys only where higher percentage of boys in grade 8 than in grade 11 provided COAD units in this construct. Association between Adolescent Depression Constructs and Depressive Symptomatology To address the association between the adolescent depression constructs and depressive symptomatology, I first present univariate analyses findings of depressive symptomatology followed by the x2 tests results. To distinguish the participants into those screened as "Depressed" and "Not Depressed," scores from the RADS-2 measure were transformed and collapsed into two values of "Depressed" and "Not Depressed." The cut-off score for "Depressed" contained potentially mild to severe depression (score of 76 and above; see also Measures section in the Methodology Chapter), and a score of 75 and below represented "Not Depressed." To assess the occurrence of screened depression, frequency counts were tabulated for those individuals with a cut-off score of 76 and above. Thirty (9%) participants were screened as depressed: 23 girls (12.5% of the girls in the study) and 7 boys (4.7% of the boys in this study). In grade 8, seven girls (10.6% of the girls) and five boys (10.4% of the boys) were screened as depressed. In grade 11, a higher percentage of girls (n = 16, 13.6% of the girls) than boys (n = 2, 2.0% of the boys) were screened as depressed. The differences in the percentages between the grade 11 girls and the grade 11 boys screened as depressed are higher than epidemiological community studies on gender differences in depressive symptomatology, which are more often in the ratio of two to three girls to a boy after puberty (see Gender section in the Literature Review). However, the percentage of adolescents screened as depressed and the overall ratio of 23 girls versus 7 boys screened as depressed reflect the findings in the literature (Kessler & Walters, 1998). Chi-square tests results revealed significant differences between adolescent depression constructs and depressive symptomatology in the categories, Feel Trapped and Anxiety Symptoms; and in subcategories, Inner Pain and Anxious/Stressed. Table 20 lists the outcomes of x2 analyses. Thirty percent of "Depressed" participants versus 9.6% of "Not Depressed" adolescents mentioned Inner Pain, a characteristic not discussed as a clinical or threshold criterion for depression in the DSM-IV-TR (APA, 2000), x2 (1, N= 332) = 11.20,p = .001. A 106 Chapter IV: Results significant association was also demonstrated between the category, Anxiety Symptoms, and depressive symptomatology. A higher percentage of "Depressed" students (33.3%) versus "Not Depressed" participants (18.2%) mentioned COAD units in this category, x2 (1, N= 332) = 3.96, p = .05. This result is in keeping with studies that have shown a link between anxiety symptoms and depression (Compas & Oppedisano, 2000) and Mixed Anxiety-Depressive Disorder in the DSM-IV-TR (APA, 2000). A significantly higher percentage of "Depressed" (16.7%) adolescents than "Not Depressed" (4.6%) respondents mentioned COAD units in Feel Trapped category, x2 (1, N= 332) = 7.32, p = .01. A significantly higher percentage of "Depressed" participants (26.7%) in contrast to "Not Depressed" adolescents (11.6%) provided COAD units in the Anxious/Stressed subcategory, x2 (1, N= 332) = 5.50,p = .02. This subcategory is not identified as a clinical criterion for depression (DSM-IV-TR, APA, 2000; K-SADS-PL, Kaufman et al, 1996) although studies have found a relationship between stress and mood disorders (Goldberg & Huxley, 1993; Rutter, 2000). More than 80% of both "Depressed" and "Not Depressed" adolescents mentioned COAD units in the category of Depressed Mood and more than 60% in the subcategory of Sadness (see Table D1-7 in Appendix D1 for complete Table results). More than 50% of the participants generated COAD units in the Social Impairment category. Subcategories Irritability, Behavioral Disconnection, and Lonely, contained 30% (or more) of the students' COAD units in both "Depressed" and "Not Depressed" groups. These high percentages in both groups of adolescents suggest that these constructs are representative of the adolescent depression concept as understood by the participants. Of note, a higher percentage of "Not Depressed" adolescents (70.2%) versus "Depressed" participants (53.3%) contributed COAD units in the Social Impairment category, x2 (1, N= 332) = 3.61, p = .06; however not significant. Regarding subcategory Aggression/Bullied, there were no "Depressed" participants in contrast to 8.9% "Not Depressed" adolescents mentioning COAD units in this area (not significant). 107 Chapter IV: Results Table 20 Percentage Distribution, Significant and Near Significant (< .10) Chi-Square Tests of Frequency in Adolescent Depression Constructs by Depressive Symptomatology, N = 332 Not Depressed (<75)^Depressed (>76) Category / Subcategory^n = 302^n = 30 % % X2 (1, N= 332) Depressed Mood Inner Pain 9.6% 30.0% 11.20, p = .001 Low Self-Worth/Guilt 27.5% 13.3% 2.83, p = .09 Feel Trapped 4.6% 16.7% 7.32, p = .01 Social Impairment 70.2% 53.3% 3.61, p = .06 Aggression/Bullied 8.9% 0.0% 2.92, p = .09 Retardation/Agitation 14.2% 3.3% 2.82, p = .09 Anxiety Symptoms 18.2% 33.3% 3.96, p = .05 Anxious/Stressed 11.6% 26.7% 5.50, p = .02 As can be seen in Table 20, Inner Pain, Feel Trapped, Anxiety Symptoms, and Anxious/Stressed were significant findings associated more with "Depressed" adolescents rather than "Not Depressed:" not found in the DSM-IV-TR (APA, 2000) criteria for MDE. Association between Adolescent Depression Constructs and Self-Recognition of Depression To examine the association between adolescent depression constructs and S-ROAD, x 2 tests were used to assess the association between adolescent depression constructs and two forms of self-recognition of depression that included S-ROAD in the past two weeks prior to the survey in both analyses. The rationale for including analyses for S-ROAD within a two week period is because this time frame is used as part of the diagnostic criteria for MDE in the DSM-IV-TR (APA, 2000). Univariate analysis of S-ROAD time recall, from within two weeks to more than 12 months, is presented initially followed by the x 2 test results. Results of the analyses for these two variables are presented in separate parts. One dichotomous S-ROAD variable consisted of those adolescents who self-recognized depression within the past two weeks prior to the survey and those who did not self-recognize depression. The second S-ROAD 108 Chapter IV: Results variable contained two values of self-recognition of depression within the past two weeks prior to the survey and self-recognition of depression beyond the two weeks. Univariate Analyses of Self-Recognition of Adolescent Depression To assess adolescents' self-recognition of depression in their lifetime, frequency counts were calculated for the S-ROAD item (Appendix F). Results indicated that more than half of adolescents (56%) identified themselves as having had depression sometime during their lifetime. To establish time recall of those participants who recognized depression in themselves (n = 185), frequency counts were computed for "Presently or up to two weeks" continuing with a range to lifetime S-ROAD (Appendix G1). Participants were able to provide more than one answer. From the study sample (N= 332), 31 (9%)18 adolescents considered themselves to have been depressed within the last two weeks prior to the survey. Thirteen out of 148 boys (8.8%) and 18 out of 184 girls (9.8%) comprised these 31 adolescents who self-recognized depression within the last two weeks prior to the survey. Most of the adolescents who self-recognized depression in their lifetime considered themselves depressed beyond the two weeks (n = 151, [83%]). Association between Adolescent Depression Constructs and Self-Recognized Depression within the Past Two Weeks versus Not Self-Recognized Depression Chi-square tests revealed significant differences in self-recognition of depression within two weeks prior to the survey versus those who did not recognize depression in themselves in association with categories Anxiety Symptoms and Contextual/Causal, and subcategory Fear (see Table 21). In all these three constructs, adolescents who recognized depression in themselves within the past two weeks prior to the survey had significantly higher percentages of COAD units than those participants who did not self-recognize depression. 18 The percentages of participants who recognized depression in themselves, within two weeks and beyond, were taken against the entire sample of 332 rather than only those who self-recognized depression, n = 185 (- 3 missing in this analysis) = 182, because the data already revealed that 143 (43%) did not self-recognize depression (missing n = 4). 109 Chapter IV: Results Table 21 Percentage Distribution, Significant, and Near Significant (< .10) Chi-Square Tests of Frequency in Adolescent Depression Constructs by Self-Recognition of Depression in Those Who Self-Recognized Depression Within two Weeks Prior to the Survey versus Those Who Did Not Self-Recognize Depression, N = 174 Category / Subcategory Did Not Self-^Self-Recognized Recognize Depression Depression, n = 143^(past 2 weeks), n = 31 oh^ % X2 (1, N = 174) Depressed Mood Irritability 32.2% 48.4% 2.94, p = .09 Indecisiveness/Impaired 14.7% 29.0% 3.68, p = .06 Concentration and Attention Social Impairment 67.1% 51.6% 2.68, p = .10 Aggression/Bullied 10.5% 0.0% 3.56, p = .06 Anxiety Symptoms 16.8% 32.3% 3.88, p = .05 Fear 5.6% 16.1% 4.09, p = .04 Contextual/Causal 19.6% 35.5% 3.71, p = .05 Note. Self-Recognition of Depression is out of those who recognized depression in themselves within the past 2 weeks prior to the survey, n = 31, and those adolescents who did not self-recognize depression, n = 143. Remaining Sample, n= 151 self-recognized depression beyond 2 weeks, missing, n = 7. As can be seen in Table 21, the constructs Anxiety Symptoms, Fear, and Contextual/Causal had significant findings. In association with Social Impairment and all four of its subcategories, adolescents who did not self-recognize depression had substantially higher percentages of COAD units in these constructs than those who did self-recognize depression within the past two weeks (although not significant) (see Table E1-8 in Appendix El for complete results). Regarding subcategory Aggression/Bullied, 0% of the self-recognized depression group of adolescents provided a COAD unit in this subcategory compared to 10.5% of those students who did not self-recognize depression (not significant). 110 Chapter IV: Results Association between Adolescent Depression Constructs and Self-Recognition of Depression within the Past Two Weeks versus Those Who Self-Recognized Depression Beyond the Two Weeks Table 22 lists the percentages of adolescent depression constructs by self-recognition of depression within versus beyond two weeks prior to the survey. Chi-square tests revealed that adolescents who recognized depression in themselves within the past two weeks had a significantly higher percentage of COAD units than those students who self-recognized depression beyond the two weeks in the category Indecisiveness/Impaired Concentration and Attention, x2 (1, N= 182) = 5.96,p = .02. In association with Social Impairment, the ones who self-recognized depression beyond the two weeks had a significantly higher percentage of COAD units in this category than those adolescents who self-recognized depression within two weeks, x2 (1, N = 182) = 5.44,p = .02 (see Table F1-9 in Appendix Fl for complete results). In all four subcategories of Social Impairment, adolescents who self-recognized depression beyond two weeks prior to the survey, had higher percentages of COAD units in these areas than those who did self-recognize depression within the past two weeks (not significant). 111 Chapter IV: Results Table 22 Percentage Distribution, Significant, and Near Significant (< .10) Chi-Square Tests of Frequency in Adolescent Depression Constructs by Self-Recognition Depression in Those Who Self-Recognized Depression Beyond Two Weeks Prior to the Survey versus Those Who Self- Recognized Depression within the Past Two Weeks, N = 182 Category/ Subcategory Self-Recognized Depression (beyond 2 weeks), n = 151 Self-Recognized Depression (within past 2 weeks), n = 31 x2 (1, N= 182) Depressed Mood Unmotivated 18.5% 6.5% 2.73, p = .10 Indecisiveness/Impaired 11.9% 29.0% 5.96, p = .02 Concentration and Attention Social Impairment 72.8% 51.6% 5.44, p = .02 Lonely 39.1% 22.6% 3.03, p = .08 Aggression/Bullied 7.9% 0.0% 2.64, p = .10 Note. Self-Recognition of Depression is out of those who recognized depression in themselves within the past 2 weeks prior to the survey, n = 31, and those adolescents who self-recognized depression in their lifetime (beyond 2 weeks), n = 151. Remaining Sample comprised n = 143 who did not self-recognize depression, missing, n = 7. Summary A significantly higher percentage of adolescents who self-recognized depression within two weeks prior to the survey than those who did not self-recognize depression, provide their COAD in the constructs, Anxiety Symptoms, Contextual/Causal, and Fear. Adolescents who self-recognized depression within two weeks prior to the survey had a significantly higher percentage of COAD units than those who self-recognized depression beyond two weeks in the category Indecisiveness/Impaired Concentration and Attention. In contrast, adolescents who did not self-recognize depression beyond two weeks had a significantly higher percentage of COAD units than those who recognized depression in themselves within two weeks, in the category Social Impairment. More than 80% of the students in both sets of self-recognition groups generated COAD units in the category of Depressed Mood (see Table E1-8 in Appendix E 1, and Table F 1-9 in Appendix Fl for complete results). More than 50% of the participants, in both sets of self- 112 Chapter IV: Results recognition groups, generated COAD units in the Social Impairment category and in the subcategory of Sadness. Subcategories Irritability and Behavioral Disconnection contained over 30% of the students' COAD units in both sets of groups. These high percentages in both self- recognition groups of adolescents suggest that these constructs are representative dimensions of the adolescent depression concept as understood by these participants. Association between Adolescent Depression Constructs and Thinking About Own Need to Talk to Someone when Depressed To examine adolescents' own need to talk to someone when they have ever thought themselves to be depressed by constructs of adolescent depression, x2 tests were conducted using one item (Appendix G) developed for this study to examine participants' self-assessed need to talk to someone. Chi-square tests revealed that significant differences in percentages of COAD units were found between those adolescents who thought that they needed to talk to someone versus those participants who did not think that they needed to talk to someone in the category Feel Trapped , X2 (1, N= 233) = 4.92, p = .03 (see Table 23). More than 80% of the students in both groups mentioned COAD units in the category Depressed Mood, and more than 60% in the category Social Impairment (see Table H1-10 in Appendix H1 for complete Table results). More than 50% of the participants in both groups generated COAD units in the subcategory Sadness. Subcategories Irritability, Behavioral Disconnection, and Lonely contained over 30% of the students' COAD units in both groups. 113 Chapter IV: Results Table 23 Percentage Distribution, Significant, and Near Significant (< .10) Chi-Square Tests of Frequency in Association between Adolescent Depression Constructs and Thinking About Own Need to Talk to Someone When Depressed, N = 233 Category / Subcategory Did Not Think Needed to Talk to Someone, n = 85 Thought Needed to Talk to Someone, n = 148 X2 (1, N = 233) 0/0 Depressed Mood Moody 4.7% 11.5% 3.03, p = .08 Pessimism 2.4% 7.4% 2.64, p = .10 Feel Trapped 1.2% 8.1% 4.92, p = .03 Contextual/Causal 29.4% 19.6% 2.92, p = .09 Note. 233 (out of N = 332) participants responded to the values, "Yes" and "No" in this item, 94 responses applied to "never been depressed" and "did not self-recognize depression," missing, n = 5. As can be seen in Table 23 there was one significant finding in the category Feel Trapped. Those adolescents who thought that they needed to talk to someone when depressed had a higher percentage of COAD units in this construct than those who did not think that they needed to talk to someone. Association between Adolescent Depression Constructs and Talking to Someone at the Time of Feeling Depressed To examine the association between adolescent depression constructs and talking to someone at the time of feeling depressed, percentages and x2 tests were calculated using one item developed for this study to examine whether participants talked to someone at the time of feeling depressed (Appendix H). Table 24 lists the percentage distribution and x2 tests results. Chi-square tests revealed that adolescent depression constructs differed significantly only in the subcategory Hopelessness. A significantly higher percentage of adolescents who did not talk to someone at the time of feeling depressed (25.5%) versus those students who did talk to someone (12.0%) mentioned COAD units in this area, x2 (1, N = 180) = 5.12, p = .02. 