21H3l805 MICHIGAN STATE “J11! Nil} ll}; “wrl’m 00571 3221 Ill LIERARY Michigan State University ‘—l This is to certify that the dissertation entitled CORRELATIONS AMONG PERCEPTIONS OF DISEASE IMPACT, PERSONAL CONTROL, AND INFORMATION IN PSYCHOSOCIAL CONTEXTS OF ADOLESCENCE presented by Randall John Varga has been accepted towards fulfillment of the requirements for Ph. D degree in Educational Psychology 7/ Aug/1% Major professor Date 8/4/88 MS U i: an A ffirmati w Action/Equal Opportunity Institution 0-12771 MSU LIBRARIES “ RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. {a O t. .{uani CORRELATIONS AMONG PERCEPTIONS OF DISEASE IMPACT, PERSONAL CONTROL, AND INFORMATION IN PSYCHOSOCIAL CONTEXTS OF ADOLESCENCE BY Randall John Varga A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Counseling, Educational Psychology and Special Education 1988 6‘63? (70ij ‘ ABSTRACT CORRELATIONS AMONG PERCEPTIONS OF DISEASE IMPACT, PERSONAL CONTROL, AND INFORMATION IN PSYCHOSOCIAL CONTEXTS OF ADOLESCENCE BY Randall John Varga Most contemporary theories of coping emphasize sensitivity to the particular psychosocial contexts in which coping occurs . Information relevant to both disease and psychosocial context may facilitate coping and thereby attenuate psychosocial disease impact . The variables in the present study were opera- tionalized very narrowly as a means of examining more closely the cognitive mediation between context-specific information and impact. Correlations among four variables relevant to the psycho- social perspectives of chronically ill adolescents were computed in each of eight psychosocial contexts and at the across-contexts level . The four variables were Psychosocial Disease Impact, Self-Efficacy, Information Adequacy, and Health Locus of Control . All variables other than Health Locus of Control referred to disease-related psychosocial information in the following contexts: appearance, peer relationships, future orientation , sexual functioning, freedom/ independence , school achievement, physical abilities, and family relationships. Thirty chronically ill adolescents, ages 13-18 years, responded to a 42-item questionnaire which contained the four operationalized variables . Correlations between Impact and Sel f -Ef f icacy generally were moderate and positive , ranging from . 187 to . 515 across contexts. Correlations between Impact and Information Adequacy were negative (range: - . 213 to - . 714) . Correlations between Impact and Health Locus of Control were negative and tended to be moderately high (range: - . 3 01 to - . 778) . The addition of Information Adequacy to Health Locus of Control generally enhanced prediction of Impact whereas the addition of Self-Efficacy did not . A moderate amount of contextual specificity was evident in the correlations between Impact and Sel f -Ef f icacy. Inter- context differences in correlations between Impact and Information Adequacy were minimal . A two-factor (age and sex) analysis of variance was con- ducted with each variable within each context and at the across- contexts level . None of the main effects of age attained sta- tistical significance. The Self-Efficacy ratings of females were significantly higher in two contexts , appearance and peer relationships , and at the across-contexts level . Younger females produced significantly lower Information Adequacy ratings in the context of peer relationships . Interpretations of the observed relationships were pro- vided based on the presumed nature of the variables . Implica- tions for psychosocial research and clinical practice were dis- cussed. LIST LIST LIST II. III. IV. TABLE OF CONTENTS OF TABLES OF FIGURES OF ABBREVIATIONS INTRODUCTION LITERATURE REVIEW A. Psychosocial Impact of Disease B. Perceptions of Personal Control C. Information and Coping D. Literature Summary and Implications for the Present Study METHOD A. Goals B. Hypotheses C. Subjects D. Measures E. Procedure F. Data Analysis G. Limitations RESULTS AND DISCUSSION A. B. Bivariate Correlations Across Contexts Bivariate Correlations Within Contexts iv PAGE vi viii ix 11 13 16 19 19 20 25 27 34 35 38 41 42 46 TABLE OF C. D. E. CONTENTS - Continued Contextual Specificity of Bivariate Correlations Supplemental Analyses Integration and Conclusions V. SUMMARY AND RECOMMENDATIONS A. B. C. D. E. APPENDIX APPENDIX APPENDIX APPENDIX APPENDIX APPENDIX APPENDIX APPENDIX Summary of Bivariate Correlations Summary of Contextual Specificity Summary of Multivariate Prediction Summary of Age and Gender Effects Recommendations A: Approval to Use Human Subjects B: Request for Study Participants C: Advertisement for Study Participants D: Teenager/Patient Questionnaire E: Subject Consent Form F: Parent/Guardian Consent Form G: Study Summary for Participants H: Variable List and Data File LIST OF REFERENCES 57 58 7O 78 79 83 84 85 88 89 91 92 99 100 102 106 109 TABLE 1. 2. 4. s. 6. 7. 10. 11. 12. 13. 14. 15. 16. LIST OF TABLES Predicted Intercorrelations Among Variables Configurations of Variable Levels That Exemplify All Hypothesized Relationships Subject Characteristics Sample Summary Layout for 2 X 2 ANOVAs Correlations Among Variables Across Contexts Correlations Among Variables In-Context: Appearance Correlations Among Variables In-Context: Peer Relationships Correlations Among Variables In-Context: Future Orientation Correlations Among Variables In-Context: Sexual Functioning Correlations Among Variables In-Context: Freedom/Independence Correlations Among Variables In-Context: School Achievement Correlations Among Variables In-Context: Physical Abilities Correlations Among Variables In-Context: Family Relationships Multivariate Predictors of Psychosocial Disease Impact Across Contexts Multivariate Predictors of Context-Specific Impact: Appearance vi PAGE 26 26 28 29 38 44 44 49 49 53 53 55 55 59 S9 61 LIST OF TABLES (Continued) 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. Multivariate Predictors of Context-Specific Impact: Peer Relationships Multivariate Predictors of Context-Specific Impact: Future Orientation Multivariate Predictors of Context-Specific Impact: Sexual Functioning Multivariate Predictors of Context-Specific Impact: Freedom/Independence Multivariate Predictors of Context-Specific Impact: School Achievement Multivariate Predictors of Context-Specific Impact: Physical Abilities Multivariate Predictors of Context-Specific Impact: Family Relationships Analysis of Variance Across Contexts Analysis of Variance In-Context: Appearance Analysis of Variance In-Context: Peer Relationships Cell Means for Peer Relationships Analysis of Variance In-Context: Future Orientation Analysis of Variance In-Context: Sexual Functioning Analysis of Variance In-Context: Freedom/Independence Analysis of Variance In-Context: School Achievement Analysis of Variance In-Context: Physical Abilities Analysis of Variance In-Context: Family Relationships Summary of Obtained Correlations in Relation to Bivariate Hypotheses vii 61 62 62 63 63 65 65 67 67 69 69 71 71 72 72 74 74 81 LIST OF FIGURES FIGURE PAGE 1 . The Self in Reference to Psychosocially 24 Relevant Effects of Illness or Treatment viii PDI SE IA HLC LIST OF ABBREVIATIONS Psychosocial Disease Impact Self-Efficacy Information Adequacy Health Locus of Control ix I . INTRODUCTION Chronic disease has the potential to significantly compli- cate the psychosocial development of children and adolescents . Prior to the mid-1970 ' s , such complication was often equated with psychopathology . The dominant perspective today, however, is that most disease-related psychological problems are normal reactions to severe stressors rather than symptoms of pervasive mental disorder. Most of the early reports were anecdotal and based primarily on observations of medical patients who had ex- hibited behaviors of significant concern to hospital staff and who had therefore been brought to the attention of mental health professionals. Epidemiological studies in this area now rou- tinely include non-referred medical patients and, to a somewhat lesser extent, persons with unimpaired health. As biases toward pathology have given way to more objective approaches , evidence attesting to the psychological normality of chronically ill children has emerged with compelling consistency (e . g . , Anderson, Asher, Clark, Orrick, & Quiason, 1979: Bedell, Giordani , Amour, Tavormina, & Boll , 1977 : Kellerman, Zeltzer, E1 lenberg, Dash, & Rigler, 1980: Tavormina , Kastner, Slater, & Watt, 1976) . It appears that investigators are becoming increasingly cognizant of the adaptiveness of diverse methods of coping. 