114 Chapter IV: Results More than 80% of the students, in both groups, mentioned COAD units in the category of Depressed Mood and more than 60% in the category Social Impairment. Subcategories Irritability and Behavioral Disconnection contained over 30% of the students' COAD units in both groups (see Table I1-11 in Appendix Il for complete results). Table 24 Percentages and Chi-Square Tests of Frequency in Adolescent Depression Constructs by Talking to Someone at the Time of Feeling Depressed, N = 180 Category / Subcategory Did Not Talk to Someone^Talked to Someone x 2 (1, N = 180) Depressed Mood Hopelessness ^ 25.5%^ 12.0%^5.12, p = .02 Note. 180 (out of N= 332) participants responded to this item "Talking to Someone at the time of feeling depressed," 118 did not self-recognize depression and did not respond to this item, 34 missing. As can be seen in Table 24, one construct, Hopelessness, was a significant finding associated with a higher percentage of COAD in the group of adolescents who did not talk to someone versus those who did talk to someone. Summary Table 25 presents the significant x2 tests results of the associations between adolescent depression constructs and the various variables discussed in this section (see Table J1-12 in Appendix J1 for complete x2 test results). Only the significant results are presented in this summary. Depressed Mood and its Subcategories In association with Depressed Mood, grade 8 boys had a higher percentage of COAD units in this category than grade 11 boys. A higher percentage of 8 th graders than 1 1 th graders provided COAD units in its subcategory Inner Pain, as well as a higher percentage of boys in grade 8 than grade 11 defined units in this area. Further, in association with Inner Pain, those adolescents screened as "Depressed" had a higher percentage of COAD units than those students screened as "Not Depressed." Regarding Depressed Appearance, higher percentages of 115 Chapter IV: Results adolescents providing COAD units in this construct were in: grade 8 versus in grade 11; girls versus boys; girls versus boys in grade 11; 8 th grade boys versus 11 th grade boys; and girls in grade 8 versus girls in grade 11. The subcategory Hopelessness had one significant x 2 test finding in that those adolescents who did not talk to someone at the time of feeling depressed had a higher percentage of COAD units than the group of students who did talk to someone at that time. In association between Pessimism and grade level, 8 th graders had a higher percentage of COAD units than the 11 th graders in this area. Social Impairment and its Subcategories Concerning the Social Impairment construct, girls more than boys, grade 11 girls more than grade 11 boys, and those students who self-recognized depression beyond a two-week period more than the group of adolescents who self-recognized depression within two weeks, mentioned this category. In association with its subcategory Perceived Disconnection, girls more than boys, and more grade 8 girls than grade 8 boys, provided COAD units in this subcategory. In association with the Aggression/Bullied subcategory, higher percentages of girls than boys, and 8th grade girls than 8 th grade boys, mentioned COAD units in this construct. Remaining Significant Categories and Subcategories In association with the category Low Self-Worth, girls more than boys, and girls more than boys in grade 8, had higher percentages of COAD units in this construct. Regarding the category Indecisiveness/Impaired Concentration and Attention, those adolescents who self- recognized depression within two weeks had a higher percentage of COAD units in this area than the group that self-recognized depression beyond the two-week time frame. In association with the category Feel Trapped, higher percentages were found in the group of adolescents screened as "Depressed" versus those screened as "Not Depressed," and those adolescents who thought of needing to talk to someone versus those who did not think that they needed to talk to someone. In association with the category Suicidal, a higher percentage of 8 th graders versus 11 th graders, and girls versus boys, provided COAD units in this construct. Regarding the Retardation/Agitation category, higher percentages in girls versus boys, and in grade 11 girls versus boys in grade 11, mentioned COAD units in this area. Its subcategory Quiet, which comprised 94% of this category, had significant x2 test results with the 116 Chapter IV: Results same variables. In association with the Anxiety Symptoms category, those adolescents screened as "Depressed" had a higher percentage of COAD than those screened as "Not Depressed." Additionally, a higher percentage of students who self-recognized depression within two weeks versus those who did not self-recognize depression, provided COAD units in the Anxiety Symptoms category. In association with the subcategory Fear, adolescents who self-recognized depression within two weeks had a higher percentage of COAD than those who did not self- recognize depression. A greater percentage of adolescents screened as "Depressed" versus those who were "Not Depressed," provided COAD units in the subcategory Anxious/Stressed. In association with the Contextual/Causal category, a higher percentage of girls in grade 8 versus grade 11 girls; and a higher percentage of adolescents who self-recognized depression within two weeks than those who did not self-recognize depression, had COAD units in this category. Regarding the Masking/Stigma category, more girls than boys provided COAD units in this area. 117 Chapter IV: Results Table 25 Significant Chi-Square Tests of Frequency in Adolescent Depression Constructs by Grade, Gender, Gender Differences in Grade 8 and in Grade 11, Grade Level Differences in Boys and in Girls, Self-Recognition of Depression, Thought about Needing to Talk to Someone, and Talking to Someone When Feeling Depressed Category / Subcategory Grade^Gen.* X 2^x 2 N=332 N=332 ^Gen.^Gen. in^Grade^Grade in Gr. 8^Gr. 11^in Boys^in GirlsxX 2 2 X2^X2 N= 114 N= 218 N= 148 N= 184 D.S. 1 X 2 N= 332 Self-Rec 2^Self-Rec 3 None vs.^2 weeks within 2^Within vs. ^ wks.^Beyond ^ 2^x 2 N= 174^N= 182 Thought Needed to^Talk Talk^X2 X2 N= 180 N= 233 Depressed^—^—^—^—^4.67, Mood p= .03 Sadness^— Irritability — Moody^— - - Inner Pain^4.67,^-^-^-^9.13,^-^11.20, p= .03 p= .00 p= .00 Depressed^11.38,^8.21,^-^4.94,^4.64,^6.45, Appearance^p= .00^p= .00 p= .03^p= .03^p= .01 - - - - Hopeless —^—^—^—^—^—^—^—^—^—^5.12, p= .02 Unmotivated^— Pessimism^3.74, p= .05 Appetite/Weight/ Sleep Changes Chapter IV: Results Category / Subcategory Grade^Gen.* X2^x2 N=332 N=332 Gen. in Gr. 8 X2 N= 114 Gen. in^Grade^Grade Gr. 11^in Boys^in Girls X2^x2^x2 N= 218 N= 148 N= 184 D.S. 1 X 2 N=332 Self-Rec 2 None vs. within 2 wks. X 2 N= 174 Self-Rec 3 2 weeks Within vs. Beyond x 2 N= 182 Thought Needed to^Talk Talk^X2 X 2 N= 180 N= 233 Fatigue/Loss of Energy Low Self- Worth/Guilt Indec./Impaired Concentration Feel Trapped Suicidal Social Impairment Behavioral Disconnection Perceived Disconnection Lonely Aggression/ Bullied Functional Impairment/Distress Retardation/ Agitation - 7.33,^9.96, p= .01^p= .00 - 5.96, p= .02 ^ 7.32,^-^-^4.92, ^p= .01 p= .03 5.68, p= .02 9.05,^-^8.28,^-^-^-^-^5.44, p= .00 p= .00 p= .02 - 8.23,^5.89, p= .00^p= .02 - 4.14,^6.28, p= .04^p= .01 - 6.15,^-^8.54, p= .01 p= 00 5.60, p= .02 Chapter IV: Results Category / Subcategory Grade^Gen.* X2^x 2 N= 332 111= 332 Gen.^Gen. in^Grade^Grade in Gr. 8^Gr. 11^in Boys^in Girls X 2 X2 X 2 X2 N= 114 N= 218 N= 148 N= 184 D.S.1 X 2 N= 332 Self-Rec2 None vs. within 2 wks. X 2 N= 174 Self-Rec 3 2 weeks Within vs. Beyond X 2 N= 182 Thought Needed to^Talk Talk^X 2 X 2 N= 180 N= 233 Quiet Anxiety Symptoms Fear Anxious/Stressed Substance Use Contextual/Causal — 5.97,^—^8.44, p = .02 p = .00 - - - - - - - - — 3.96,^3.88, ^p = .05^p = .05 — —^4.09, p = .04 — 5.50, p = .02 4.44,^—^3.71, p = .04 p = .05 Masking/Stigma^—^5.90, p = .02 Note. * Gen. = Gender, (df = 1), — = not applicable. ' D.S. = Depressive Symptomatology 2 Self-recognition of depression within two weeks versus no self-recognition of depression groups. 3 Self-recognition of depression within two weeks versus self-recognition of depression beyond two weeks groups. Chapter IV: Results Association of Self-Recognition of Depression to Depressive Symptomatology, and to Thinking About Own Need to Talk and Talking to Someone at the Time of Feeling Depressed To address the third question concerned with examining the association of self- recognition of depression to depressive symptomatology, and to approach-oriented coping specifically via adolescents' thinking about own need to talk to someone and talking to someone at the time of feeling depressed, frequencies, percentages, and a x 2 tests were calculated at a significance level of .05. Gender frequencies are also sequentially presented in these two tables and one figure to illustrate these separate associations. To examine the first part of the third question, the association between self-recognition of depression and depressive symptomatology, self-recognition of depression was separated into those adolescents who self-recognized depression in the past two weeks prior to the survey and those who did not recognize depression in themselves. The two-week time recall was used to reflect the same time frame used as a criterion for depression in the DSM-IV-TR (APA, 2000). Table 26 displays the findings of this association between self-recognition of depression and depressive symptomatology. Thirty adolescents were screened as "Depressed" using the RADS- 2 (Reynolds, 2002): 7 (5%) boys and 23 (13%) girls. In the past two weeks prior to the survey 31 adolescents self-recognized depression. Out of the 31 adolescents who self-recognized depression in the past two weeks, 14 (45%) adolescents were screened as "Depressed" and 17 (55%) participants as "Not Depressed." A higher percentage (46.7%) of "Depressed" adolescents versus "Not Depressed" participants (5.7%) self-recognized depression, x 2(1, N= 328) = 53.44, p = .00. There were 13 (9%) boys and 18 (10%) girls who self-recognized depression in the past two weeks. Seventy-one percent of the seven boys screened as "Depressed," self-recognized depression in the past two weeks prior to the survey. In contrast, 39% of the 23 girls screened as "Depressed" self-recognized depression in the past two weeks prior to the survey. 121 Chapter IV: Results Table 26 Association between Self-recognition of Depression Within the Past two Weeks and Depressive Symptomatology Self-Recognized Depression PAST 2 WEEKS Depressive Symptomatology Not Depressed^Depressed (RADS-2 < 75) (RADS-2 > 76) Totals NO 281 (94.3%) 131 boys, 150 girls 16 (53.3%) 2 boys, 14 girls 297 (90.5%) 133 boys, 164 girls YES 17 (5.7%) 8 boys, 9 girls 14 (46.7%) 5 boys, 9 girls 31 (9.5%) 13 boys, 18 girls Totals 298 139 boys, 159 girls 30 7 boys, 23 girls 328 146 boys, 182 girls Note. Percentages calculated down the columns, N= 332 (-4 missing) = 328. As can be seen in Table 26, only about half (45.2%) of the 31 adolescents who self- recognized depression in the past two weeks prior to the survey were screened as "Depressed." Table K1-13 in Appendix K1 provides an overview of the association of lifetime self-recognition of depression to depressive symptomatology illustrating that 15 adolescents screened as "Depressed," self-recognized depression beyond the two weeks prior to the survey, and 1 adolescent screened as "Depressed" did not self-recognize depression. To calculate the differences in those adolescents who self-recognized and did not self- recognize depression for the RADS-2 (Reynolds, 2002) Depression Total Score means, the 30 adolescents screened as "Depressed" with a score of greater or equal to 76 were removed from the analysis, and a t test was conducted for the remaining participants. For the means of the Depression Total Score, those adolescents who self-recognized depression (M= 57.63, SD = 9.39) had a significantly higher mean on the RADS-2 (Reynolds, 2002) Depression Total Score than those adolescents who did not self-recognize depression (M= 49.01, SD = 9.30), t (296) = 7.95, p .000 (two-tailed). Further, to examine the differences between those adolescents who self-recognized depression and those who did not self-recognize depression for the RADS-2 (Reynolds, 2002) Depression Total Score, the medians and their range were calculated for the participants who 122 Chapter IV: Results were screened as "Not Depressed." For the medians of the Depression Total Score, those adolescents who self-recognized depression (Mdn = 58.50, range = 36 — 75) had a higher median that those participants who did not self-recognize depression (Mdn = 50.00, range = 31 — 74). Figure 2 displays the frequency findings of the analyses examining the second part of the third question, the association of lifetime self-recognized depression (two and beyond two weeks prior to the survey) to adolescents' thinking about own need to talk to someone and talking to someone at the time of feeling depressed. 123 Olt! NOT Talk to rrteorie Chapter IV: Results Figure 2 Self-recognized Depression, Thinking About Own Need to Talk to Someone when Depressed, and Talking to Someone when Feeling Depressed Self-Recognized Depression in Lifetime (PAST 2 WEEKS AND BEYOND) N = 185 71 boys, 114 girls Thought Needed to Talk to Someone  Thought NOT Needed to Talk to Someone n = 124/183(-2) (67.8%) 40 boys, 84 girls n = 59/183(-2) (32.2%) 29 boys, 30 girls Note. Sample size numbers vary depending on the results of the analyses. As can be seen in Figure 2, the majority (67.8%) of adolescents who self-recognized depression in their lifetime thought that they needed to talk to someone when depressed. Furthermore, the majority (78.5%) of adolescents who thought that they needed to talk to someone when depressed actually did talk to someone. Fifty-six percent of the boys (n = 40) and 74% of the girls (n = 84) who self-recognized depression in their lifetime, thought that they needed to talk to someone when depressed. When it came to talking to someone when feeling depressed, 70% of the boys (n = 28) and 80% of the 124 Chapter 1V: Results girls (n = 67), who thought that they needed to talk to someone, actually did. Chi-square test revealed that a higher percentage of those adolescents who self-recognized depression and thought that they needed to talk to someone when depressed, talked to someone versus those who did not talk to someone, x 2(1, N 149) = 33.93, p = .00. In differentiating the group of adolescents who talked and those who did not talk to someone when feeling depressed by depressive symptomatology using the RADS-2 (Reynolds, 2002), there were no "Depressed" adolescents in the group who did not self-recognize depression. The 24 (6 missing) screened as "Depressed" adolescents were all in the group who self-recognized depression, and 54% of these "Depressed" adolescents talked to someone versus 72% who were "Not Depressed." In the "Depressed" group of adolescents who talked to someone, 50% of the 6 boys and 56% of the 18 girls talked to someone. Relations of Social Perspective Coordination, Emotion Awareness, Expression of Emotion, and Emotion Regulation, to Adolescents' Dimensions of Depressive Symptomatology To address the fourth question concerned with examining the relations of social perspective coordination, emotion awareness, expression of emotion, and emotion regulation, to adolescents' dimensions of depressive symptomatology, the data were first screened for violations of assumption. Next, correlational analyses were done to measure the associations of social perspective coordination, emotion awareness and expression, and emotion regulation, to dimensions of depressive symptomatology. Finally, linear hierarchical regression analysis was conducted to explore the influence of social perspective coordination, emotion awareness, expression of emotion, and emotion regulation, to dimensions of adolescents' depressive symptomatology. In the initial part of this fourth section preliminary analysis to screen the data is presented. Next, a presentation of the correlational analysis among the selected variables is reported. In the final part of this section, I present the results of hierarchical regression. 125 Chapter IV: Results Preliminary Analyses Prior to statistical analysis, the data were examined using the procedures outlined by Tabachnick and Fidell (2001) for screening data. Results from the variables implemented in the regression analyses were screened for violations of assumptions using the procedures suggested by Tabachnick and Fidell (2001) and de Vaus (2002). In developing the various measures, participants were included if they had answered at least 80% of items for each scale. Less than 5% of the participants were missing any data points (Tabachnick & Fidell, 2001) in this study. SPSS (Version 13) Frequencies histograms were further used to check on accuracy for the normality of distributions, and their shapes were examined using SPSS Graphs (Version 13). The means (Al), minimum and maximum values, standard deviations (SD), skewness, kurtosis, and range of the scores, were inspected for plausibility. Both visual inspection and the coefficients for skewness and kurtosis were examined to detect departures from normality (de Vaus, 2002; Stevens, 1992; Tabachnick & Fidell, 2001). Skewness values were found to be within the acceptable range (Tabachnick & Fidell, 2001). The kurtosis on social perspective coordination, the 4+ Rel-Q, had a minor deviation from normality (see Table 27). However, the impact of skewness and kurtosis diminishes for larger samples of 100 or more (de Vaus, 2002; Tabachnick and Fidell, 2001) or 200+ (Waternaux, 1976), leading to the conclusion that the data set could be used without transforming the scores on this measure to obtain more normal distributions. Upon visual inspection, the distributions of the emotion regulation scales: Anger Management: Coping; Anger Dysregulated-Expression; Sadness Management: Inhibition; and Sadness Dysregulated-Expression, showed a departure from normal distribution (Tabachnick & Fidell, 2001). These variables were not retained in multivariate analyses to help avoid Type I error (see also Methods Chapter for reliabilities of scales). The emotion regulation scales that were kept in multivariate analyses were the Anger Management: Inhibition (a = .79) and the Sadness Management: Coping (a = .53). Although the reliability for the Sadness Management: Coping was low, this scale remained in multivariate analyses as it was the only sadness regulation scale that did not depart from normal distribution, and reliability of a = .50 on emotion scales have been used in previous studies (J. L Zeman, personal communication, March 6, 2007). The scales: depressive symptomatology; social perspective coordination; poor emotion awareness; expressive reluctance; sadness management coping; and anger management 126 Chapter IV: Results inhibition; were considered acceptable for this study. Table 27 includes the means, standard deviations, skewness, kurtosis, and range for the measures used in regression analysis. Table 27 Descriptives of Measures used in Regression analyses, N = 332 Variable M SD Skewness Kurtosis Range Depressive Symptomatology (RADS-2) 56.27 13.30 .65 .46 31-103 Social Perspective Coordination (4+Rel-Q) 2.17 .14 -.85 2.21 1.44-2.50 Poor Emotion Awareness (EESC) 2.00 .78 .93 .32 1.00-4.50 Expressive Reluctance (EESC) 2.44 .84 .45 -.53 1.00-4.75 Sadness Management: Coping (CSMS) 2.16 .40 -.18 -.34 1.00-3.00 Anger Management: Inhibition (CAMS) 1.74 .51 .41 -.45 1.00-3.00 Box plots were used to identify univariate outliers in the measures that identify social perspective coordination (4+Rel-Q), depressive symptomatology (RADS-2), emotion awareness (EESC scale) and expression of emotion (EESC scale), and sadness (CSMS scale) and anger regulation (CAMS scale). Out-of-range values were checked against the original data for accuracy of input. Upon visual inspection of the distributions, one univariate outlier was identified in the 4+Rel-Q, but it disappeared upon examination for multivariate outliers and hence did not affect multivariate distribution. As noted by Tabachnick and Fidell (2001) and others (Stevens, 1992), in samples with more than 100 cases, a few outliers can be expected by chance. This one case was retained for all multivariate analyses. Using Mahalanobis distance with p < .001 (Tabachnick & Fidell, 2001) scatterplot, three multivariate outliers were identified. Mahalanobis distance was also calculated as chi-square with degrees of freedom equal to the number of observed variables in the hypothesized model (Tabachnick & Fidell, 2001). For the present study, which implemented seven independent variables for the regression analysis, the X2 critical value at p <.001 was 24.32. Thus if the Mahalanobis distance for a single case exceeded 24.32, it was considered an outlier. The three cases in this data set were identified as multivariate outliers having Mahalanobis distance values as 36.82, 27.35, and 24.86. When using Cook's distance in a pairwise plot (Fox, 1991; Tabachnick & Fidell, 2001), to establish leverage and discrepancy influence on the regression, 127 Chapter IV: Results two multivariate outliers were identified. These two outliers were the same two out of the three identified using Mahalanobis distance. Cohen and Cohen (1983) suggest that if outliers represent less than one or two percent of the total number of cases, they can remain in the data set. Hence, the three multivariate outliers were retained in the multivariate analyses. A bi-scatterplot was performed between each independent variable against the dependent variable depressive symptomatology, and the linearity was reasonable and not homoscedastic (Tabachnick & Fidell, 2001). The assumptions have been met. Although Poor Emotion Awareness and Expressive Reluctance scales were strongly correlated (r = .67), they were still implemented as separate independent variables in regression analysis because of their theoretical distinctions and their differences in measuring emotion objectives (see Self-Awareness of Emotion and Expression of Emotion section in Methodology Chapter; J. L. Zeman, personal communication, March 6, 2007). According to Tabachnick and Fidell (2001), careful thought to using correlated variables is advised for those with .70 and above. There are also differences in thought as to the general rule in acceptance of any variable that could indicate problems with multicollinearity. According to de Vaus (2002), any variable with a Tolerance of .2 or less or a Variance Inflation Factor (VIF) of 5 or more is deemed problematic within regression procedures. Others (Tabachnick and Fidell, 1996) note that a variable with a Tolerance of .8 to 1 and the VIF between 1 and 3 are the more appropriate choices. The Tolerance of these independent variables implemented in regression analysis were .48 and .41 respectively, and the VIF were 2.10 and 2.4 respectively. No multicollinearity was evident with the remaining variables implemented in regression analysis. Correlational Analyses A series of zero-order correlations were used to measure the associations of social perspective coordination, emotion understanding and expression, and emotion regulation, to dimensions of adolescents' depressive symptomatology. The results of the correlational analysis presented in Table 28 demonstrate that several significant relationships were found. The dimension of depressive symptomatology was strongly and positively associated with Poor Emotion Awareness and Expressive Reluctance of Emotion to others. To a lesser but significant positive correlation, the dimension of depressive symptomatology was positively related to inhibition of Anger Expression. In contrast, depressive symptomatology was negatively 128 Chapter IV: Results correlated with Social Perspective Coordination and coping with sadness regulation. Social Perspective Coordination was significantly and negatively correlated with Poor Emotion Awareness and Emotion Expressive Reluctance. Poor Emotion Awareness was strongly positively correlated with Expressive Reluctance (see Preliminary Analysis) and moderately with inhibition of anger regulation. A significant negative correlation was found between Poor Emotion Awareness and strategies for coping with sadness. Expressive Reluctance of Emotion was strongly positively correlated with Anger Regulation Inhibition and less so correlated with Sadness Regulation Coping. Anger Regulation Inhibition was positively correlated with coping with sadness regulation. Taken together, these findings generally reflect that with severity of depressive symptomatology, adolescents have increasing difficulty in identifying internal emotional experiences, or expression of emotions to others, or suppressing anger expression. Greater social-cognitive maturity or coping with sadness regulation are negatively correlated with increasing levels of depressive symptomatology. Table 28 Intercorrelations among Social Perspective Coordination, Emotion Measures, and Depressive Symptomatology Variable M SD 1 2 3 4 5 1 Depressive Symptomatology 2 Social Perspective Coordination 3 Poor Emotion Awareness 4 Expressive Reluctance 5 Anger Regulation Inhibition 6 Sadness Regulation Coping 56.27 2.17 2.00 2.44 1.74 2.16 13.30 .14 .78 .84 .51 .40 -.14* .63** .50** .20** -.20** ____ -.13 * -.18** .04 .06 .67** .27** -.15** _ .45** .12* .23** Note. * p < .05 (2-tailed). ** p < .01 (2-tailed). Regression Analysis In order to further assess the links among these variables, linear hierarchical regression analysis was conducted to explore the influence of Social Perspective Coordination, Emotion Awareness, Expression of Emotion, and emotion regulation, to dimensions of adolescents' 129 Chapter IV: Results depressive symptomatology. To control for individual demographic differences (Carey, Carey, & Kelley, 1997; Mazza & Reynolds, 1999; Ozer & McDonald, 2006), grade and gender were entered into the analysis as a block on the first step. Second, to differentiate between social- cognitive maturity and emotion factors because of their negative correlation with each other and with dimension of adolescents' depressive symptomatology, only scores from the 4+Rel-Q were entered as a block on this step. Finally, scores from the emotion measures were entered as a block on the third step. The RADS-2 (Reynolds, 2002) scores were entered as the dependent variable in this regression analysis. The betas, t values, sig. t, R 2, and AR 2 values are presented in Table 29. Table 29 Summary of the Hierarchical Regression Analysis for Social Perspective Coordination, Emotion Awareness, Emotion Expression, and Emotion Regulation, in Predicting Depressive Symptomatology, N = 320 Variable 13 t Sig. t R2 AR2 Step 1 Grade .017 .305 .760 .026 .026 Gender .162 2.929 .004 Step 2 Grade .023 .423 .673 .055 .028 Gender .190 3.421 .001 Social Perspective Coordination - .171 - 3.082 .002 Step 3 Grade .074 1.747 .082 .462 .407 Gender .140 3.089 .002** Social Perspective Coordination - .057 - 1.312 .191 Poor Emotion Awareness .458 7.610 .000** Expressive Reluctance .230 3.536 .000** Anger Regulation: Inhibition - .003 - .071 .943 Sadness Regulation: Coping - .110 - 2.407 .017* Note. * p < .05. ** p < .01. The results of the hierarchical regression analysis indicated that the combined influence of grade and gender at the first step, accounted for 2.6% of the variance in adolescents' dimensions of depressive symptomatology, F(2, 317) = 4.312, p = .014. In the second step, the 130 Chapter IV: Results model now with three predictor variables that contained grade, gender, and Social Perspective Coordination, accounted for 5.5% of the variance in adolescents' dimensions of depressive symptomatology, F(3, 316) = 6.12, p = .000. The unique contribution of Social Perspective Coordination was 2.8% and was statistically significant (AR 2 = .028, p = .002). In the third step, the model now with seven predictor variables, accounted for 46.2% of the variance in adolescents' dimensions of depressive symptomatology, F(7, 312) = 38.225,p = .000. The unique contribution of the four emotion variables, Poor Emotion Awareness, Expressive Reluctance of Emotion, Anger Regulation Inhibition, and Sadness Regulation Coping, was statistically significant, AR 2 = .407, p = .000. Hierarchical Regression Analysis for Boys Only To assess for any distinctions in gender regarding the influence of the independent variables on predicting adolescents' dimensions of depressive symptomatology, hierarchical regression analysis was done separately for boys and for girls. The rationale for separating the analysis by gender rather than interactions was for ease of interpreting the findings. The betas, t values, sig. t, R 2, and AR 2 values are presented in Table 30 for boys only. 131 Chapter IV: Results Table 30 Summary of the Hierarchical Regression Analysis for Social Perspective Coordination, Emotion Awareness, Emotion Expression, and Emotion Regulation, in Predicting Depressive Symptomatology in Boys Only, N = 142 Variable 13 t Sig. t R2 AR2 Step 1 Grade - .101 - 1.196 .234 .010 .010 Step 2 Grade - .096 - 1.149 .252 .026 .016 Social Perspective Coordination - .128 - 1.528 .129 Step 3 Grade .041 .605 .546 .418 .391 Social Perspective Coordination - .109 - 1.625 .107 Poor Emotion Awareness .446 5.055 .000** Expressive Reluctance .259 2.763 .007** Anger Management: Inhibition - .011 -^.152 .880 Sadness Management: Coping - .037 -^.522 .603 Note. * p < .05. ** p < .01. The results of the regression analysis indicated that grade at the first step accounted for 1.0% of the variance in adolescents' dimensions of depressive symptomatology, F(1, 140) = 1.430, p = .234 (nonsignificant). In the second step, the model now with two predictor variables, grade and Social Perspective Coordination, accounted for 2.6% of the variance in adolescents' dimensions of depressive symptomatology, F(2, 139) = 1.889, p = .155 (nonsignificant). The unique contribution of Social Perspective Coordination was 1.6% and was statistically nonsignificant (AR2 = .016, p = .129). In the third step, the model now with six predictor variables, accounted for 41.8% of the variance in boys' dimensions of depressive symptomatology, F(6, 135) = 16.128,p = .000. The unique contribution of the four emotion variables, Poor Emotion Awareness, Expressive Reluctance of Emotion, Anger Regulation Inhibition, and Sadness Regulation Coping, was statistically significant, AR 2 = .391, p = .000. 132 Chapter IV: Results Hierarchical Regression Analysis for Girls Only The betas, t values, sig. t, R 2, and Al? 2 values are presented in Table 31 for girls only. Table 31 Summary of the Hierarchical Regression Analysis for Social Perspective Coordination, Emotion Awareness, Emotion Expression, and Emotion Regulation, in Predicting Depressive Symptomatology in Girls Only, N=178 Variable (3 t Sig. t R2 AR2 Step 1 Grade .097 1.294 .197 .009 .009 Step 2 Grade .105 1.423 .157 .049 .040 Social Perspective Coordination - .200 - 2.709 .007 Step 3 Grade .088 1.552 .122 .476 .427 Social Perspective Coordination - .029 - .491 .624 Poor Emotion Awareness .480 5.584 .000** Expressive Reluctance .196 2.132 .034* Anger Management: Inhibition .004 .066 .948 Sadness Management: Coping - .151 - 2.577 .011* Note. * p < .05. ** p < .01. The results of the regression analysis indicated that grade at the first step accounted for 0.9% of the variance in adolescents' dimensions of depressive symptomatology, F(1, 176) =- 1.676, p = .197 (nonsignificant). In the second step, the model now with two predictor variables, grade and Social Perspective Coordination, accounted for 4.9% of the variance in adolescents' dimensions of depressive symptomatology, F(2, 175) = 4.538,p = .012. The unique contribution of Social Perspective Coordination was 4.0% and was statistically significant (AR 2 = .040, p = .007). In the third step, the model now with six predictor variables, accounted for 47.6% of the variance in girls' dimensions of depressive symptomatology, F(6, 171) = 25.913, p = .000. The unique contribution of the four emotion variables, Poor Emotion Awareness, Expressive 133 Chapter IV: Results Reluctance of Emotion, Anger Regulation Inhibition, and Sadness Regulation Coping, was statistically significant, AR2 = .427, p = .000. 134 Chapter V: Discussion CHAPTER V: DISCUSSION Depression: A Voice of an Adolescent in Grade 8 Depression. Such a simple word, yet it means so much...It's something small, but on the other hand --- big and powerful. It's when you're alone, and your whole world stops. Sometimes, it's good. You can look at your past and see your mistakes. But most of the time you can't. You try to correct and leave your problems behind. But they will follow you, like winter shadows. Every step you take, they'll be there. And you can't run away. It's when you fall, and there's no one to catch you. When there's no one you can lean on. When the only person you ever trusted, let you down. When your heart has been broken, stepped on, and finally thrown out the window. It's when you become a stranger, to your own reflection. When you just hope you were prettier or slimmer. When, even the best things about you start fading away. When you sit in the rain, without an umbrella, praying for sunshine. It's when you try to talk to someone, but they don't listen, and you have to keep it all inside. When you try to let it all out, but no one takes you seriously, because you've been such a goof before. It's when you go home every night and you cry until you just can't cry anymore. It's when you think "Just one more day, one more hour, one more moment...," and when the dark corner has been your favorite place to stay. It's when you are your only friend. The adolescent who wrote this poem slipped the neatly folded sheet that this poem was written on into my hand, after survey completion on the way out of the classroom. One of the aims of this study was to understand depression in adolescents from adolescents' own voices. This poem represents one of the voices. This study examined four research objectives to better communicate with young people, and to enable early detection and intervention of depression in adolescents. In order to extend the literature on adolescent depression, the first objective was to develop categories and subcategories of adolescent depression generated from the adolescents' own conceptions of depression. The aim of this objective was to clarify the adult criteria presently used to diagnose depression in adolescents from adolescents' own voices. 