1 2 Greater sensitivity to the uniqueness of coping has contributed to the curtailment of a priori evaluations of profound responses to disease (e.g. , denial) . What is important today is to view disease-related challenges from the perspective of normal adolescent development rather than from perspectives in which psychopathology is central . Most studies of disease-related adj ustment problems have focused on adults . Downward extensions of what is known of chronically ill adults are of limited relevance to the situation of the child or adolescent. Fortunately, several frameworks for examining relationships between development and health have recently been proposed (e.g. , Maddux, Roberts, Sledden, & Wright, 1986; Susman, Hollenbeck, Nannis, & Strope, 1980) . Such frameworks have spawned much research on relationships between cognitive development and conceptualizations of health and ill- ness (e.g. , Bibace & Walsh, 1980: Campbell, 1975; Potter & Roberts, 1984: Simeonsson, Buckley, & Monson, 1979: Susman, et a1 . , 1982) . Undoubtedly, a child' s emotional and behavioral reactions to disease and treatment are determined, in part, by his or her ability to understand what is happening . However, few investigators have taken the important additional step of detailing, with data, the effects of serious, chronic illness on the psychosocial tasks associated with particular developmental periods. The determinants of psychosocial impact remain unknown . Previous research has revealed that disease types and person- ality traits account for minimal variance in psychosocial im- 3 pact . Are measures of personality traits simply too broad to be sensitive to specific manifestations of disease-related impact? This possibility seems at least partially responsible for the current trend toward the identification and measurement of narrower variables . Adolescence was chosen as the period for this inquiry because many psychosocial ly relevant aspects of chronic illness (e.g. , changes in appearance, increased dependence, reduced physical abilities , etc.) are of heightened importance to the identity-seeking teenager. A point of significant departure from previous research is that the present study focuses on relationships among variables in each of eight relatively narrow psychosocial contexts integral to adolescent development; these contexts are appearance, peer relationships , future orientation, sexual functioning, freedom/ independence, school achievement, physical abilities, and family relationships. The following variables were selected for study because of their relevance to coping: Psychosocial Disease Impact (PDI) , Self-Efficacy (SE) , Information Adequacy (IA) , and Health Locus of Control (HLC) . Each variable other than Health Locus of Control relates specifically to each of the eight psychosocial contexts (i . e . , each variable is "crossed" with each context) . This study is the first systematic exploration of the contextual specificity of variables related to coping. Of primary interest are the correlations between Psychosocial Disease Impact and the other three variables and the extent to which such correlations are context specific . The two main 4 research questions are: 1 . Are Self-Efficacy, Information Adequacy, and Health Locus of Control significant correlates of Psychosocial Disease Impact in various psychosocial contexts of adolescence? 2 . Is the correlation between Psychosocial Disease Impact and each of the other variables highly context specific or is there a high degree of continuity across contexts? Of secondary interest are answers to the following questions: 3 . Is the variance in Psychosocial Disease Impact that is explained by each predictor variable highly unique or do two or three of the predictors account for essentially the same portion of Impact variance? 4 . Are there significant age effects? 5 . Are there significant sex effects? Although the design of this project is nonexperimental (i . e. , survey/correlational) and therefore incapable of providing direct evidence of causal mechanisms , the resulting correlations are an important foundation for refined hypotheses about the determinants of psychosocial impact . Knowledge of such determinants would be very relevant to clinicians involved in promoting the use of effective coping strategies . For example, if it becomes known that high self-efficacy often strengthens coping (e.g. , increased utilization of disease- related psychosocial information) and thereby reduces psychosocial impact, the clinician would have a much clearer therapeutic focus (i.e. , alleviate self-efficacy deficits) . If the determinants of psychosocial impact are found to vary sig- 5 nificantly from one psychosocial context to another, clinicians would be well advised to develop therapeutic strategies from within a contextually sensitive framework. The origins of the four variables currently under study are traced in Chapter II . Included therein are reasons for the increasing dissatisfaction with the broad trait measures that dominate assessment practice in both clinical and research settings . In response to the need for increased specificity in assessments of coping, three context-specific variables are introduced and operationalized in Chapter III . A fourth vari- able, Health Locus of Control , is relatively broad and serves as a point of contrast to the specificity of the other three variables . A description of the study ' 3 subjects , procedures and data analysis is followed by an account of some of the study ' s most significant limitations . Most of Chapter IV consists of data and discussion relevant to the primary research issues (correlations among variables, contextual specificity) . Supplemental analyses are then presented to address issues of secondary interest (linear combinations, effects of sex and age) . An integrative overview of the results concludes Chapter IV. Chapter V begins with a final look back at the study and closes with suggestions for utilizing the project ' 3 contributions . II . LITERATURE REVIEW Three areas of psychological research have significant bearing on the research questions posed in Chapter I . These areas are psychosocial impact of disease, perceptions of per- sonal control , and information as it relates to coping. It is from these three areas that the four variables currently under investigation (Psychosocial Disease Impact, Self-Efficacy, Information Adequacy, and Health Locus of Control) were derived . The first area reviewed, psychosocial impact of disease, provides the groundwork for the variable Psychosocial Disease Impact . One of the most salient issues that emerges in the dis- cussion of impact is the insensitivity of broad trait measures and the consequent need to define more precisely the contexts in which psychosocial 1y relevant effects of disease occur. In response to this need, three of the four variables in the present study (1 . e . , all variables other than Health Locus of Control) are crossed with each of the following eight psycho- social contexts: appearance, peer relationships, future orientation, sexual functioning, freedom/ independence , school achievement, physical abilities and family relationships. The rationale for selecting these particular contexts concludes the review of impact-relevant literature. The second area reviewed provides a critical account of the 6 7 evolution of theory and measurement of perceptions of personal control . Theoretical and empirical work in the area of personal control provides the foundation for two variables of the current study. One such variable has a broad scope (Health Locus of Control) whereas the other is context-specific (Self-Efficacy) . The final area of relevant research relates to the relationship between information and coping. Perceptions of the adequacy of one ' s current level of disease-related psychosocial information are of particular interest and are represented by the fourth variable, Information Adequacy. A. Psychosocial Impact of Disease The adolescent ' 8 understanding of the nature of various social relations is qualitatively distinct from earlier forms of social awareness (Ellis & Davis, 1982: Selman, 1980) . Hypothesizing about relations between self-perspectives and the perspectives of others requires the capacity for abstract logic which, according to most theories of cognitive development, is generally unavailable prior to adolescence . These abilities enable the adolescent to integrate various perceptions of the self into a psychosocial identity. A well established identity shows continuity between past and present and provides an orientation toward the future. According to Erikson (1963) , failure to achieve an integrated sense of self results in "role confusion. " Various parenting practices and social mil ieus , past and present , greatly affect the adolescent ' s preparedness to succeed in establishing a strong personal identity. Illness- 8 imposed barriers to independence and threats to body image are perhaps greater obstacles in adolescence than in any other period of development (Hofmann, Becker, 8 Gabriel, 1976; Willis, Elliott, & Jay, 1982) . Most investigations of disease impact have focused on broad personality traits with little regard for the developmental con- texts in which such traits are expressed. One important excep- tion is the work of Kellerman et al . (1980) . Three hundred forty-nine healthy adolescents and 168 adolescents with various chronic diseases were compared on measures of trait anxiety, 1 sel f-esteem, and locus of control related to health and illness. These particular traits were chosen because of their apparent relevance to the primary psychosocial tasks of adolescence . No significant differences in trait anxiety or self-esteem were found between the healthy and ill subjects or among the various illness groups. With regard to the health locus of control variable, patients with renal , cardiac, rheumatologic, and oncologic disorders perceived less self-control over health than did healthy peers and subjects with diabetes mellitus or cystic fibrosis . Kellerman et al . (1980) suggested that the patient with diabetes mellitus or cystic fibrosis has a relatively greater degree of direct control over his or her diet, medication intake, and other factors associated with disease management. The greater external ity among patients with cancer or renal , cardiac, or rheumatologic disease appears to represent accurate self-appraisal rather than psychopathology . In a closely related study, Zeltzer, Kellerman, Ellenberg, 9 Dash, and Rigler (1980) administered an 18-item "illness impact questionnaire" to the subjects described above . Healthy sub- jects presumably responded to questionnaire references to illness (e.g. , "Illness hasn 't changed the way I look") based on experiences with colds and other common maladies . There were essentially no differences between chronically ill and healthy respondents in total impact of illness. Among