135 Chapter V: Discussion The second objective was to differentiate the developed constructs of adolescent depression by age, gender, depressive symptomatology, adolescents' self-recognition of depression, and the pathways to talking to someone when feeling depressed. The purpose of this exploration was to uncover the constructs of depression that are relevant to young people to improve communication with them, and to find clues to the early detection of depressive symptomatology. The third objective of this study was to examine the association of adolescents' self- recognition of depression to depressive symptomatology. The goal of this third objective was also to examine the pathways adolescents initiate to talking to someone once they self-recognize depression. The pathways include one's own thought of needing to talk to someone when depressed, and choosing to talk to someone when feeling depressed. This aim was to examine whether S-ROAD, or self-diagnosis of depression within a two week time frame, is meaningful or closely associated with screened presence of depressive symptomatology. Self-recognition of depression can be linked to early detection of depression and can enhance our understanding of adolescent depression. The fourth aim of this study was to examine how several factors, such as social perspective coordination, emotion awareness, expression of emotion, and emotion regulation, act as potential mechanisms in predicting depressive symptomatology. This knowledge can be important in detecting the precursors that can affect the development of depressive symptomatology. In turn, this information can be central in planning interventions suited to adolescents' needs. The present study extends and adds depth to the existing literature on these four objectives. Constructs of adolescent depression based on participants' COAD add new information and depth to the previous research and the diagnostic criteria (APA, 2000) for adolescent depression. The occurrence of the categories and subcategories generated by adolescents' COAD provides a map of the themes and subthemes that pervade adolescents' personal philosophies regarding adolescent depression. Various categories and subcategories were significant when considered with regards to the various characteristics (e.g., age, gender, depressive symptomatology). Less than half of the respondents (45%) self-recognized depression within the past two weeks prior to the survey when examining the association of self-recognition of depression to depressive symptomatology. When examining lifetime self-recognized 136 Chapter V: Discussion depression in connection to pathways to talking to someone at the time of feeling depressed (see Figure 2), most adolescents who self-recognized depression in their lifetime, thought that they needed to talk to someone (68%) when they have thought they were so depressed. Further, most adolescents, when they thought that they needed to talk to someone, talked to someone (79%) when feeling depressed. Finally, emotion understanding, particularly Poor Emotion Awareness, reluctance to express emotion, decreasing levels of coping with sadness, and gender are important factors in predicting the severity of depressive symptomatology. Elaborations of these findings are presented in the following six sections. The first four sections are presented in sequence with the four research questions around which this study is organized: (1) the construction, generation, frequencies, and discussion of the descriptives of the categories and subcategories generated by adolescents' conceptions of adolescent depression (COAD); (2) differentiating specific categories and subcategories, generated from adolescents' COAD in connection with grade, gender, depressive symptomatology, S-ROAD, thinking about own need to talk to someone, and talking to someone when feeling depressed; (3) distinguishing the associations of S-ROAD to the presence and non-presence of screened depression, as well as to thinking about own need to talk and to talking to someone when depressed; and (4) discussing how Social Perspective Coordination and emotion capacities relate to the severity of depressive symptomatology. In the fifth section, I discuss the strengths and limitations of this study. In the sixth section, I consider the implications and importance of the findings to research, policy, and practice, and offer further considerations and suggestions for future research. What are Adolescents' Conceptions of Depression? Construction of Adolescents' COAD In the present study, the categories and subcategories were developed from participant- generated COAD using content analysis, and predominantly guided by the DSM-IV-TR (APA, 2000) criteria for depression and infrequently by K-SADS-PL. These developed constructs of adolescents' COAD clarify, bridge, and extend prior research on adolescent depression in important ways. The interest centers on the adolescents' COAD to gain greater understanding of adolescent depression which may be a common path for other problems such as anxiety, increases in health risks, social and functional problems, and possible suicidality. 137 Chapter V: Discussion Because adolescents were asked to define "depression" and not "clinical or threshold" depression, their COAD encompassed a holistic approach to describing depression. The categories and subcategories generated in this study echo research findings that point to dimensional (Goldberg, 2000; Slade & Andrews, 2005), categorical DSM-IV-TR (APA, 2000), and new constructs (e.g., Inner Pain). This finding may also reflect depressive symptomatology of a MDE, where the condition can develop over weeks and may last for weeks to months before the onset of a full MDE episode (DSM-IV-TR; APA, 2000). The constructs based on adolescents' COAD included appearances, behaviors, cognitions, emotions, social relationships, and abstract states that seem to be informed by Piaget's theory of cognitive development. Piaget suggests that children attain objectivity when they become aware of their own subjectivity, and once this relationship is properly understood, they can apply this knowledge to relations existing between entities and the environment (Chapman, 1988). Generation, Distribution, and Range of COAD Units Grade Distribution of COAD Units In documenting grade differences in generated frequencies of adolescents' conceptions of depression, although the 8 th graders provided more COAD units than the 11 th graders, the 8 th graders provided a narrower range of COAD units than the 11 th graders representing the different constructs. This wider range generated by grade 11 participants underscores findings from previous literature and from Piaget's fourth and final formal-operational stage (adolescence to adulthood) of development. Conceptions, or the number of possible meanings, can vary from younger to older adolescents (Chapman, 1988). At this higher equilibrated stage, the ability to coordinate more perspectives or possible solutions (Chapman, 1988; Noam et al., 1995), or coping strategies expand. The older students' wider range of different conceptions can also be explained by findings from a study on 483 fifth to ninth graders (ages 9 to 15) by Larson and Ham (1993). These researchers reported that older students, both boys and girls, encountered more negative life events than younger ones, and that experience was more strongly associated with daily negative affect among these older students, indicating that older students experienced and 138 Chapter V: Discussion reported more mildly negative states daily (Larson & Lampman-Petraitis, 1989). Keating (1990) suggests that from early adolescence, cognition tends to involve abstract, self-reflective, self- aware, and multidimensional interpretations. This has been shown in adolescents' range of COAD in both grades. Gender Distribution of COAD Units Girls provided significantly more definitions of depression than boys and contributed a greater range of different conceptions of depression. These findings reflect gender comparisons where adolescent girls more than boys report more adversity in their lives (Goldberg & Goodyer, 2005; Goodyer, Tamplin, & Altham, 2000), are more self-reflective and aware of inner states (Allgood-Merten et al., 1990), and are more likely than boys to identify emotional disorders in themselves (Chen et al., 1998). As puberty emerges, and girls experience puberty earlier than boys, the rate of depression increases and is twice as common for adolescent girls compared to boys (DSM-V-TR; APA, 2000). This suggests that the way girls perceive their environment and interpret it in relation to their personhood may be important (Goldberg & Goodyer, 2005). Further, some studies report that depressed adolescent males tend to exhibit more externalizing symptoms (observed behavior), whereas depressed adolescent females show more internalizing symptoms (subjective) (Frydenberg, 1997). However, Goldberg and Goodyer (2005) report that internalizing and externalizing disorders are only weakly correlated with one another (e.g., reflecting that it is possible to have an externalizing disorder, such as alcohol dependence, and also be depressed). Taken together, girls are expressing their experiences in their greater number and range of COAD. Amidst the tipped balance in the direction of girls' greater number and range than boys' contributions of different conceptions of depression, there were a couple of aberrations where the boys generated COADs that produced different findings. Boys in grade 8 provided significantly higher percentages than boys in grade 11, of conceptions of depression in Inner Pain, and higher percentages than girls in both grade levels. This finding may be a window to boys' willingness to express themselves in terms of depression more readily at lower grade level than at a higher one. Hopelessness was another construct that was important to boys' definitions of depression at both grade levels: boys generated higher COADs than girls in grade 8 and in grade 11. Although assumptions cannot be made as to the meaning and importance of the constructs Inner Pain and 139 Chapter V: Discussion Hopelessness, it seems that boys are expressing themselves differently to girls in terms of these subcategories of Depressed Mood that may also be captured at a younger grade level. Generated Constructs of Adolescent Depression Adolescents have a holistic understanding of depression based on the conceptions they report. In examining the type and frequency of adolescents' generated conceptions of depression, the present research revealed that more than half of their definitions of depression were contained within the two most prominent categories, Depressed Mood and Social Impairment. In the category Depressed Mood, the leading subcategories were Sadness and Irritability. In the category Social Impairment, at least 20 percent of adolescents' COAD emerged in each of its three subcategories, Behavioral Disconnection, Lonely, and Perceived Disconnection. The other categories with at least 20 percent of adolescents' COAD were Low Self-Worth, Suicidal, Anxiety Symptoms, and Contextual/Causal. These findings support the contemporary diagnostic criteria for depression but extend and deepen the current criteria of the literature, and contribute unique meanings to depression as adolescents define it; except for the Contextual category, these categories are found in the DSM and the K-SADS-PL. In the category Depressed Mood, the students' COAD extended its definition by including familiar (e.g., Sadness) and unique (e.g., Inner Pain) subcategories. In Social Impairment, this study's findings expand the concept of depression as defined by adolescents. The DSM criteria for diagnosing MDE requires a change from previous functioning, and social impairment is merged into an area of functioning impairment (caused by depressive symptomatology) that includes distress, and social, occupational, and functional impairment. These adolescents added depth and unique information by identifying Social Impairment as one of two main concepts of adolescent depression, and distinguishing such detail in Social Impairment as to create four subcategories generated by their COAD. Although Contextual/Causal refers to inferences pertaining to the environment, adolescents' generated COAD in this category indicates that context is important in their definitions of depression. Low Self-Worth, Suicidal, and Anxiety Symptoms are factors commonly linked with depression. Further discussion is developed separately in this section on each of these categories. 140 Chapter V: Discussion Depressed Mood Almost nine in ten adolescents contributed a COAD within the category Depressed Mood. In the DSM, Depressed Mood is one of the criteria integral for MDE. The literature, including the DSM-IV-TR (APA, 2000), contains an array of labels such as sadness that describe Depressed Mood. Depressed Mood is also referred to as affect on the subthreshold dimensions on a continuum of depression (Angold, 1988a) and the core symptom along this dimension. Adolescents clearly identify Depressed Mood as a critical characteristic of depressive symptomatology, as the literature suggests. However, what Depressed Mood means to adolescents differs with the literature in several of its subcategories and in the level of their importance. Such subcategories clarify and extend the meaning of Depressed Mood, and are presented in the following paragraphs. Sadness. Almost two in three adolescents included the subcategory Sadness (e.g., sad, sadness, unhappy) in their definitions of depression; this coincides with the DSM-IV-TR (APA, 2000) descriptions (e.g., sad, depressed, or down in the dumps). DSM-IV-TR (APA, 2000) and the literature indicate that irritability may be the mood adolescents report (APA, 2000); however, adolescents in this study identified Sadness as the prominent characteristic defining Depressed Mood. Irritability. Just over one in three adolescents described the subcategory Irritability (e.g., anger, frustration, upset, annoyed, disappointment) in their COAD, reflecting the term as it is used in the literature as a mood equivalent for depression (DSM-IV-TR; APA, 2000; K-SADS-PL; Kaufman et al., 1996; Mueller & Orvaschel, 1997; Parker & Roy, 2001; Russell & Fehr, 1994). Although a substantial number of adolescents include Irritability as a definition of depression which clarifies similarities in the literature, questions remain as to which definitions clustered within Irritability are more important than others. For example, does upset precede frustration, which may come before anger? Differentiating meanings of Irritability and the severity of Irritability across distinct contextual points may also be similar to distinctions made for Sadness. 141 Chapter V: Discussion In distinguishing the dimensions of Irritability or Sadness, comparisons may be made between these aspects and may reveal clearer dimensions of depression. According to the adolescents, Irritability and Sadness are integrally associated with Depressed Mood, and distinguishing the relationship between these moods, and whether one mood may precede another (e.g., if anger precedes sadness) may be valuable. No adolescent used the actual word irritability (or irritable) to define Irritability in this study. Inner Pain. More than one in ten adolescents provided Inner Pain as their COAD (e.g., inner pain, broken heart, hurting inside). Inner Pain does not reflect somatic symptoms of pain (e.g., abdominal pain, headache). Their descriptions of Inner Pain connote, rather, a metaphoric type of pain. It may be a term that is closer in meaning than Sadness to their definition of depression, in that Inner Pain may involve sadness and possible other aspects in context. The DSM-IV-TR (APA, 2000) mentions that the motivation for suicide may include a painful emotional state that is perceived by the person to continue without end. Perhaps, adolescents' inclusion of Inner Pain in their COAD may also link to pain in connection with motivations for suicide. Other Subcategories of Depressed Mood. Two in five adolescents defined depressed cognitions that included subcategories such as Hopeless, Unmotivated, Pessimism, and Bored/Indifferent. Hopeless and Unmotivated were the prominent subcategories in this area. Matching descriptions are found in the literature under cognitive facets such as hopelessness, boredom, indifferent, pessimistic (Zuckerman, 2005); under depressed mood (e.g., hopeless) (DSM-IV-TR; APA, 2000); in association with depression (e.g., poor motivation) (US Department of Health and Human Services, 1999); as negative thinking (Garber et al., 1993); hopelessness (Kreuger, 2002; Stanard, 2000); and in reference to ideational content (e.g., hopelessness, pessimism) (K-SADS-PL). In connection to ideational content, hopelessness may relate to a disruption of an individual's future goals or self-continuity, that may also be linked to suicidality (Ball & Chandler, 1989; Chandler & LaLonde, 1998). Low motivation and loss of interest or pleasure (anhedonia) are included under the heading Depressive Disorders in the K-SADS-PL, whereas only loss of interest or pleasure is reported as an integral criteria for MDE in the DSM-IV-TR (APA, 2000). In the K-SADS-PL 142 Chapter V: Discussion under Depressive Disorders, boredom is also referred to as loss of interest or ability to enjoy. Although Pessimism is not reported as a criterion for MDE criteria in the DSM-IV-TR (APA, 2000), it is mentioned as a criterion in Dysthymia. However, in adolescents' COAD, Boredom is distinguished without knowing the individual's previous condition. Hence, it is difficult to determine whether there is a "loss" of interest or pleasure based on their descriptions, and it is not known what state precedes the subcategory Unmotivated. Almost two in five adolescents included Depressed Appearance in their definitions (e.g., crying, teary, stop caring about their appearance). The DSM-IV-TR (APA, 2000) refers to facial expression and demeanor to help determine a diagnosis of MDE, with crying or tearfulness used as frequent presentations of MDE. In the K-SADS-PL, references are made in how an individual looks or acts when observations are made in interviews. Davidson et al. (1990) distinguish the importance of facial expression for the study of emotion, and difficulties in emotion management have been associated with poor social functioning (Shipman et al., 2003). Recognizing facial expressions indicate that social cognition or social competence is related to social cues or competence. The degree and meaning of Depressed Appearance can be important relative to these characteristics, however careful consideration needs to be made before attaching these meanings to Depressed Mood because adolescents did not provide that kind of detail in the subcategory Depressed Appearance. Adolescents identified Moody in their COAD but this term is not reported under MDE criteria in the DSM-IV-TR (APA, 2000). However, the DSM-IV-TR (APA, 2000) uses the term affective lability under Depressive Disorders Not Otherwise Specified, and as emotional under Dysphoria (Zuckerman, 2005). Deciphering the intended meaning of Moody as adolescents used it seems complex, but possible references to the environment and irritability may be more useful attributes to explore. In this study, adolescents provided one definition of Somatic Symptoms (e.g., difficulty breathing). This single reference, provided by the adolescents, contrasts with an emphasis that is put on somatic complaints (e.g., bodily aches, headaches, or joint pain, abdominal pain) in the DSM-IV-TR (APA, 2000) criteria for depression, rather than feelings of sadness. 143 Chapter V: Discussion Social Impairment More than two in three adolescents provided descriptions of depression in the category Social Impairment. Three out of its four subcategories, Behavioral Disconnection (e.g., hardly socializes), Perceived Disconnection (e.g., feeling misunderstood), and Lonely, had at least 20 percent of adolescents contribute definitions in each of these constructs, extending the previous literature. In the DSM-IV-TR, no rules or distinctions specifically refer to social impairment as requirement criteria for a diagnosis for MDE, rather a general functioning impairment is tied in with symptoms causing distress, social, occupational, functional, or other important areas of functioning. Others must observe the functioning impairment (that includes social impairment) to account for its presence in the affected individual. References in the literature allude to social impairment as being linked with depression; however, they are labeled as generalizations in association with depressive symptomatology. Some of these findings link depression with: individuals who are rejected by others (Joiner, 2000), loneliness (Brage & Meredith, 1994), alienation (Gjerde et al., 1988), lack of self-belonging (Stanard, 2000), maladaptive coping patterns (Kreuger, 2002), behavioral inhibition and shyness (Cherny et al., 1994), social information processing deficit (Crick & Dodge, 1994), and peer difficulties (Hankin & Abramson, 2001; Nolen-Hoeksema, Girgus, & Seligman, 1992; Petersen et al., 1991). In defining depression, most adolescents are saying in their COAD that rather than a general functioning impairment, it is more specifically about Social Impairment. Individuals seem to recognize Social Impairment in themselves and in others. Another critical factor is that adolescents recognize Social Impairment in terms of overt or behavioral disconnection, perceived disconnection from others, subjective feelings of loneliness, and in aggressiveness or bullying. As described in the DSM-IV-TR (APA, 2000; Criteria C in Table 5 in the Methodology chapter), functioning impairment must be overt and accompany depressive symptomatology to make a diagnosis in MDE. However, adolescents are defining Social Impairment in terms of both overt and covert terms, and reinforce this area as an integral part of depression definition. The Importance of Social Impairment. The importance of Social Impairment can be broadened in scale by linking it with other categories and subcategories generated by adolescents' COAD and the related descriptions 144 Chapter V: Discussion discussed in the DSM-IV-TR (APA, 2000) and in the literature. In this study, the subcategory Quiet (e.g., quiet, silent) is just one of these constructs that can belong to both its presently assigned category Retardation/Agitation and to Social Impairment. Recall that almost all of adolescents' COAD Table 12 in Results chapter) identified subcategory Quiet in the category Retardation/Agitation. In the DSM-IV-TR (APA, 2000), Retardation includes definitions such as speech that is decreased in volume, inflection, amount, slowed down in thinking (or slowed down in speech in the K-SADS-PL), or muteness, as one of the symptoms others observe that is used as a criterion for MDE. Hence, symptoms of Retardation (Criteria A-5 in the DSM-IV-TR; APA, 2000; Table 5) can exclude these individuals from also having a functional impairment (e.g., quiet, silent as overt criteria in Criteria C in the DSM-IV-TR; APA, 2000; see Table 5) in making the diagnosis for MDE. However, when social relational withdrawal (e.g., being quiet or silent) is observed, the individual may be slowed down in their thinking or retardation (overt evidence); they cannot talk because they cannot think. Keeping adolescents' COAD units in one category Retardation/Agitation lessened the impact of the Social Impairment category, already an important area as defined by adolescents and the possible missing link when and if impairment changed from a previous state. Other categories and subcategories may represent social indicators and link with Social Impairment. Some of the features of Irritability seem to connect with discontent with one's environment or relationships (e.g., are angry, blaming others, frustration). These features involve interpersonal factors and may contribute to Social Impairment. Further categories that can be socially construed and linked with Social Impairment are: Low Self-Worth, Masking/Stigma, Escaping Reality, or Feel Trapped. For example, in Low Self-Worth, comparisons are made about an individual's perceptions of himself or herself in relation to others. Masking/Stigma, too, is socially constructed, where an individual places importance on their understanding, privacy, expectations, and own actions in relation to others. Escaping Reality and Feel Trapped also have social relational bases. The importance of Social Impairment is critical when other constructs are examined in combination. Differentiating Qualities of Social Impairment. In further exploring Social Impairment as a link to depressive symptomatology, researchers need to exercise caution: being alone does not equate with lonely, and feeling lonely 145 Chapter V: Discussion does not readily equate with Social Impairment. Disconnecting from someone can be viewed in two lights, both positively and as an impairment. The element of change is also important. A normally shy person is not necessarily depressed, but an outgoing person who starts to exhibit this behavior may be. Hence, dimensional, contextual, and critically subjective factors attached to this quality need to be considered before social disconnection (e.g., behavioral or perceived) is addressed as an impairment. Social Impairment can be linked to Social Perspective Coordination, a core component of social competencies. Individuals can be vulnerable to outcomes of social risks if they do not have a mature perspective on what those risks mean to them personally (Selman, 2003). Even if they do have the maturity to understand social competencies as defined in Social Perspective Coordination, something else can occur that may entail their level of emotion understanding, and/or context, and/or other underlying physical dynamics that interconnect with each other. Other Categories Defined by Adolescents At least one in five adolescents contributed definitions in each of the categories Low Self-Worth/Guilt, Suicidal, Anxiety Symptoms, and Contextual/Causal. This study's category Low Self-Worth/Guilt (e.g., worthlessness, lack of self-esteem, guilt, ashamed, insecure) is defined as a criterion for MDE in the DSM-IV-TR (APA, 2000), and by other researchers who link low self-worth with depressive symptomatology (Kreuger, 2002; Loehlin, 1992; Lykken & Tellegen, 1996). In the Suicidal category (e.g., thinks of death, attempts suicide, hurt themselves, suicidal) adolescents' COAD used descriptions similar to those found in the DSM-IV-TR (APA, 2000); the literature is rich in linking suicide with depression. Adolescents' COAD in the category Anxiety Symptoms (e.g., fear, feel scared, stressed, always worry, anxiety) were similar to the descriptions of anxiety in the DSM-IV-TR (APA, 2000). Although anxiety is not a criterion for the diagnosis of MDE, the DSM-IV-TR (APA, 2000) mentions that adolescents with MDE frequently present with anxiety disorders (e.g., complain of feeling anxious, stressed, fear, worry). The literature shows that anxiety typically first begins in childhood and precedes the onset of adolescent MDD (Rutter et al., 2006), and adult MDD is most likely to be preceded by adolescent overanxious disorder (Pine et al., 1998). Other findings in the literature indicate that anxiety disorder often accompanies depressive symptomatology (Compas & Oppedisano, 2000; Compas et al., 1997; Rivas-Vazquez, Saffa- 146 Chapter V: Discussion Biller, Ruiz, Blais, & Rivas-Vazquez, 2004). In this study, adolescents identify Anxiety Symptoms as a definition of depression which clarifies previous literature in linking anxiety as a comorbidity to depressive symptomatology, and extends the DSM-IV-TR (APA, 2000) criteria. The category Suicidal (e.g., may resort to self-mutilation, thinks of death, attempts suicide, hurt themselves) is another construct identified by at least one in five adolescents' generated COAD, and the descriptions are similar to the criteria for MDE in the DSM-IV-TR (APA, 2000) (e.g., self-harm behaviors, recurrent thoughts of death, recurrent suicidal ideation, or suicide attempt) and in the K-SADS-PL (Kaufman et al., 1996), where self-mutilation is mentioned under the section Non-Suicidal Physical Self-Damaging Acts in Depressive Disorders. Adolescents' definitions of depression reflect prior investigations in associating suicidality with depression. Finally, one in five adolescents defined the category Contextual (e.g., stress, pressure, loss, rough life, pushed too hard) as depression. This finding contrasts with the criteria outlined in the DSM-IV-TR (APA, 2000), which displays that MDE often follows psychosocial stressors but in which context, as a criterion for the diagnosis of depression, seems to be ignored. Findings in the literature are closer to adolescents' COAD in the Contextual category. For example, studies indicate that life events are found to impact adolescent depression (Ge, Lorenz, Conger, Elder, & Simons, 1994; Goodyer, 1990; Lewinsohn et al., 1994), and the adverse impact of stressful events and strains in reduced social support affects psychological well-being (Mitchell & Moos, 1984). Further, the number and timing of stressful life events may provoke depression (Goldberg & Huxley, 1993; Rutter, 2000; US Department of Health and Human Services, 1999), and the level of stress was greater in a clinically depressed group than in a non-depressed group of students (Fuks Geddes, 1997). Adolescents seem to see things in a holistic sense interconnected with the context, Contextual/Causal. An individual's negative space seems to be important to their definition of depression, and these environmental factors play a role in their depression. Contextual/Causal factors and the degree of social change may overlap with social interactions in contributing to adolescent depression. Perhaps exploring Contextual/Causal relationships to depression in adolescents can be important in understanding depressive symptomatology in adolescents. 147 Chapter V: Discussion Remaining Categories and Subcategories in Descending Order of Percentages as Defined by Adolescents' COAD Adolescents provided definitions of depression in several categories reported and not mentioned in the DSM-IV-TR (APA, 2000). The categories generated by adolescents' COAD and corresponding to findings in the DSM-IV-TR (APA, 2000) criteria for depression (e.g., Retardation/Agitation, Loss of Interest or Pleasure, Indecisiveness/Impaired Concentration and Attention, and Appetite/Weight/Eat/Sleep Changes), will not be elaborated on. In the following paragraphs, I discuss only the categories defined by adolescents that were not described in the DSM-IV-TR (APA, 2000). I also discuss the category Functional Impairment/Distress because the adolescents' COAD differentiated it from the general functioning impairment described in the DSM-IV-TR (APA, 2000). In the DSM-IV-TR (APA, 2000), Functional Impairment (e.g., their marks might drop, thinking there is no use in school, distressed) is part of a general criterion (as is Social Impairment) that includes distress, social, occupational, or other important areas of functioning impairment (e.g., truancy, school failure, activities decline, distress), where overt evidence is needed for a diagnosis of MDE. For example, when an individual has lost interest (e.g., lose interest in school), the evidence of Functional Impairment could be that his or her marks might drop or truancy occurs (as defined in the DSM-IV-TR [APA, 2000]). In this study, Functional Impairment/Distress covers the remaining impairments after Social Impairment is distinguished in its own category. Adolescents' COAD revealed that Functional Impairment/Distress was considerably less important in defining depression. This paragraph discusses the remaining categories that were defined by adolescents' COAD but were not found in the DSM-IV-TR (APA, 2000) as criteria for depression. Although each category was minor in terms of its frequency compared to other more prominent constructs, adolescents contributed important information about them. The category Masking/Stigma (e.g., too scared to tell anyone about their own depression, false outgoingness; see Table 8 in Methodology chapter) implies that adolescents placed importance on their perceived image or identity in relation with others. This construct seems to reflect adolescents' inner conflicts in their emotions, perceived expectations, and privacy issues. Adolescents seem to be saying that stigma occurs, and that it is part of depression. The category Feel Trapped (e.g., being trapped in your life) may represent a point of hopelessness where there is no way out of their difficult or 148 Chapter V: Discussion impossible condition, and may be associated with suicidal thinking. Tulloch, Blizzard, Hornsby, and Pinkus (1994) in their article on suicide and self-harm in adolescents in Tasmania, Australia, suggest that common motivations for self-harm or suicide were to escape an impossible situation, rather than to attract the attention of others. Although a minor category, as defined by adolescents' COAD, Substance Use (e.g., take drugs, alcohol) was still described by them as a definition of depression. In the DSM-IV-TR (APA, 2000), alcohol or other substance use may be associated with criteria for MDE, but by definition, MDE is not due to the direct physiological effects of abuse of drug or substance. Rutter et al. (2006) suggest that there is sometimes a progression from substance abuse to major depression from possible psychopharmacological effects of substance abuse or interference with psychosocial functioning. What the adolescents may be saying is that someone who is on drugs, or alcohol, or other medication, may be depressed, or substance abuse may be a sign that he or she is depressed. Several adolescents' COAD were clustered within the category Escaping Reality (e.g., escape reality, numb out pain). Perhaps these terms may be the outcome of self-harm coping strategies, and may be related to adolescents' Inner Pain. In the category Biological (e.g., chemical imbalance, hormones out of control) adolescents' COAD suggests that it is of minor importance in their definitions of depression. The US Department of Health and Human Services (1999) defines biological influences on depression to include genes, infections, hormones, and toxins. In one response, an adolescent reported that acting out sexually is part of a definition of depression. These categories, not found in the DSM-IV-TR (APA, 2000) as criteria for depression, indicate that adolescents are aware that a multitude of factors can affect depressive symptomatology, and these factors form part of their understanding and descriptions of depression. Summary Adolescents' definitions of depression were dominated by subjective, holistic interpretations that are intrinsically related to overt characteristics and their context or environment. Depressed Mood and Social Impairment are the core categories reported in adolescents' definitions of depression. Adolescents highlight Social Impairment and its more intricate subcategories (e.g., Perceived Disconnection and Behavioral Disconnection) in their COAD as a definition of depression, not emphasized in the DSM-IV-TR (APA, 2000). Most adolescents define Sadness as the main subcategory of Depressed Mood followed by Irritability. 149 Chapter V: Discussion DSM-IV-TR (APA, 2000) highlights irritability as an adolescent depressive symptom. Although Depressed Mood is an integral criterion in determining a MDE diagnosis in the DSM-IV-TR (APA, 2000), adolescents have a more comprehensive definition of what Depressed Mood means to them. Associations of Adolescent Depression Constructs to Grade, Gender, Depressive Symptomatology, Self-Recognition of Depression, and Pathway to Talking to Someone When Depressed Chi-square tests, categories and subcategories developed from adolescents' definitions of depression were examined against grade, gender, presence and non-presence of screened depression, self-recognition of depression, adolescents' thinking about own need to talk to someone, and talking to someone when feeling depressed. Findings were significant in various associations between the constructs of adolescent depression and these different variables. This section presents the specific findings from these analyses. Associations of Categories and Subcategories of Adolescent Depression to Grade and Gender Association of Categories and Subcategories of Depression to Grade Suicidal, Depressed Mood, Inner Pain, Depressed Appearance, and Pessimism were significant findings between the grade levels. In each analysis, a higher percentage of the 8 th versus the 11 th graders provided definitions in these constructs. It seems that students in grade 8 generated their COAD both in overt and covert constructs when defining depression. The covert constructs, such as Inner Pain or Pessimism, reveal both cognitive and emotional understanding in the younger grade students. Cicchetti, Rogosch, Toth, and Spagnola (1997) suggest that cognition provides the structure, but that emotion is a separate developmental domain exerting reciprocal influences. Interacting with adolescents' cognition, emotion may exert its own force, providing the structure at the abstract developmental phase of thinking. These constructs may lead to the detection of specific characteristics that are important in the way students in grade 8 define depression; however, the meanings of these constructs need further elaboration from adolescents. 150 Chapter V: Discussion Association of Categories and Subcategories of Depression to Gender The constructs Low Self-Worth, Suicidal, Social Impairment, Retardation/Agitation, Masking/Stigma, Depressed Appearance, Perceived Disconnection, Aggression/Bullied, and Quiet were significant when differentiated between boys and girls. In each significant construct, girls had generated higher percentages of COAD than boys. These findings suggest that girls have a more mature level of development as seen in the greater number of different constructs girls provided, and their wider range of responses compared to boys. These findings may also have a connection with gender and socialization differences (Gjerde et al., 1988). All these constructs could reflect overt (e.g., Depressed Appearance, Quiet), covert (e.g., Perceived Disconnection), or both overt and covert (e.g., Masking/Stigma) social implications. The higher percentages of girls' versus boys' COAD in the significant category and subcategory findings may be linked with differences in pubertal development. Angold et al. (1998) report that pubertal development, as measured by Tanner stages, predict emergence of gender difference in depression, as girls report increased rates of depressive disorders after Tanner Stage III. Further, Allgood-Merten et al. (1990) report that more adolescent girls than boys are aware of inner states, are more self-reflective, more publicly self-conscious, and exhibit greater social anxiety. The importance of the underlying meaning of these significant constructs, however, are difficult to determine based only on girls' higher percentages than boys' in these significant constructs. Association of Categories and Subcategories of Depression to Gender in Grade 8 In grade 8, significantly higher percentage of girls contributed definitions more than boys, in the category Low Self-Worth, and the subcategories Perceived Disconnection and Aggression/Bullied. These findings suggest that there may be a social link that is more closely associated with girls as shown in their percentages of COAD in these constructs. Low self-worth may involve perceived inequality in relation to others and can overlap with perceived disconnection from others. Aggression/Bullied aspects may extend the perceived disconnection from others to an aversive relationship. Aversive behaviors may elicit negative reactions (Gjerde et al., 1988), may propagate the cycle of perceived low self-worth and disconnection that may lead to hostility in relationships. 