the explanations offered by the authors for this rather striking finding is that for the healthy adolescent the onset of a minor illness disrupts accustomed routines whereas the chronically ill adolescent may de-emphasize all but the most significant manifestations of his or her disease, having become inoculated by its milder presenta- tions . Of course denial must be considered a plausible explanation , particularly with regard to the adolescents who died within a few months after participating in the study. Regardless of its origin, this finding illustrates clearly that knowledge of a patient ' s diagnosis or other index of health status is of very limited value in understanding the psycho- logical mediation involved. Although Kellerman and Zeltzer defined illness impact in terms of developmental tasks , they used only broad classifi- cations (i . e . , disease types and personality traits) in their search for meaningful correlates of impact: context-specific perceptions and attitudes were not assessed. Classifications of disease or personality are too general to be of significant value in predicting how a patient is likely to deal with his or her disease in particular situations . Given that for any 10 particular patient chronic disease may impact self-perceptions in one or more aspects of psychosocial functioning but not in others , it may be more informative to study impact (and poten- tial correlates of impact) in terms of relatively narrow psychosocial contexts . As Erikson (1963) suggested, the primary developmental tasks of adolescence are those which relate to the estab- lishment of a psychosocial identity. The pursuit of an identity is a process that usually involves establishing independence from parents , adjusting to sexual maturation, forming mature relationships with peers of both sexes, and preparing for the future . The developmental tasks most relevant to identity formation can, for the sake of conceptual clarity, be cate- gorized in terms of the psychosocial contexts to which they correspond. Eight such contexts appear to encompass nearly all aspects of psychosocial functioning relevant to adolescent identity formation. These contexts are: appearance, peer rela- tionships , future orientation, sexual functioning, freedom/ independence, school achievement, physical abilities, and family relationships . Although various theorists and researchers have emphasized different subsets of this group of psychosocial contexts , most in-depth accounts of adolescent development reflect the authors ' acknowledgement of the importance of most , if not all, of these eight contexts (e.g. , Erikson, 1959: Havighurst, 1951; Kimmel & Weiner, 1985: McKinney, Fitzgerald, & Strommen, 1982) . The developmental tasks which correspond to these contexts are no less relevant to adolescents with chronic ll diseases (e.g. , Blumberg, Lewis, & Susman, 1982: National Institutes of Health, 1984: Pfefferbaum & Levenson, 1982: Spinetta et al. , 1982: Zeltzer et a1. , 1980) . B. Perceptions of Personal Control There is now a trend toward redefining broad constructs , such as locus of control , in terms of narrower dimensions. The recent development of health locus of control scales (e. g. , Parcel & Meyer, 1978: Wallston, Wallston, & DeVellis, 1978) reflects this more refined focus . However, these scales deal with perspectives of health-related personal control only in very general terms. The following items are from the Multi- dimensional Health Locus of Control scales (Wallston et al . , 1978) : "If I take care of myself, I can avoid illness. " "My good health is largely a matter of good fortune. " "I am in control of my health . " From what frame of reference would a person with an intense , chronically obtrusive disease respond to such items? Although the health locus of control scales that have been designed to date primarily address preventative health behavior, they have been used to assess the perceived personal control of individuals with serious chronic diseases (e.g. , Felton & Revenson, 1984: Nagy s. Wolfe, 1983) . Evidence for the validity of such applications has yet to be reported. Rotter (1954: Rotter, Chance & Phares, 1972) developed the social learning theory on which most locus of control scales are based . He postulated that generalized expectancies are most likely to be relied upon in situations that are novel or 12 ambiguous . It would seem that a person who has had a serious disease for several years is likely to have health-related expectancies that have been shaped by experience with the disease. It follows that, because of their greater relevance, experiential ly based expectancies would take precedence over general predispositions . As Strickland (197 8) noted, "the severity of the disorder, the time of onset, the current status of the patient, the support that he/ she receives, and so on, all interact with what is probably a complex set of cognitions about the disorder" (p. 1198) . The amount of variance in health- related behavior that can be explained by locus of control alone is unimpressive (Strickland, 1978: Wallston et al. , 1978) . There appears to be a consensus that in order to accurately predict the health-related behavior of a chronically ill person in a particular circumstance, influences specific to the situation must be taken into account . One approach would be to determine general predispositions using a locus of control measure, then generate hypotheses as to the impact of contextual factors on those general tendencies . This method would involve a great deal of speculation and consequently would yield imprecise predictions . A much more direct method would be to assess the patient ' s perceptions of self-efficacy in regard to each context of interest. According to Bandura ( 1977) , "an efficacy expectation is the conviction that we can successfully execute the behavior required to produce the outcomes" (p. 193) . Efficacy expectations are derived largely from the person's experiential 13 history, current interpersonal influences, and physiological state. Bandura (197 7) hypothesized that "expectations of per- sonal efficacy determine whether coping behavior will be initiated, how much effort will be expended, and how long it will be sustained in the face of obstacles and aversive experiences" (p. 191) . Tests of the relationship between expectancy and performance are hindered whenever expectancies are assessed globally as if they were static and unidimensional . As Kendall and Korgeski ( 1979) noted, "a specific self-efficacy instrument would need to be devised for each behavior under investigation” (p. 11) . Given the attitudinal lability that characterizes the struggle for an identity, specificity in expectancy assessment is particularly critical when dealing with adolescents. C . Information and Coping According to Lazarus and Folkman (1984) , coping refers to the ”process of managing demands (external or internal) that are appraised as taxing or exceeding the resources of the person" (p. 283) . One of the most significant aspects of this perspec- tive is the emphasis on processes (i . e . , coping strategies) rather than on styles or traits . An important implication of this is that a method of coping which is helpful in a particular context at a particular time may be less beneficial or even maladaptive on other occasions or in other settings . Although the relative adaptiveness of any strategy cannot be fully determined independent of the context in which it is applied, 14 some methods are generally more adaptive than others . One strategy that is common to most theories of coping and that has frequently been associated with positive outcome is "information seeking” (e.g. , Cohen, 1980: Moos & Tsu, 1977; "Proj ect on Coping, " 1983) . In studies of pediatric popula- tions , disease-related information has been associated with decreased anxiety (Katz , Kellerman, & Siegel , 1980: Pfefferbaum, Levenson & van Eys , 1982) and improved compliance with medical advice (Francis , Korsch, 5. Morris, 1969) . However, such associ- ations have been’ weak. In a study of hospitalized adults , Cohen and Lazarus (1973) found that a preoperative measure of a trait believed to be related to information needs , "repression- sensitization, " failed to predict coping behavior or patient recovery. Both the desire for additional information and the effects of such information on behavior and psychological state appear to be mediated by situation-specific influences . Willis et al . (1982) suggested that our understanding of the role of disease-related information might be enhanced significantly by addressing the patient ' s "perceived sel f—ef f icacy in being able to apply that knowledge" (p. 43) . Self-efficacy and information seeking are significantly interdependent; information facilitates competence in dealing with daily tasks which then increases the expectancy that additional information would be advantageous . Of course, there are circumstances in which additional disease-related information would be incapable of increasing competence and to 15 believe otherwise would be merely wishful and perhaps maladaptive. However, it would seem that only a small minority of the population of chronically ill adolescents would be in such grave straits at any given time . Information seeking may be particularly important during adolescence because it facilitates self-suf f iciency and a sense of independence. Parents, even more than their ill adolescents, may feel a need for increased patient knowledge and independence (Levenson, Copeland, Morrow, Pfefferbaum, & Silberberg, 1983) . Perhaps it is because adolescents who are aggressively involved in their own treatments seem to survive longer (Achterberg, Matthews-Simonton, & Simonton, 1977) , that most studies