151 Chapter V: Discussion Association of Categories and Subcategories of Depression to Gender in Grade 11 All the significant categories and subcategories linked with gender in grade 11 were different from the significant constructs associated with gender in grade 8. In grade 11, girls had significantly higher percentages than boys in Depressed Appearance, Social Impairment, Retardation/Agitation, and Quiet. Although the finding was significant in the subcategory Depressed Appearance, percentages of their COAD dropped for both boys and girls from grade 8 to grade 11, suggesting that this subcategory had less importance in adolescents' definitions of depression at the upper grade level. This may correspond to the level of development where abstract thinking is more pronounced. Social Impairment and Quiet revealed the effect of social relationship in defining depression. Association of Categories and Subcategories of Depression to Grade Levels for Boys Only A significantly higher percentage of boys in grade 8 than in grade 11 contributed definitions of depression in the category Depressed Mood, and its subcategories Inner Pain and Depressed Appearance. There may be several reasons for the increased percentages of generated COAD by the younger boys. The significant percentages in these constructs may reflect the younger boys' willingness to divulge their feelings at a younger age, possibly related to prepuberty. It may also be associated with peer changes in the transition to high school, where adjustments of interpersonal interactions can bring negative experiences that may persist over time (Rutter, 2000). Regarding Inner Pain, the significance of the drop in boys providing definitions in grade 11 (6%) from grade 8 (23%), seems to connect to the research that indicates depressed adolescent males tend to exhibit more externalizing symptoms (Frydenberg, 1997; Gjerde, 1995; Schonert-Reichl, 1994). The significant differences in that grade 8 boys more than grade 11 boys provide their COAD in Depressed Mood and its two subcategories suggest that they are expressing themselves differently in terms of depression. Association of Categories and Subcategories of Depression to Grade Levels for Girls Only Eighth grade girls had significantly higher percentages than 11 th grade girls in reporting depression definitions in the constructs Contextual/Causal and Depressed Appearance. Both 152 Chapter V: Discussion constructs reflect overt characteristics associated with a less mature level of cognitive development and approaching puberty, which coincides with the younger age in grade 8. Association of Depressive Symptomatology to Constructs of Adolescent Depression Depressed adolescents had significantly higher percentages of their COAD in the constructs Inner Pain, Feel Trapped, Anxiety Symptoms, and Anxiety/Stressed. Inner Pain is a distinct construct which contrasts the criteria found in previous literature including the DSM-IV-TR (APA, 2000). Inner Pain seems unique to depressed adolescents' experience of the condition and different from present DSM-IV-TR (APA, 2000) symptoms of Depressed Mood such as sadness or somatic complaints. The Inner Pain that adolescents may be defining is closer to the type of pain they are experiencing. Feel Trapped, another construct revealed in adolescents' COAD, is not found in the literature in connection with depressive symptomatology. Adolescents who feel trapped may feel so ensnared in their environment that this thinking may precede suicidal ideation. Anxiety Symptoms reflect the literature findings where anxiety disorder often accompanies depressive symptomatology (APA, 2000; Compas et al., 1997; Compas & Oppedisano, 2000). Perhaps depressed adolescents' COAD in the construct Anxiety Symptoms and the condition of being stressed or anxious may be connected to their vulnerability and link with depressive symptomatology. The change in the decreased percentage of adolescents screened as "Depressed" who provided definitions in Social Impairment suggests that, in the process of destabilization to depressive symptomatology, it may be more difficult for the adolescent to distinguish social characteristics when they are depressed. Just over half the "Depressed" adolescents provided their COAD in this construct, yet the majority of the "Not depressed" students contributed their definitions in Social Impairment. This extends even further to Aggression/Bullied, where not one "Depressed" adolescent was able to provide a COAD in this construct. Perhaps depressed adolescents are in a state of disequilibration or possibly a fixed-type of state based on their levels of maturation and social interaction in the way they represent themselves, in their relationships, and their social world. Being anxious and stressed can fixate the individual to concern himself or herself with their own state of anxiousness or being stressed, separating them from their context of relationships or their social world. 153 Chapter V: Discussion In association with Anxious/Stressed construct, adolescents seem to provide a holistic definition of being stressed, stress, and anxious as part of depression itself. Anxious/Stressed is not identified as a clinical criterion for depression in the DSM-IV-TR (APA, 2000), although studies have found that depressive symptomatology can be precipitated by psychological stressors (Post et al., 2003). Adverse early experiences can have a lasting impact on the way individuals react to future stresses and their vulnerability to depression (Boyce & Essau, 2005). It appears that most adolescents screened as "Not depressed" and "Depressed" have similar definitions of Depressed Mood and Sadness. However, the new construct that "Depressed" adolescents are using significantly more often than "Not Depressed" participants to define depression is Inner Pain. Self-Recognition of Depression and Its Association with Constructs of Adolescent Depression Just over half of adolescents recognized depression in themselves. Out of this group, 17 percent recognized depression within two weeks and the rest beyond the two weeks. Adolescents who self-recognized depression within two weeks had higher percentages of COAD than those who did not self-recognize depression in the constructs Anxiety Symptoms, Contextual, and Fear. As previously noted, significantly more depressed than non-depressed adolescents generated their COAD in Anxiety Symptoms category. In association with the Contextual category, more particularly for those adolescent who self-recognized depression within two weeks, is indicative that the context or environment plays a part in their definition of depression which may have connection to Anxiety Symptoms or Fear in their environment. Concerning Social Impairment and each of its subcategories, the findings are similar to associations found with depressive symptomatology. Higher percentages of adolescents who did not self-recognize depression versus those who self-recognized depression within the past two weeks provided their COAD in Social Impairment and each of its subcategories. Again, no individual in the group who self-recognized depression within two weeks provided a COAD in the construct Aggression/Bullied. These findings suggest that overt and covert social characteristics are integral to the self-recognition of depression, and adolescents seem more aware of Social Impairment when not thinking about the possibility that they may be depressed. A significantly higher percentage of adolescents who self-recognized depression within two weeks versus those who self-recognized depression beyond the two weeks, contributed their 154 Chapter V: Discussion definitions in Indecisiveness/Impaired Concentration and Attention. As a criterion for MDE in the DSM-IV-TR (APA, 2000), this construct may represent an early sign of depressive symptomatology in these adolescents. Perhaps self-recognized depression may foster conceptions that continue to feed upon the interrelationships between an individual's COAD, their self-recognition of depression, and depressive symptomatology. A higher percentage of participants who self-recognized depression beyond two weeks versus those who self-recognized within two weeks, provided their COAD in Social Impairment and in each of its four subcategories. Changes registered in the decreased number of adolescents defining Social Impairment and its subcategories seem important in early detection of self- recognized depression; when depression is self-recognized within two weeks, there may be some loss of social distinctions. Other studies show that altered patterns and meanings of interpersonal interactions can bring about negative experiences or vulnerability to depressive symptomatology (Brown & Harris, 1978; Rutter, 2000). Deficits in social functioning are predictive of antisocial behavior and mental disorders (Schultz & Selman, 2004), and an observed criterion used to assess and diagnose MDE in the DSM-IV-TR (APA, 2000). Association of Adolescents' Thinking About Own Need to Talk to Someone When Depressed to Constructs of Adolescent Depression A significantly higher percentage of adolescents, who thought that they needed to talk to someone when depressed versus those who did not think that they needed to talk to someone, provided definitions in the construct Feel Trapped. Perhaps these adolescents may be having a thoughtful cry for help. However, only 5.7 percent of this study sample provided definitions within this construct. Association of Talking to Someone when Feeling Depressed to Constructs of Adolescent Depression A significantly higher percentage of adolescents who did not talk to someone, versus those who did, contributed definitions in the Hopeless subcategory. It seems that when it comes to the point when adolescents' behavior is examined in an effort to find out how to help them, it may be too late in the process. Perhaps these adolescents may think that, at this stage, there is no hope in talking to someone. The decision whether or not to talk to someone would require 155 Chapter V: Discussion previous skills and knowledge about the support network, and it appears that when Hopelessness comes into adolescents cognitions, they tend to withdraw from social connectedness. There is less clarity in construct differentiations in connection with approach-oriented coping, specifically via talking to someone when feeling depressed. Similar Percentages in the Constructs Generated by Adolescents' COAD Several constructs of adolescent depression contained similar and high percentages of adolescents' definitions in each of the differentiated groups of participants. The majority of adolescents in all the groups provided definitions within the categories Depressed Mood and Social Impairment, and in the subcategory Sadness. The depressed group was the only group with just over half rather than the majority of adolescents who provided their COAD in the construct Social Impairment. This may indicate that both overt and covert social disconnection is associated with depressive symptomatology. Other constructs defined by at least one in five adolescents (Bigelow & Zhou, 2001) in each group were Irritability, Behavioral Disconnection, and Lonely. These findings suggest that these terms are common for all groups of adolescents. Associations of Self-Recognition of Depression to Depressive Symptomatology, and Pathways to Talking to Someone When Feeling Depressed Frequencies, percentages, and z2 tests were conducted to examine the associations of self-recognition of depression to depressive symptomatology, and to adolescents' thinking about own need to talk to someone and talking to someone at the time of feeling depressed. The finding in associating self-recognition of depression within the past two weeks to depressive symptomatology, extend prior research in this area. Lifetime self-recognized depression to pathways to talking to someone when depressed provides insight into a new territory of research that connects to approach-oriented coping in connection to depressive symptoms. In this section, I discuss the findings separately. 156 Chapter V: Discussion Association ofSelf-Recognition of Depression within the Past Two Weeks to Depressive Symptomatology Thirty one adolescents self-recognized depression within the past two weeks, 13 boys and 18 girls. Out of the 31 adolescents, only 14 or 45% of these 31 participants were screened as "Depressed" using the RADS-2 (see Table 26). Out of the 13 (9% of the boys) boys who self- recognized depression within the past two weeks, 5 (39%) were screened as "Depressed." Out of the 18 (10% of the girls) girls who self-recognized depression within two weeks, 9 (50%) were screened as "Depressed." It seems that the percentage of boys and girls who self-recognize depression within two weeks is fairly equal for boys and girls. Screening out "Depressed" adolescents qualified only about half of the boys and girls who self-recognized depression within two weeks. In other words, the key finding revealed by this calculation is that just under half of these adolescents who self-recognized depression in the same time frame that their depressive symptomatology was assessed qualify into the potential threshold depression criteria based on the DSM-IV-TR (APA, 2000). Hence, this discrepancy raises further questions that need to be addressed. These questions can have relevance to early detection and early intervention of depression in adolescents, as well as possible lessons in how to communicate with them. Several questions that need to be addressed as a result of these findings that the screening tool for depressive symptomatology only picked up less than half of the adolescents who self- recognized depression within the two-weeks, are presented in this paragraph. The findings in this study revealed that the mean and the median for the RADS-2 (Reynolds, 2002) total score were higher in those adolescents who self-recognized depression than in those adolescents who did not self-recognize depression. Hence, are these adolescents who self-recognized depression in the subthreshold state of their depressive symptomatology and can self-recognize their vulnerability to depression but are screened out because the assessment tool can only screen out the potential threshold individuals with depressive symptomatology? How do we want to communicate to these young people when addressing their depressive symptomatology, particularly when they are telling us that they recognize depression in themselves? Do we want to involve all these adolescents who self-recognized or self-diagnosed their depression within the past two weeks or continue to select out only the potential threshold half by simply staying within the boundary of "screened depression" according to a measure? We already know that most adolescents do not access support services and those who do, do not receive adequate care. Hence, we need to think 157 Chapter V: Discussion about the other half of adolescents who self-recognized depression within the two weeks who are screened as "Not Depressed." If the aim is to better to communicate with adolescents and for adolescents to take ownership of their own understanding of what depression means to them, we need to redirect our thinking according to what these young people are telling us. The finding that the self- recognizers of depression have a significantly higher mean and median than the non self- recognizers of depression on the Depression Total Score on the RADS-2 (Reynolds, 2002) can be an integral part of the important factors that translate into early detection and early intervention of depression in these adolescents. This finding suggests that adolescents' self- recognition of depression as part of the critical factor in detecting depressive symptoms in themselves can be verified just by asking them if they think they are depressed. Self-recognition of depression involves adolescents' integration of their understanding and conceptions of depression, including their rating of the importance of their depression. To self-recognize depression, the individual needs to be able to conceive of what depression means to him or her personally. This may be the initial step to early detection of depression in the adolescent. The potential for early detection of depression through self-recognition of depression can help individuals maintain control of their depression. Recent studies indicate that depressive symptoms or subthreshold depression exist on a continuum (Judd & Akiskal, 2000; Lewinsohn et al., 2000a; Lewinsohn et al., 2000b; Ruscio & Ruscio, 2000), and that dimensions on this continuum are appropriate models for diagnosis (Goldberg, 2000; Slade & Andrews, 2005). Once individuals move into the more severe dimensions of depressive symptoms, they may not be able to recognize the condition as readily. In these findings, self-recognition of depression is associated significantly with the presence of screened depression; it may be the critical factor that informs early detection of depression before any help-seeking begins. In a theoretical paper, Cauce et al. (2002) contend that help-seeking cannot begin in earnest until the problem of mental health need is recognized; and according to Piaget, thought precedes action as cited in Chapman (1988). 158 Chapter V: Discussion Lifetime Self-Recognized Depression to Reported Pathways to Talking to Someone When Feeling Depressed Those adolescents who self-recognized depression in their lifetime (185 [56%]) were selected to examine the association of self-recognized depression to pathways to talking to someone when feeling depressed. First, association of self-recognized depression to thinking about own need to talk to someone when depressed, was examined. In this analysis, the majority (68%) of adolescents thought that they needed to talk to someone. In further analysis, the findings revealed that the majority of adolescents who thought that they needed to talk to someone when depressed, actually did talk to someone (79%). These findings reflect the results of Seiffge-Krenke and Klessinger (2000), where fewest depressive symptoms were reported with approach-oriented coping style. Other studies examining self-perceived need for mental health treatment found that self-perceived need, independent of the DSM diagnoses, was associated with suicidal ideation (Sareen et al., 2002), and mood disorders and suicidality were strong predictors of self-perceived need (Mojtabai et al., 2002). Correlates of health may also indicate indirectly adolescents' level of help-seeking abilities (Compas et al., 2001), and that all forms of avoidant coping are linked with high levels of depressive symptomatology (Seiffge-Krenke & Klessinger, 2000). In reference to talking to someone, this study's findings reflect the Seiffge-Krenke and Klessinger (2000) study, where approach-oriented coping was reported with fewest depressive symptoms, but where avoidant- copers reported the most depressive symptoms. Two years after the study, higher levels of depressive symptoms were found in all adolescents who used avoidant coping. In this study, the majority of adolescents who thought that they needed to talk to someone would not meet the threshold requirement for depressive symptomatology. This finding can reflect Seiffge-Krenke and Klessinger (2000) study results where approach-oriented coping was reported with fewest depressive symptoms. Further, in this study, adolescents screened as "Depressed" had lower percentages (53%) of conceptions of depression versus those screened as "Not Depressed" (70%) in the category Social Impairment. Similar findings were found in those adolescents who self-recognized depression within two weeks (had lower percentages of COAD) versus those who did not self-recognize depression and who self-recognized beyond two weeks. Also, in this study, a higher percentage of "Not Depressed" adolescents (72%) talked to someone versus those who were screened as "Depressed" (54%). These findings seem to echo Seiffge- 159 Chapter V: Discussion Krenke's and Klessinger's (2000) and other findings (Herman-Stahl et al., 1995; Murberg & Bru, 2005), where avoidant-copers reported the most depressive symptoms. The study's findings show that adolescents' approach-oriented coping style specifically via thinking about own need to talk and talking to someone can indirectly extend from their conceptions of depression and their self-recognition of depression. Logically, impaired social functioning would be related to impaired help-seeking pathways, which require greater maturity in social and emotional understanding. Subjective thought of needing to talk to someone when depressed precedes talking to someone and may be the reference point for early intervention of depression in adolescents because by the time adolescents decide to talk to someone, it may be too late. Moreover, self-recognition of depression can precede adolescent's thought about own need to talk to someone. This study's findings in adolescents' approach-oriented coping style when feeling depressed, show that the majority of adolescents included Social Impairment in their definition of depression. Social impairment works against an individual's ability to seek help. As a result, affected young people who are socially impaired have more difficulty in seeking help on their own. In essence, adolescents are less likely to seek help when depressed because their social impairment is part of the problem that defines their depression. This finding reinforces studies showing that most young people do not seek out support services. Approach-oriented coping specifically via adolescents' thought about own need to talk and talking to someone when feeling depressed may involve other factors such as family influence, peer influence, school referrals, as well as the context itself. Increased understanding and a higher capacity to coordinate relationships would suggest that contextual factors are important considerations in adolescent reluctance to talk to someone when feeling depressed. Adolescents may recognize that they may be depressed, but once they look for help in dealing with depression, they confront barriers within themselves and their environment. However, the findings in this study, presented in this section, also point out the factors that can enable communication with adolescents, early detection, and intervention. 160 Chapter V: Discussion Social Perspective Coordination, Emotion Understanding, and Emotion Regulation, as Predictors of Severity of Depressive Symptomatology This study placed adolescents' depressive symptomatology within a theoretical framework by examining the ways that adolescents' levels of Social Perspective Coordination, emotion awareness, emotion expression, and emotion regulation were associated with the dimensions of their depressive symptomatology. Relations of Social Perspective Coordination and emotion understanding to depressive symptomatology were examined both correlationally and in a series of hierarchical regression analyses. In this section, I begin with a discussion of the findings from the correlational analyses and continue with a discussion of the regression analyses. Intercorrelations among Social Perspective Coordination, Emotion Understanding and Depressive Symptomatology The correlational analyses revealed several significant relationships. A significant correlation between severity of depressive symptomatology and Poor Emotion Awareness, including the reluctance to express emotion, provided support for this complex link and the importance of examining these interconnections further (Davis, 1999). This relationship is also linked to reports pointing out that the reluctance to express emotion can play an important role in disrupting social relationships which could lead to psychopathology (Lane & Schwartz, 1987; Penza-Clyve & Zeman, 2002; Saarni, 1999). A lesser but significant positive correlation between increasing levels of Anger Regulation with depressive symptomatology revealed support for previous literature showing that the essential building block to emotional competence is emotion awareness (Saarni, 1999). Significant negative correlations revealed that the presence of a more sophisticated Social Perspective Coordination and the ability to cope with sadness decrease as the severity of depressive symptomatology increases. Though caution should be exercised in suggesting causal links, these results may indicate that greater awareness and expression of emotion, mature Social Perspective Coordination, and coping with sadness may serve as a hedge to the increasing severity of depressive symptomatology. 161 Chapter V: Discussion Hierarchical Regression Analyses Predicting Depressive Symptomatology A multiple regression analysis further examined relations of levels of Social Perspective Coordination, emotion awareness, emotion expression, and emotion regulation as predictors to dimensions of depressive symptomatology. An examination of the hierarchical regression analyses revealed Poor Emotion Awareness to be the strongest contributor to severity of depressive symptomatology. This finding supports and extends previous research that has shown a positive correlation between the self-reported difficulties with emotion awareness in children (aged 9 to 13) and depressive symptomatology (Penza-Clyve & Zeman, 2002). Although assumptions regarding emotion awareness and emotion understanding cannot be confirmed without longitudinal data, this finding extends the literature in that increasing difficulty in identifying internal emotional experiences are related to severity of depressive symptomatology. Further, this finding provides weight to the correlational finding suggesting that, as Poor Emotion Awareness increases, it is associated with severity of depressive symptomatology. A reluctance to express emotion was another variable that was moderately significant in predicting the severity of depressive symptomatology. This finding supports previous research in children, where a correlation was found between reluctance to express emotion and depressive symptomatology (Penza-Clyve & Zeman, 2002). This study's data advances the view that increasing reluctance to express emotion is associated with concomitant levels of depressive symptomatology in adolescents. The emotion expression variable was weaker than emotion awareness as a contributor to depressive symptomatology, reinforcing previous reports indicating that a willingness to express emotion is not considered as essential a block to emotional competence as is emotion awareness (Penza-Clyve & Zeman, 2002; Saarni, 1999). This finding provides weight to the correlational finding suggesting that increasing reluctance to express emotion is associated with increasing severity of depressive symptomatology. Increasing ability to cope with sadness was another significant variable inversely related to severity of depressive symptomatology. This study's finding reflect other studies showing emotion regulation associated with lower levels of depression (Salovey et al., 1995), although there has been no differentiation of emotion regulation (e.g., sadness or anger). Based on a sample of children in grades four and five, Zeman et al. (2001) suggest that children who report using effective strategies for coping with sadness tend to experience lower levels of depressive 162 Chapter V: Discussion symptomatology. This result clarifies and broadens our view that varying dimensions of coping with sadness are associated with lower levels of depressive symptomatology. The final variable that contributed significantly to the severity of depressive symptomatology was gender. This finding confirms previous studies where the rate of depression emerges and increases in adolescent girls at about a 2 to 1 female-to-male ratio (Cairney, 1998; Hankin et al., 1998; Nolen-Hoeksema & Girgus, 1994; Parker & Roy, 2001). This result further reinforces the significant difference in depressive symptomatology between the boys and girls in grade 11, a 6 to 1 girl-to-boy ratio in this study. This finding makes intuitive sense in that girls have a higher ratio than boys in experiencing depression, girls are more likely to experience internalizing symptomatology (Frydenberg, 1997; Gjerde, 1995; Schonert-Reichl, 1994), and they more often report emotional disorders (Chen et al., 1998). Another finding in this study, although not significant in the final analysis, was that maturity in Social Perspective Coordination contributed to lower levels of depressive symptomatology. This variable was significant in correlational analysis, and in the second of the three steps in hierarchical regression analysis. However, when the emotion variables were added in the third step of hierarchical regression analysis, the emotion variables overrode Social Perspective Coordination. These findings suggest that Poor Emotion Awareness and Expressive Reluctance of emotion are critical in predicting severity of depressive symptomatology. Maturity in Social Perspective Coordination may well be interconnected with emotion understanding. This consideration is strengthened by other researchers who report that social competence requires emotion awareness and a willingness to express emotions (Halberstadt, Denham, & Dunsmore, 2001; Penza-Clyve & Zeman, 2002; Saarni, 1999; Zeman et al., 2006). Emotion awareness or willingness to express emotion is essential for emotional competence (Halberstadt et al., 2001; Penza-Clyve & Zeman, 2002; Saarni, 1999; Zeman et al., 2006), while the reluctance to express emotions is critical in disrupting social relationships and can lead to psychopathology (Lane & Schwartz, 1987; Saarni, 1999; Penza-Clyve & Zeman, 2002). Difficulties in managing emotion can also be associated with poor social functioning (Shipman et al., 2003). Most adolescents define Social Impairment and Depressed Mood as the two integral concepts of adolescent depression. Depression is termed an emotional disorder (Goldberg & Goodyer, 2005) or a mood disorder in the DSM-IV-TR (APA, 2000). While DSM-IV-TR (APA, 2000) includes impaired functioning that involves social difficulties as one of its criteria in 163 Chapter V: Discussion making a diagnosis for MDE, other studies on social adaptive patterns indicate that deficits in social functioning are predictive of mental disorders (Schultz & Selman, 2004). Emotion and social factors are reflected in adolescents' integral definitions of depression, Depressed Mood and Social Impairment. This study extends the framework of theories of social and emotional understanding by asserting that emotion awareness seems to be the linchpin for emotion expression and regulation, displacing growth in Social Perspective Coordination as the core contributor to lower levels of depressive symptomatology. Since Poor Emotion Awareness is an important contributor to the severity of depressive symptomatology, early intervention programs that stress awareness and management of emotion may have positive implications for depression prevention. Gender Differences in Social Perspective Coordination and Emotion Understanding as Predictors of Severity of Depressive Symptomatology A multiple regression analysis was conducted to further examine the relationship dimensions of Social Perspective Coordination, emotion awareness, emotion expression, and emotion regulation as predictors of dimensions of depressive symptomatology differentiated by gender. Hierarchical regression analyses again revealed that increasing Poor Emotion Awareness and reluctance to express emotion significantly predicted the severity of depressive symptomatology for both boys and girls. Although conclusions regarding changes in emotional development cannot be confirmed without longitudinal research, these findings show that increasing levels of depressive symptomatology are associated with concomitant levels of Poor Emotion Awareness and Expressive Reluctance. These findings support the correlational results suggesting that increasing Poor Emotion Awareness and reluctance to express emotion are associated with severity of depressive symptomatology. When gender was differentiated, coping with sadness was the only other significant variable that contributed inversely to the severity of depressive symptomatology in girls only. Taken together, these findings indicate that emotion awareness can have implications for depression prevention and early intervention programs by stressing awareness and management of emotions for boys and girls. Also, coping with sadness can have implications for emotion- focused, health promotion programming for girls. 