of self-help have focused on treatment issues . This focus also seems to parallel the information preferences of the adolescents themselves . In a study of 55 adolescent cancer out- patients, Levenson et al . (1983) found the need for disease- related information to be greatest in relation to the disease itself and least with regard to implications of the illness in terms of feelings, appearance, and relationships. In a re- lated study, Pfefferbaum and Levenson (1982) described the information preferences of 63 adolescent oncology patients . Although information directly related to the disease and its treatment was given the highest priority, there was also a strong indication of need for information in psychosocial areas . For example, the following questionnaire items were rated as either "important" or "extremely important" (rather than "slightly important" or "unimportant") to the extent indicated 16 by the corresponding percentages: "The effects of cancer and its treatment on current appearance" (73%) : "The effects of cancer on future appearance" (77%) : "The kinds of physical activity that can be done" (79%) : "How cancer will affect future family role" (68%) : "Ways to talk to friends and relatives about my illness" (46%) . The need to address the adolescent ' s psychosocial concerns becomes acute when such concerns interfere with medical treat- ment (Kellerman & Katz, 1977) . For example, in interviews with nine adolescents who refused cancer treatment , the primary reasons given for refusal were : fear of how they would look and fear of friends ' reactions (Blotcky, Cohen, Conatser, & Klopovich, 1985) . D. Literature Summary and Implications for the Present Study Three areas of research relevant to the present study have been discussed: 1. Psychosocial impact of chronic illness during adolescence, 2 . Perceptions of personal control , and 3 . Disease-related information and coping . The first section described aspects of adolescent develop- ment which affect subjective responses to chronic illnesses and related medical treatments . Other than studies of cognitive development as related to conceptualizations of health issues, the available literature offers few empirical accounts of how disease impact is mediated by developmental processes . Per- spectives which emphasized typologies of illness and personality were found to add little to the understanding of the psycho- 17 social concomitants of illness. The literature reviewed in the second section outlined the evolution of the theory and measurement of perceptions of per- sonal control . From its broad roots in Rotter ' 3 social learning theory, locus of control has been refined to address particular domains such as general perceptions of personal control over one ' s own health. However, "health" is such a broad content area that general measures of health control have provided little insight into behaviors that occur in specific, disease- affected contexts . Bandura ' s (1977) seminal paper on self- efficacy has prompted calls for context-specific investigations of perceptions of personal control with pediatric patients . Such applications have yet to appear in the literature . The final section of the literature review addressed the relationship between disease-related information and coping. Information seeking was found to be an effective coping strategy for many patients . It appears that both information needs and the effects of such information often relate to specific mani- festations of the illness and the psychosocial contexts in which they are most problematic. Published surveys indicate that chronically ill adolescents have strong needs for psychosocial information which takes into account the effects of their illnesses. In all three areas (impact, personal control , infome- tion) there is a very salient trend toward increased recognition of the influence of contextual factors. The primary impetus for this more refined view is the need to predict behavior more 18 precisely. Lazarus and Folkman (1984) cogently noted the need for greater sensitivity to the contextual specificity of coping: Quite clearly, we need to identify the situational context of the coping process if we are to make advances in our understanding. . . .This means that we should be centering our assessments of coping, at least some of the time, on actual , specific encounters (p. 307) . This trend is much more evident in conceptual rather than methodological aspects of recent work: broad trait measures continue to dominate assessment practice . The present study is distinct from previous work in terms of the contextual speci f i- city with which disease-relevant psychosocial variables are measured and analyzed. This proj ect provides the first true test of the suspected advantage of context-specific measures over trait measures as correlates of psychosocial disease impact . III . METHOD A. Goals The goals of this study are as follows: 1 . To estimate correlations among four major variables relevant to the psychosocial perspectives of chronically health- impaired adolescents . These variables are outlined below and operationalized (transcribed into rating scales) in Section D of this chapter. WW (PDI) refers to disease- induced changes in how the self is regarded in each of eight psychosocial contexts commonly believed to be integral to adolescent development. These contexts are: appearance, peer relationships, future orientation, sexual functioning, freedom/ independence, school achievement, physical abilities, and family relationships . 521131112991 (SE) refers to perceived ability to derive benefit from disease-related information in each of the afore- mentioned psychosocial contexts . Self-efficacy in terms of disease-related information is a highly specific derivative of the self-efficacy construct elaborated by Bandura (1977) . 1W (IA) refers to perceived sufficiency of current disease-related knowledge in each of the psychosocial contexts. 19 20 WW (HLC) refers to the extent to which one believes his/her health to be under sel f-control . This variable relates to personal control over health generally (e.g. , prevention, recovery, etc.) , rather than in specific disease or psychosocial contexts. 2 . To provide substantive interpretations of the observed intercorrelat ions based on the presumed nature of the variables . B. Hypotheses The first five hypotheses address the first primary research question asked in Chapter I: Are Sel f-Ef f icacy, Information Adequacy and Health Locus of Control significant correlates of Psychosocial Disease Impact in various psycho- social contexts of adolescence? Hypothesis 6 addresses the second primary research question: Is the correlation between Psychosocial Disease Impact and each of the other variables highly context specific or is there a high degree of continuity across contexts? The secondary research questions asked in Chapter I (regarding multivariate prediction of Impact and the effects of age and sex) are addressed separately in the Supple- mental Analysis section of Chapter IV. Each formally stated hypothesis is preceded by a brief substantive description of the relationship (s) involved . 1 . Subjects who have high Sel f-Ef f icacy (believe that they could make good use of disease-related information) are likely to have used information effectively in the past, thereby lowering their 2 1 Psychosocial Disease Impact. Subjects with low Self- Efficacy are likely to have high Impact as a conse- quence of not using psychosocial information in the past. This relationship will be expressed as a high negative correlation between Sel f-Ef f icacy and Psychosocial Disease Impact. Hypothesis 1: A significant, negative correlation will exist between Psychosocial Disease Impact and Self- Efficacy. The relationship between Sel f -Ef f icacy and Psycho— social Disease Impact stated in Hypothesis 1 will be stronger than any other hypothesized relationship . Hypothesis 2: The magnitude of the negative correlation between Psychosocial Disease Impact and Sel f-Ef f icacy will be greater than the correlation between Psychosocial Disease Impact and any other variable. Subjects who have adequate levels of disease-related information are likely to have benefited from such information in terms of Psychosocial Disease Impact. Subjects with very inadequate levels of disease- related information have had little information to bring to bear on the psychosocially relevant effects of their diseases. This relationship will be 22 expressed as a moderate, negative correlation between Information Adequacy and Psychosocial Disease Impact . Hypothesis 3: A significant, negative correlation will exist between Psychosocial Disease Impact and Informa- tion Adequacy . Subjects who perceive a great deal of control over their health in general will tend to have relatively low Psychosocial Disease Impact whereas subjects who perceive little control over their health will tend to have relatively high levels of Impact. This relation- ship will be expressed as a low, negative correlation between Health Locus of Control and Psychosocial Disease Impact. Hypothesis 4: A nonsignificant negative correlation will exist between Psychosocial Disease Impact and Health Locus of Control . Subjects who have high Sel f-Ef f icacy will tend to have high (internal) Health Locus of Control (and vice versa) . Subjects who have high levels of Information Adequacy will also tend to show high levels of Health Locus of Control . These relationships will be expressed as low, positive correlations . Hypothesis 5: A nonsignificant positive correlation will exist 23 between: (a) Health Locus of Control and Self— Eff icacy: (b) Health Locus of Control and Information Adequacy . 6 . The relationship between Sel f-Ef f icacy and Psycho- social Disease Impact will vary significantly from one psychosocial context to another. The same will be true for the relationship between Psychosocial Disease Impact and the other context-specific variable , Information Adequacy. This variability will be evi- denced by significant differences between the eight psychosocial contexts in terms of the correlations which involve Psychosocial Disease Impact. Hypothesis 6: There will be significant variability between psychosocial contexts with regard to bivariate correlations which involve Psychosocial Disease Impact. The relationship between Psychosocial Disease Impact and the other three variables is depicted in Figure 1 . The width of each arrow represents the corresponding variable ' s influence relative to the other identified influences . The amount of influence exerted by unexamined variables is unknown, as indicated by the undefined lines of the arrow at the right of the figure . 24 . The self in reference to psychosocially relevant effects of illness or treatment. The opening at the figure ' s apex illustrates the point at which psychosocial impact dissipates or increases in accord with gains and pitfalls in the coping process . Certainly the most desirable outcome is decreased psychosocial impact as coping progresses over time. However, even in such an adaptive scen- ario, situational influences, such as disease state or availa- bility of social support, are likely to fluctuate and alter the balance between challenges and resources . Therefore , the extent 25 to which the self is impacted by psychosocially relevant aspects of the illness will vary over time: the curved line of Figure 1 would vacillate accordingly . The hypotheses are summarized in Table 1 . Configurations of variable levels for hypothetical cases consistent with all hypotheses are exhibited in Table 2 . C . Subjects The sample consisted of 30 adolescents, ages 13 through 18 years , each of whom had been diagnosed as having a serious chronic disease . Because of the great variability throughout the psychosocial literature in criteria used to define adolescence, the age span of 13 through 18 years was chosen " as the period to describe the adolescent experience, rather than as a method to encapsulate all of adolescence" (Ellis & Davis, 1982 , p. 698) . The only other criterion for inclusion was the absence of diagnosed or suspected mental retardation. Subject recruitment began soon after the study proposal was approved by the University Committee on Research Involving Human Subjects at Michigan State University (Appendix A) . Since the investigator was not affiliated with a health care facility, subjects were recruited from diverse sources . A standard letter (Appendix B) requesting assistance in securing subjects was distributed to 65 physicians in the northwestern United States. The medical specialties represented in the physician group included: pediatrics, adolescent internal medicine, hematology, oncology, rheumatology, neurology, and 26 Table 1 r i 1 Variable SE IA HLC 1 PDI H (-) M (-) L (-) 2 SE -- M (+) L (+) 3. IA -- L (+) 4 HLC - EQIQ- H=high (_>_.60); M=moderate (30-59); L-—-low (5.29). Table 2 WW 1] H l . I E l . l . Hypothetical Case Variable 1 2 1 PDI L H 2 SE H L 3. IA M-H L-M 4 HLC M M Note. H=high; M=moderate; L=low. 27 cardiology. Physician referrals accounted for seven study participants . An advertisement (Appendix C) in the January, 1988 issue of the W, published by the Oregon chapter of the National Cancer Society, yielded 10 sub- jects . The remaining 13 subjects were located with assistance from various professionals in social work, psychology, and education. Subject characteristics are profiled in Table 3 and summarized in Table 4 . D. Measures The data-gathering instrument developed for this study con- sists of 42 items in questionnaire format (Appendix D) . The introduction to the "Teenager/ Patient Questionnaire" instructs the subject to respond in terms of the six-point continuum of relative agreement that appears below each statement . This instrument consists of four distinct scales . Three of the four scales originate in this project. Each item of these three scales is framed in terms of one of the eight psychosocial contexts (i . e. , one item per context per scale) . The main interest is on the relationships among the variables within each context, rather than on relationships among the variables across contexts . For example, self-efficacy across contexts is of much less interest than sel f-ef f icacy with regard to appearance , sel f-ef f icacy with regard to peer rela- tionships, self-efficacy with regard to physical abilities , etc. The many psychometric "unknowns" and other limitations inherent in one-item measures are discussed in Section III G. Table 3 S l . Q r . . Subject Sex Age State Illness 1 M 17 CA Lymphoma,Leukemia 2 M 15 CA Diabetes mellitus 3 M 17 CA Diabetes mellitus 4 F 18 OR Thyroid cancer 5 F 15 OR Leukemia 6 M 15 CA Leukemia 7 M 13 OR Leukemia 8 F 13 CA Leukemia 9 M 17 CA Seizure disorder-mixed 10 F 16 CA Neuroblastoma 1 1 F 16 CA Muscular dystrophy 12 F 17 CA Leukemia 13 F 13 OR Diabetes mellitus 14 M 14 OR Lymphoma, Leukemia 15 M 16 CA Osteogenesis imperfecta 16 M 15 CA Spina bifida 17 F 14 CA Leukemia 18 M 18 CA Leukemia 19 F 18 CA Leukemia 20 F 15 CA Leukemia 21 M 15 MI Muscular dystrophy 22 M 13 MI Muscular dystrophy 23 M 14 OR Rhabdomyosarcoma 24 M 16 OR Hodgkin’s disease 25 F 13 OR Leukemia 26 M 16 OR Leukemia 27 M 13 CA Torsion dystonia 28 F 15 CA Friedreich’s ataxia 29 M 15 WA Ewing’s sarcoma 30 F 13 OR Osteogenic sarcoma 29 u-a a?! O .5 a PSQMPPNPEPPNEEPNPENE‘E’ Male Female 13, 14 15, 16 17, 18 California Michigan Oregon Washington Cancer Diabetes mellitus Friedreich’s ataxia Muscular dystrophy Osteogenesis imperfecta Seizure disorder Spina bifida Torsion dystonia 17 13 10 l3 17 10 19 Hut—no: p—ag—ap—L £91m 56.7 43.3 33.3 43.3 23.3 56.7 6.7 33.3 3.3 63.3 10.0 3.3 10.0 3.3 3.3 3.3 3.3 30 However, one psychometric property , " content val idity" was given particular priority in the construction of the measures of Psychosocial Disease Impact, Self-Efficacy, and Information Adequacy and therefore warrants consideration here . "Content validity is built into a test from the outset through the choice of appropriate items” (Anastasi, 1968 , p. 100) . According to Salvia and Ysseldyke ( 1985) , content validity is evaluated not only by examining the appropriate- ness of the items ( i. e. , Does this item really tap the domain of interest?) but also by determining the completeness of the item sample . Certainly a one-item measure of a personality trait such as locus of control would fall far short on the complete- ness criterion . Even constructs which are much narrower in scope , such as general sel f-ef f icacy, require many appropriate items to ensure adequate representation. However, the domains of interest in the present project are so narrow (i . e. , con- strained by both relevance to the subject ' s disease and rele- vance to a particular psychosocial context) that large numbers of items are not necessary to ensure adequate domain sampling. Such domains include the impact of one ' 8 disease on his or her appearance , the extent to which one believes that he or she could make use of disease-related information to improve family relationships , perceived adequacy of one ' 3 current level of disease-related knowledge about maximizing his or her school achievement , etc . 31 W. This scale relates to disease-induced changes in how the self is regarded in each of eight psychosocial contexts of adolescence. These contexts are: appearance, peer relationships, future orientation, sexual functioning, freedom/ independence , school achievement, physical abilities, and family relationships. The order of all questionnaire items was determined via a randomization process . The items that compose the Psychosocial Disease Impact scale are: 1, 18, 23, 27, 29, 32, 35, and 36. The 18-item "Illness Impact Questionnaire" developed by Zeltzer, et al . (1980) is the only other instrument which addresses the impact of disease on particular aspects of adoles- cent psychosocial functioning. Each of seven of the eight psychosocial contexts examined in the present study (i . e. , other than physical abilities) is represented by one or more items in the Illness Impact Questionnaire. In the interest of optimal content validity and parallel form among the three context- specific measures , the Illness Impact Questionnaire served only as a very general prototype in the development of the Psycho- social Disease Impact items. Selfflfflm. This scale relates to perceived ability to derive benefit from context-specific disease-related information. The items that compose the Self-Efficacy scale are: 4, 11, 12, 13, 19, 25, 30, and 39. As Strickland (1978) and others have noted, assessing perceptions of personal control broadly, as in the case of locus of control scales, tells us little about actual behavioral 32 functioning. Using locus of control scales which relate to a particular aspect of functioning, such as health, is a step toward greater specificity yet still provides little information about how one copes with disease manifestations (Wallston et al. , 1978) . Although Bandura (1977) argued that self-efficacy is a personal control construct specific enough to explain observed coping behavior, no such applications have been reported. Given that acquisition of disease-related information is often an important aspect of the coping process (Moos & Tsu, 1977) , self-efficacy in the current project was framed in terms of disease-related psychosocial information. Since there were no prototypical disease-related sel f-ef f icacy items to rely upon, the tenets of content validity and the interest in parallel form were given