164 Chapter V: Discussion Study Strengths and Limitations Study Strengths First, a vital strength of this study is that it adds to an important area of research on adolescent depression by shifting the focus to adolescents themselves to provide us with their understanding and conceptions of depression. Moreover, this investigation enabled adolescents' conceptions of depression to be differentiated against various factors including age, gender, and depressive symptomatology. This type of investigation has been unexplored and extends our insight into what adolescent depression means to these individuals. Adolescents were able to generate rich definitions of depression from a holistic perspective which can allow us to communicate with young people and to help us detect their potential vulnerability to depression. Adolescents' definitions of depression generally support, clarify, and extend present criteria for depression in the DSM-IV-TR (APA, 2000) and the K-SADLS-PL (Kaufman et al., 1996). Several categories, generated by adolescents' conceptions of depression, were significant when analyzed against various factors. Second, an additional contribution of this investigation was to establish adolescents' capacity to self-recognize or self-diagnose depression through time recall, and determine adolescents' understanding and management of depression that could translate into practice. Adolescents' self-recognized depression within two weeks expanded our attention beyond their depressive symptomatology. Adolescents' management of depression may also be linked with correlates of coping. Third, a significant strength of this study is in placing depressive symptomatology within a theoretical framework that was drawn from the literature on theories of social and emotional understanding. This approach advanced our knowledge in this area to broaden our focus into developmental constructs that might underlie and contribute to the severity of depressive symptomatology. Both correlational analyses and hierarchical regression analyses revealed that Poor Emotion Awareness was the strongest contributor to the severity of depressive symptomatology, followed by Expressive Reluctance, and gender. Not coping with sadness was also a significant contributor to severity of depressive symptomatology, especially for girls. Fourth, both a strength and limitation, self-report/recall data were used in this study. Their strength is discussed in this paragraph and their limitations under a separate Limitations 165 Chapter V: Discussion heading. This study's main focus was to obtain adolescents' definitions of depression, self- recognition, and pathways to talking to someone when depressed. Particularly for this purpose, a self-report study has a number of advantages which include adolescents' perspectives and cost efficiency because data is collected from one source. Self-reports are also an important method to measure cognitive responses that are unobservable by others (Compas et al., 2001). Hankin and Abramson (2001) suggest that children may be best informants after the age of nine. Finally, rigorous data collection methods were firmly adhered to throughout the course of this study. Prior to the main research study, a pilot study and a focus group were implemented to ensure the questions, timing of the survey, and terminology in the questionnaire were appropriate for this age group. Communication was established with school personnel and presentations were given as requested. Extreme care was taken to ensure that the students and parents were informed about the purpose of this study, and a representative return rate on parental permission slips giving consent were obtained. Two research assistants, one male and one female, were trained by this investigator to help administer the questionnaires. Protocol was established to have follow- up within 24 hours for those adolescents who scored with potential depression. Resource pamphlets were developed and provided to all students in the classrooms. I also made myself available for any personal questions immediately after the survey. A large sample of participants was obtained for this investigation to ensure an adequate number of adolescents with depressive symptomatology could be found. Limitations While this study has made significant contributions to adolescent depression literature it is not without limitations. First, I discuss some of the methodological limitations in this investigation. The data of this research rely on adolescents' self-reports that stemmed from a survey. Interviews may have clarified, refined, and provided more in-depth information as to adolescents' definitions of depression. Although nine percent of the sample of participants was screened with depression, reflective of the prevalence found in the community (Kessler & Walters, 1998), those who were clinically depressed, or who may have scored with potential depression, or provided different responses to the questions, may have not participated in this study. Non-participation may have also been due to other factors such as school, ethnicity, or these non-participants may have been depressed. Only those adolescents who chose to do so, 166 Chapter V: Discussion along with their family's permission, participated in this study. Another methodological limitation to this study was that the study sample was ethnically diverse and teachers were the only ones to identify students who were able to comprehend oral and written English. Reading literacy was not examined to establish if there were any differences in participants' abilities to express themselves in English. Further, requiring the participants to understand written and spoken English may have excluded some immigrant students from this study. Second, adolescents' conceptions of depression were limited to descriptive definitions of depression, and did not provide the dimensions of their meaning. For example, the subcategory "Sadness" lacked the information as to its severity of sadness (e.g., deep sadness, very sad). Further, definitions were limited in that the context of "sadness" was missing (e.g., sad with their life, sad at school). Gaining knowledge about these contextual factors that adolescents feel sad about may have provided a fuller understanding of their "Sadness" in defining depression. Third, the assessment measures used to measure depressive symptomatology present a limitation. The use of other informants for the assessment of depression in adolescents avoids the inherent difficulty of self-report and the possibility of distortion of self-report observations. Hence, the data from this study may be questionable because it was based on the subjective appraisal of the participants, and can produce biased results where distortions arise as recall periods are longer (Mechanic, 1978). However, adolescent depression should not be considered as a condition independent of the respondent, and adolescents are in the best position to know their feelings and behaviors in different situations. The inclusion of a third subthreshold depressive symptomatology group between the non-depressed and the depressed group, may have supported a richer concept with more levels of depression in adolescence. This means that various factors analyzed against the three classified groups, could have refined and broadened the dimensional differentiations among the groups and detection of depressive symptomatology. Moreover, because this study was on adolescents' understanding of depression, the participants may have been hesitant to express their status via self-reports in the survey. Another item, adolescents' thinking about own need to talk to someone when depressed, lacked time recall, hence can be subject to memory and other distortions (Mojtabai et al., 2002). Limiting the recall period to the past 12 months may have provided more depth to this question. Fourth, ethnicity and socioeconomic status were not delineated in adolescents' conceptions of depression or other analysis in this study. Hence, generalizability of these 167 Chapter V: Discussion findings to adolescents from other racial/ethnic backgrounds and/or family compositions and socioeconomic status is limited. This study sample was ethnically diverse. Ethnicity differentiations, particularly in adolescents' definitions of depression, their self-recognition of depression, and their pathways to talking to someone when depressed, could provide a richer and broader knowledge. This study's sample of participants stemmed only from schools located in middle to upper income communities. Consequently, generalizability of these findings to adolescents from lower income neighbourhoods is limited. Different cultural backgrounds and neighbourhoods demarcated by different socioeconomic status may have influenced experiences, symptoms of depression, and communication of responses. Fifth, the school principals and teachers chose to participate in this research. Out of the four schools taking part in this study, one school participated in depression screening six months prior to this study's data collection and may have been interested in the study. A second school did not have any depression programs. It is not known if there were any depression awareness programs for the remaining two participating schools. Further, three out of four participating classrooms took place during regularly scheduled Social Studies classes. Although it is not known about the discussion topics in these classes, knowledge from possible classroom discussions on depression may have influenced adolescents' contributions of their definitions of depression, self-recognition of depression, and approach-oriented coping style for depression. Sixth, the sample of participants was not randomly selected for this study. Further, the adolescents in the participating schools had the final choice in deciding whether or not to take part in this research. The participation rate in this study was 63% and it s unknown why the other 37% non-participating adolescents did not take part in this study. Studies show that avoidant- copers reported having the most depressive symptoms whereas fewest depressive symptoms were reported with approach-oriented coping style (Herman-Stahl et al., 1995; Murberg & Bru, 2005; Seiffge-Krenke & Klessinger, 2000). The non-participants may have reflected those adolescents with avoidant coping style or those who may be clinically depressed and who may not be at school. Finally, it is acknowledged as a limitation and as a suggestion for future research, that this study's findings were determined with a single sample. Replication of these findings with a similar sample of adolescents should continue, to ensure that the patterns revealed in this investigation reflect the nature of this study. Further, because this was a correlational, cross- 168 Chapter V: Discussion sectional study, the connections can be made to grade but not to development. A longitudinal study may be important in differentiating the relationships of cognitive, emotional, and social development to adolescents' conceptions and self-recognition of depression, as well as their coping styles when depressed. Implications and Considerations for Research, Policy, and Practice The findings of this study can have direct implications for research, policy, and practice strategies aimed to better communicate and help young people with and without depression. The present research advances our knowledge in adolescents' conceptions of depression, self-recognition of depression, pathways to talking to someone when feeling depressed, and in the contributions of social and emotional understanding to depressive symptomatology. The results indicate that adolescents clarify, provide depth to, and extend our understanding of depression. Moreover, depressive symptomatology is closely connected with adolescents' self- recognition of depression, however self-recognition of depression extends beyond the boundaries of screened "Depression." The majority of adolescents who self-recognized depression in their lifetime, thought that they needed to talk to someone when depressed and actually did talk to someone. However, also when it comes to talking to someone, it is those young people screened as "Not Depressed" who are more likely to talk to someone. Finally, Poor Emotion Awareness, reluctance to express emotion, gender, and decreasing levels of coping strategies with sadness, are important contributors to severity of depressive symptomatology. Depression may be a common pathway for a variety of other problems that can overlap with it, hence intervention at all the three dimensions may extend across the boundaries of depression in helping these young people. In this section, these implications and considerations are discussed under their respective headings. Research As discussed previously, adolescent depression is a complex mental health problem offering diverse challenges in its modes of presentation and intervention, and posing societal concern and research needs. The findings of this study beg for further investigations to build upon its framework. Future research should continue to differentiate and examine adolescents' understanding and conceptions of depression, linked with their cognitive, and social and 169 Chapter V: Discussion emotional capacities. Further insight into depression in adolescents can inform affected individuals, family, peers, school personnel, service providers, and health practitioners in ways that young people can connect to. One of the goals for future research should be to decipher what "negative cognitions" or "negative emotions" really mean. What depth of "sad" is associated with depressive symptomatology? Is it a deep sadness, very sad, more sad? What do we mean by depressed mood? In this study, although categories and subcategories were developed based on adolescents' conceptions of depression, future research should continue to examine the depth of these constructs and develop a hierarchical-type structure where subcategories contain sub- subcategories of definitions of depression. This type of structure may provide more refined conceptions of depression that can inform both practice and potential early detection of depressive symptoms. For example, irritability, a criterion for Depressed Mood in adolescents in the DSM-IV-TR (APA, 2000), and generated by adolescents' conceptions of depression, contains several definitions such as frustration, anger, or hatred. Refining the meanings of adolescents' conceptions of depression can advance our understanding of depression and how it relates to their world and distinct dimensions of depression, as well as provide a deeper insight to its overall concept that can translate into early detection of depression. The results of this research indicate that different conceptions of depression are associated with age, gender, depressive symptomatology, self-recognition of depression, and pathways to talking to someone. Future investigations should be concerned with examining these distinctions and their meaning. Conceptualizations form the meanings that can interact with self- recognition of depression and their pathways to talking to someone when depressed. Future research should be concerned with examining the contextual factors (e.g., social or structural changes, or media) that can influence the distinctions in adolescents' conceptions of depression associated with depressive symptomatology. The study's findings indicate that contextual factors are important to adolescents' definitions of depression and self-recognition of depression. Self-understanding and self-preservation is essential in times of social change. Changes that alter the structure of social life can overburden or impact adolescents' coping resources (Crockett & Silbereisen, 2000). Perhaps the meanings of adolescents' definitions of depression reflect the interconnection between their conceptualizations and their environment. Contextual factors can also have implications for adolescents' cognitive, social and emotional 170 Chapter V: Discussion development, and adjustments in cultural beliefs and social institutions. This study's finding indicates that emotion awareness is a significant contributor to depressive symptomatology. Hence, how can emotion awareness, also a critical component to social competence, be experienced and refined within social relationships? Future research should continue to examine adolescents' recognition of their vulnerabilities, and emotion-focused interventions. Differentiations of adolescents' self- recognition of depression may be valuable in early detection of depression. Poor Emotion Awareness, emotion Expressive Reluctance, decreasing dimensions of coping with sadness, and gender contributed to increasing levels of depressive symptomatology. Pilot investigations in different forms of practice strategies can explore emotion-focused interventions for boys and girls. These strategies would need to include intersectoral teams who develop, build on, and interconnect the research findings that can translate into relevant practical interventions that communicate with adolescents. Theory of Mind (ToM) researchers are interested in understanding children's own understanding of their mental lives or mental states (Chandler & Carpendale, 1998; Taylor, 1996). Theory of Mind may interconnect with formal operational thinking and interconnect with personal experiences, understanding, conceptions of depression, and self-recognition of depression; however, research has not explored this area. Future research can initiate investigations with these links. Adolescents' conceptions of depression included Stigma, and exploration of this area would be valuable in addition to research examining the stigmatizing labels we attach to depression. In attaching labels such as "negative mood" or 'negative cognitions," are we distinguishing what depression really means to adolescents to help them with their depressive symptomatology? What information do we impart when distinguishing a "negative" label and attaching it to a characteristic of depression? How do we want to differentiate depression and what value do we want to attach to these labels? These questions relate to Stigma. Studies that integrate cultural differences in conceptions of depression and pathways to talking to someone when depressed would be a valuable addition to research strategies that can inform each other. Ethnicity was diverse in this study sample. Further differentiating cultural differences in adolescents' conceptions of depression and their pathways to talking when 171 Chapter V: Discussion depressed, may be an important component in connecting and communicating across cultures. Cultural differences can include marginalized groups such as homeless adolescents. Future investigations should continue cross-sectional designs and initiate longitudinal designs to illuminate further relations among emotion understanding, social perspective coordination, depressive symptomatology, and dimensions of conceptions of depression. More particularly, longitudinal studies are relevant in distinguishing the relationships of cognitive, emotional, and social development to adolescents' conceptions of depression. These factors can be followed in terms of family structures, attachments, and practices in influencing development in adolescents. The influence of cultural and family backgrounds cannot be overlooked in how adolescents come to understand and conceptualize depression within their social and structural world, recognize it in themselves, and develop their pathways to talking to someone when depressed within their contexts. Policy The sheer scope of adolescent depression which overlaps intersectoral teams and potential initiatives suggests that the task is complex, and decision-makers must attempt to be comprehensive yet realistic. Decision-makers can build upon existing structures and expertise on intersectoral teams to try to develop practice and funding strategies that not only provide continuity of care but also depth of support at the different levels of practice and funding. Policy should attempt to anchor partnering with universities, schools, and community. Funding opportunities should be more available at the research level and training level within the disciplines, inter-disciplines, and multi-disciplines. Decisions need to be inclusive of a more complete understanding of the systems, as well as of the family working out towards the community, and their interrelationships with each other. This type of decision-making can pave the way to greater understanding of depression, other mental health problems, and harmful behaviors. Understanding trauma, violence, victimization, stigma, school failure, and underemployme