priority in the development of the Sel f-Ef f icacy items . The close correspondence between each item and the domain of interest (disease-related, context-specific sel f-ef f icacy) is clearly evident in the structure common to all Self-Efficacy items: "I could make good use of information about , despite the effects of my illness (or treatment) . " W. This scale relates to perceived adequacy of current level of disease-related knowledge in each psychosocial context. The items that compose the Information Adequacy scale are: 2 , 3, 10, 14, 21, 26, 33, and 38. As with the other context specific measures , Information Adequacy items are logical derivatives of the construct of interest and have no empirical history in the psychological literature . The content validity and parallel form of the 33 Information Adequacy items are apparent in the structure common to all Information Adequacy items: "I currently have adequate information about , which takes into account the effects of my illness (or treatment) . " W1. This scale relates to perceived self-control over health in general . The Health Locus of Control scale is actually Form A of the Multidimensional Health Locus of Control scales (Wallston et al . , 1978) . The items that compose the Health Locus of Control scale are: 5, 6 , 7 , 8 , 9 , 15 , 16, 17, 20, 22, 24, 28, 31, 34, 37, 40, 41, and 42. The only psychometric data available for the Health Locus of Control scale are those reported by its developers , Wallston, et al. (1978) . However, the Wallston et al. (1978) study is replete with methodological limitations and is focused primarily on adults. To illustrate, 354 persons, ages 16 and above, waiting at gates in a Nashville airport were asked to complete and return a booklet that contained 115 items from which the Health Locus of Control Scales would later be derived. Eighty percent of those approached accepted the booklet and 44% of those individuals returned them. Of those returning the booklet, the mean age was 42 years, 74% had at least some college education, and most were residents of Tennessee. In summary, the Health Locus of Control scale has yet to receive validation beyond that which can be inferred from the seemingly close match between item content and the construct purportedly measured. According to a method developed by Fry (19 68) , the 34 questionnaire is approximately at the 7th grade level of reada- bility. E . Procedure Upon receipt of the name and telephone number of each pro- spective subject (from one of the referral sources described in Section III C) , an introductory telephone call was made to the adolescent ' 8 parent or guardian. A brief description of the study was provided, with an emphasis on what participation would entail . The parent or guardian was informed that the question- naire deals with potentially sensitive topics such as sexual functioning, family relationships, and appearance. If the parent/guardian indicated a willingness to allow the adolescent to participate, a similar introduction was provided to the potential subject. In 12 instances, the study was introduced by someone other than the researcher: each representative had received instruction in how the introduction was to be handled. All of the referred adolescents and their parents or guardians indicated a willingness to examine the participant packet . This packet consisted of a consent form for the adolescent (Appendix E) , a consent form for the parent or guardian (Appendix F) , the questionnaire, and a postage paid return envelope . These materials were sent to the parent or guardian to be reviewed and approved before being passed on to the adolescent . To encourage participation and rapid return of the consent forms and questionnaire, a ten dollar incentive was offered to all but the first two participants (i. e. , it was initiated as 35 soon as the need became apparent) . The incentive check was mailed within one day of the return of the completed study materials. All adolescents who were asked to participate completed and returned the questionnaire . Each subject ' s questionnaire was complete: there were no omissions or ambiguous responses . Subjects received a "debriefing" telephone call soon after participating in the study. The purposes of the call were to thank the subject, provide "appropriate reassurance. . . about nor- normal reactions, including disappointment" (APA, 1982 , p. 66) , and to detect any adverse effects of having participated in this project . A note relevant to these matters was sent to subjects who could not be contacted by telephone: the option of placing a collect call to the researcher was included. Subjects were again contacted by telephone (or mail) approximately one month after participation to ascertain whether or not participation-related effects existed. There was no in- dication whatsoever of any adverse participation-related effect . Each subject will receive a brief summary of the project (Appendix G) when it is complete. F . Data Analysis Prior to analysis, the ratings of 12 of the 18 Health Locus of Control items (5, 7, 9, 16, 17, 22, 24, 28, 31, 34, 41, 42) were reversed so that all high ratings would correspond to "internal” control whereas all low ratings would correspond to "external" control . The primary analysis consists of (a) 36 bivariate correlations conducted within each psychosocial con- text and across contexts, and (b) a test of the contextual specificity (homogeneity) of the bivariate correlations which involve Psychosocial Disease Impact. When the population correlation rho (9) has a value other than zero, the bivariate distribution is skewed. With p# 0 , the smaller the sample size, the greater the skewness. Given the relatively small sample size in the present study (N=30) , the direct use of r in normal distribution tests of Ho:,o= O perhaps would have produced misleading results . Consequently , R. A. Fisher' 3 (1915) logarithmic transformation of r was used in testing each Ho: ,0 = 0 , since the sampling distribution of this statistic (Zr) is approximately normal irrespective of sample size or the value of r. For each obtained value of r, the following test was conducted: where 37 To test the contextual homogeneity of the bivariate corre- lations involving Psychosocial Disease Impact (e.g. , PDI/SE in appearance vs. PDI/SE in peer relationships) , the following statistic, as described by Olkin (1967) , was computed. This test was conducted for all possible context comparisons (28) for each of the two pairs of correlated variables of interest . finfizfi- r34) - (42 #53; 6712334 whergz ’ 64 ’ 0 A _ 3-22-22 (7712 r34 (1 r12) +(l r34) +r +2Er 2 2 2 2 12’34Er13+r14+‘23+r24J 13rz4+r14r23J _V ‘ZE”12’131'14Ir12r23r24l'r13‘23r34l'r14r24r343 Several supplemental analyses were conducted. The corre- lations between Psychosocial Disease Impact and various linear combinations of the other three variables (1 . e. , multiple R) were explored in each psychosocial context and across contexts . A two-factor (2 x 2) analysis of variance (ANOVA) was con- ducted with each of the four variables to detect effects of age, sex, and the interaction of age and sex. ANOVAs were conducted both within and across psychosocial contexts . Table 5 illus- trates the layout of cases in each ANOVA. All correlations and ANOVAs were performed using Psychostat-3 , release 2 . 1 for IBM. 38 The variable list and data file are in Appendix H . Table 5 W Age 13-15 16-18 Male n = 10 n = 7 Sex Female n - 8 n - 5 G . Limitations The most significant limitation of the present study is the relatively small size of the sample. Results based on 30 sub- jects are considerably less stable or conclusive than results based on large samples. A second major limitation involved the nonrandom nature of the subject selection process . This limitation is most costly in terms of inferential aspects of the obtained results. The pool of potential participants was limited by factors other than the research requirements . For example, several referral sources , particularly physicians , indicated that they knew of chronically ill adolescents who would not be approached for possible participation due to the grave nature of their current health status. Additionally, all subjects were volunteers, some of whom cited the ten dollar incentive as the primary reason for their participation. Volunteerism and variables related to financial need may have , in some unknown way, biased the 39 selection process . Perhaps the obtained results are represen- tative only of persons very similar to study participants in terms of other unmeasured variables , such as health status or intelligence. Three of the study ' 3 four scales originate in this project . A fourth measurement tool , the Health Locus of Control scale, has been the focus of very little prior research. The rel i- ability of these instruments is unknown. Given that Psycho- social Disease Impact, Self—Efficacy and Information Adequacy were measured by only one item per psychosocial context , a reliability coefficient could not be computed . (Only one type of reliability coefficient, a stability coefficient, can be computed for one-item measures: the necessary readministration of the items was not included in this project. ) Without a reliability coefficient, it is not possible to compute the standard error of measurement or to estimate true scores . For a test to be valid, it must be reliable. The instruments used in the present study may or may not be sufficiently reliable . The extent to which each of these instruments measures its parent construct is open to debate. Content validity, the concept most prominent in the development of the items in this study, is judgmental in nature (Salvia & Ysseldyke, 1985) . Items were not pilot tested prior to inclusion in the study. The absence of empirical validity studies necessitates parti- cular caution when describing relationships , even correlational relationships , among the variables purportedly measured . Inter— 40 pretations offered for the observed correlations are therefore particularly tenuous . There was no objective measure of the presence or absence of psychosocially relevant effects of disease or treatment. Low ratings on coping variables could, for example, simply indicate that there were no psychosocial challenges with which the sub- ject had to cope . The fit between the type of data and the type of analysis is less than ideal . Variables to be correlated or used (as de- pendent variables) in ANOVAs must be continuous . Although there is no one definitive distinction between discrete and continuous variables, the range of possible values at the context level (1- 6) only marginally qualifies as continuous . The final matter concerns the distinction between statis- tical significance and differences (between genders or age groups) that are of practical or theoretical use . This distinc— tion is particularly important given the large number of ANOVAs conducted (28) with a lenient level of alpha (. 05) and the consequent likel ihood of observing one or two instances of statistical significance due to chance . IV. RESULTS AND DISCUSSION Obtained correlations which relate to the hypotheses will be discussed at the "across contexts" level , then within each psychosocial context. These two levels of bivariate analysis are followed by an examination of the contextual specificity of the bivariate correlations which involve Psychosocial Disease Impact. Questions of secondary interest (regarding multivariate prediction of Impact and the effects of age and sex) are addressed under a separate heading, Supplemental Analyses . The chapter concludes with a section devoted to integration and derivation of tentative conclusions . First, however, a few comments regarding the classifica- tion of correlation coefficients are warranted. Substantive descriptions of the hypotheses were framed in terms of "high, " "moderate, " or "low" correlations. There is great diversity among various classifiers in the use of these categories . Such diversity is warranted because the evaluation of any given correlation is bound by the nature of its constituent variables . To illustrate this relativity, Blommers and Forsyth ( 197 7) suggested that coefficients as high as . 5 between measures of a physical trait and a mental trait are extremely rare whereas correlations of this magnitude between reliable measures of two mental traits are quite common and would be considered only moderate. Similarly, a correlation of . 9 between two indepen- ' 41 42 dent measures of the same trait, as in the case of alternate forms of an academic achievement test, might be considered only moderate or perhaps even low. Given the nature of the variables currently under study, the adjective high will be applied to correlations of . 60 or greater, moderate will denote correlations of . 30 to . 59 , and coefficients of . 29 or less will be regarded as low. The lower end of the high category accounts for more than one-third (36%) of the total variance, whereas the upper limit of the low category accounts for less than one tenth (8 . 4%) of the vari- ance. Values considered high are, without exception, statisti- cally significant (given N = 30) whereas low values are not. Values near the boundaries of these somewhat arbitrarily defined categories will be qualified accordingly (e . g. , moderately high). A. Bivariate Correlations Across Contexts Bivariate correlations based on data aggregated from all psychosocial contexts are displayed in Table 6 . Using the classification system described above, none of the hypotheses received support at the across contexts level . However, if a less categorical approach is taken, the relationship between Information Adequacy and Impact implicit in Hypothesis 3 (a moderate, negative correlation) , received strong affirmation (r= -. 709) . Perhaps adolescents who have illnesses with relatively minor psychosocial 1y relevant concomitants simply have little need for disease-related psychosocial information 43 (i . e. , they consider their current knowledge sufficient) . This pattern is most clearly evident in the case of Subject 3 . This 17-year-old male with very well controlled diabetes mellitus reportedly is very popular, athletic, and in his school ' 8 Gifted and Talented program. He gave a rating of one to every Impact item and a rating of six to every Information Adequacy item. Even persons such as Subjects 4 , 6 , and 28 with more serious illnesses (thyroid cancer, leukemia, and a degenerative muscle disorder, respectively) gave low ratings (average 5 2) to Impact items and high ratings (average 2 4) to Information Adequacy items . Perhaps the reason that these individuals currently experience a low level of impact is because they have ample information about how to cope with the psychosocial ly relevant aspects of their illnesses. The high, inverse relationship between Impact and Information Adequacy appears to characterize chronically ill adolescents independent of variability in objective health status . The strongest relationship observed was between Impact and Health Locus of Control (r= -. 769) . Clearly, adolescents with a more "internal” focus perceived less impact relative to their more "external" counterparts . This relationship was much stronger than that which was anticipated in Hypothesis 4 (a low, negative correlation) . Although objective health status was not systematically assessed, this high, negative correlation appears to hold true for the terminally ill adolescent as well as for those with well controlled illnesses. The magnitude of this Table 6 Cl'! 1,.”3 C Variable SE IA HLC 1. PDI .412(2.275)"' -.709(-4.588)“"" -.769(-5.273)“'" 2. SE --- -.389(-2.127)"' -.355(-l.928) 3. IA -- .676(4.270)*" 4. HLC --- Note; Values of z corresponding to r are in parentheses. *p<.05, "*p<.001 Table 7 Commons Among yan'amgs Ill-comm; Appearaggg (AB) Variable SE IA HLC 1. PDI .355 (1.928) -.714(-4.652)"* -.546(-3.184)** 2. SE -- -.457(-2.564)* -.586(-3.490)"* 3- 1A .666(4.l75)*** 4. HLC Note; Values of z correspondin to r are in parentheses. *p<.05, """p<.01, "' *p< 1 45 coefficient is somewhat surprising given the ambiguity inherent in interpreting the general and primarily prevention focused Health Locus of Control items ( i .e. , "If I take care of myself, I can avoid illness") from the perspective of someone with a major, perhaps life threatening illness . The mother of Subject 30 described on the questionnaire how her daughter responded to this ambiguity: Jenny was confused about how to answer the questions dealing with her control over her illness . She believes her cancer was a matter of chance but that she has control over minor illnesses and over the effects of treatment . The positive correlation between Information Adequacy and Health Locus of Control (. 676) was much stronger than anticipated in Hypothesis 5b (a low, positive correlation) and may be the product of bidirectional influences . A strong sense of personal control is likely to promote information acquisition which then may increase the person ' s perceived competence . The moderate, positive correlation between Impact and Sel f - Efficacy ( . 412) is the result in sharpest contrast to a hypothe- sized relationship (i.e. , a high, negative correlation) . The low Sel f-Ef f icacy ratings of many subjects may have been a reflection of minimal need for disease-related psychosocial information. Similarly, an adolescent confronted with signifi- cant psychosocially relevant manifestations of a disease may be highly motivated to seek and apply relevant information so as to improve his or her circumstances . The fact that sel f-ef f icacy 46 tended to correlate positively with Impact and negatively with Information Adequacy suggests the possibility of such a rela- tionship. However, within the sample is a subset of subjects which produced responses more consistent with the hypothesized relationships. Included in this subset are Subjects 4 and 6 who, despite low Impact ratings , had relatively high Self- Efficacy ratings . Chronically ill adolescents such as these may have used information effectively in the past and are now confident about being able to apply additional information productively. This notion is evident in comments Subject 24 wrote on his questionnaire: When the phrase, "I could make good use of " , is used, I feel that any additional information could be used. The same is true about the phrase , "I currently have adequate information" . The information may be adequate but more is always being looked for, used, and appreciated when found . B. Bivariate Correlations Within Contexts Appearance Bivariate correlations among the four variables within the context of appearance are displayed in Table 7 . The correla- tions among the appearance variables relate to the hypotheses in the same manner as did the correlations observed across contexts (i . e. , no support according to the classification system, but stronger-than-predicted support for the negative correlation between PDI and IA specified in Hypothesis 3 , the negative correlation between PDI and HLC specified in Hypothesis 4 , and 47 the positive correlation between HLC and IA specified in Hypothesis 5b) . There are two exceptions to the across contexts pattern that are noteworthy, at least from a mathematical perspective . The correlation between Impact and Health Locus of Control for appearance is considerably smaller than the correlation between these variables at the across contexts level (- . 546 vs . - . 769) . In terms of the classification system, the correlation between Impact and Health Locus of Control within the context of appearance is high rather than moderate as is the case at the across contexts level . The other exception is that the correlation between Sel f -Ef f icacy and Health Locus of Control for appearance is greater than the corresponding correlation at the across contexts level (-. 586 vs. -. 355) . Waning Interrelationships among the four variables within the context of peer relationships are displayed in Table 8 . The direction of each correlation at the peer relationships level is the same as it was at the across contexts level . However, three of the coefficients at the peer relationship level are only moderate , in contrast to their high counterparts at the across contexts level . As discussed below, only Hypothesis 3 (a moderate, negative correlation between PDI and IA) received support according to the coefficient classification system whereasthe relationships specified in Hypotheses 4 (a low, negative correlation between PDI and IA) and 5b (a low, positive correlation between HLC and IA) were stronger than anticipated . 48 Two of the more moderately correlated pairs of variables involve Health Locus of Control (PDI, HLC: IA, HLC) . Even this more moderate correlation between Health Locus of Control and Information Adequacy is stronger than the relationship predicted in Hypothesis 5b. The coefficients at the across contexts level are much greater than at the peer relationships level (for PDI , HLC: -.769 vs. -.422: for IA, HLC: .676 vs. .358) . Perhaps the Health Locus of Control scale is too broad to be sensitive to peer related implications of illness. Only one of the 18 items that compose the Health Locus of Control scale includes a reference to peers (item 31) . The weaker relationship between Impact and Information Adequacy at the peer relationships level , relative to the across contexts level , supports Hypothesis 3 (a moderate, negative correlation) . However, this coefficient may have been moderated by certain aspects of impact which cancel each other out and result in a rating which does not fully represent the actual degree of impact. Again, the comments offered by Subject 24 provide insight: In regard to the negative affect my disease and treatment have on my social life, I must comment that it has brought some people closer to me and others further away . W Bivariate correlations among the four variables within the context of future orientation are displayed in Table 9 . Hypotheses 3 (a moderate, negative correlation between PDI 49 Table 8 or.” an”: et-J‘ n-ru‘a'“ i'e'l!’ ’1 Variable SE IA HLC 1. PDI .509(2.915)""" -.412(-2.246)* -.422(-2.339)* 2. SE --- -.329(-1.255) -.463(-2.603)" 3. IA --- .358(1.946) 4. HLC --- M9; Values of 2 corresponding to r are in parentheses. ‘p< .05, “p<.01 Table 9 0,t'-.H ant! quI’ t- n"... .‘01'1101 .0 Variable SE IA HLC 1. PDI .406(2.239)* -.4S7(-2.564)" -.697(-4.474)"* 2. SE -- .092(.480) .258(1.372) 3. IA --- .348(1.887) 4. HLC --- H919; Values of z oorrespondin to r are in parentheses. *p<.05, "p<.01, * *p< 1 50 and IA) and 5a (a low, positive correlation between HLC and SE) received support according to the classification system whereas the relationships specified in Hypotheses 4 (a low, negative correlation between PDI and HLC) and 5b (a low, positive corre- lation between HLC and IA) were stronger than predicted. There is less association among the variables at the future orientation level than at the across contexts level . The coef- ficient of each of two pairs of correlated variables , Psycho- social Disease Impact/ Information Adequacy and Information Adequacy/Health Locus of Control , is only moderate within the context of future orientation, rather than high as was the case at the across contexts level . In the same trend, Self-Efficacy within future orientation bears negligible association with either Information Adequacy or Health Locus of Control . Orientation toward the future is perhaps based on factors which are less specific or less observable than the referents on which perceptions of appearance or peer relationships are based. Future orientation ' s more abstract , philosophical nature may be why the broadest variable, Health Locus of Control , is the best predictor (r= -. 697) of Impact. mainline Interrelationships among the four variables within the context of sexual functioning are displayed in Table 10 . None of the hypotheses received support according to the classifica- tion system although the relationships described in Hypotheses 4 (a low, negative correlation between PDI and HLC) , 5a (low, 51 positive correlation between HLC and SE) , and 5b (a low, positive correlation between HLC and IA) were stronger than anticipated. Other than the strong, positive correlation between Infor— mation Adequacy and Health Locus of Control ( . 667) , each corre- lation within this context is considerably different from its across contexts counterpart in terms of either magnitude or direction. In contrast to the larger, statistically signifi- cant coefficients at the across contexts level , only a negli- gible correlation exists between Impact and each of the other three variables . The relative nonassociation between sexual functioning Impact and each of the other three variables may be due to a lack of attention to sexual matters . Comments written on the questionnaires suggest that at least a few subj ects were confused as to how to respond to questionnaire references to sexual functioning, given their current lack of sexual relationships. It is likely that many of the subjects, particularly those with severe muscle disorders or active can- cers, would consider sexual matters relatively foreign as compared with family relationships, non-sexual physical abili- ties , and other matters of inescapable prominence. Consequently, high levels of Information Adequacy were endorsed. The moderate and positive correlation between Sel f-Ef f icacy and Information Adequacy may indicate that within this context both variables essentially relate to the same perspective: disinterest . 52 WW Bivariate correlations among the four variables within the context of freedom/ independence are displayed in Table 11 . None of the hypotheses was supported according to the classification system. The relationships specified in Hypothesis 4 (a low, negative correlation between PDI and HLC) and 5b (a low, positive correlation between HLC and IA) were stronger than predicted. Although the moderate tendency for increased levels of Information Adequacy to be associated with decreased levels of Self-Efficacy is evident within freedom/ independence (-. 460) , as well as within most other contexts , each of these variables appears to have little bearing on Impact (i . e. , the coefficient for PDI/SE is . 293 and the coefficient for PDI/IA is -. 293) . The strongest associations involve Health Locus of Control . As was the case at the across-contexts level , the degree of concordance between Information Adequacy and Health Locus of Control was high and the magnitude of the inverse relationship between Self-Efficacy and Health Locus of Control was moderate . MAW; Interrelationships among the four variables within the context of school achievement are displayed in Table 12 . Only Hypothesis 3 (a moderate, negative correlation between PDI and IA) received confirmation from the school achievement data whereas the relationships described in Hypotheses 4 (a low, negative correlation between PDI and HLC) and 5b (a low, S3 Table 10 o -.. 5110 ~ .-_1._l-. - u-a '.1.ql_-!‘110 'n it Variable SE IA HIE 1. PDI .187 (.938) -.259 (-1.377) -.301 (-1.600) 2. SE --- .508 (2.910)" .434 (2.415)"' 3. IA -- .667 (4.184)*** 4. HIE --- Note; Values of z correspondiazg to r are in parentheses. *p<.0 , "p<.01, "' *p<. 1 Table 11 03‘qu an”: 4.0- .1- ma "nut r-‘u' Variable SE IA HIE 1. PDI .293 (1.569) -.293(-1.569) -.679(-4.297)*" 2. SE --- -.460(-2.584)** -.498(-2.840)*"‘ 3. IA --- .605(3.645)*** 4. HIE --- Note; Valueos of 2 corresponding to r are in parentheses. It *1!!! p<. 1, p<.00 54 positive correlation between HLC and IA) were stronger than anticipated. There is a somewhat stronger association between Sel f - Efficacy and Impact with regard to school achievement, relative to the other psychosocial contexts. In general , students im- pacted the most perceived the greatest potential benefit from additional achievement-related information that would take into account the effects of illness and/or treatment. Given that the mean of school achievement Impact (3 . 30: SD=1. 78) is in the middle of the rating continuum, the moderately strong correla- tion between Impact and Sel f-Ef f icacy also reflects the perceptions of adolescents with low achievement Impact who consequently do not need disease-accommodating, school-related information. Although many of the subjects were receiving educational services designed to accommodate their health problems (e . g. , home-based instruction) , Impact was relatively independent of whatever information such programs provided (i . e. , the correlation between PDI and IA is only -. 301) . This relatively minor degree of association is in sharp contrast to the very substantial correlation coefficient at the across contexts level (-.709). PhysicalinzLLities Bivariate correlations among the four variables within the context of physical abilities are exhibited in Table 13 . Only Hypothesis 5b (a low, positive correlation between HLC and IA) was confirmed whereas the relationship described in Hypothesis 4 55 Table 12 or.” an”: -_o' a H‘A no :1'V'u‘1 :. Variable SE IA HLC 1. PDI 5150.959)" -.30l(-l.6l3) -.476(-2.691)** 2. SE --- -.366(-1.94) -.343(-l.858) 3. IA --- .482(2.731)"‘* 4. HLC --- Note; Values of 2 corresponding to r are in parentheses. "p<.01 Table 13 Ot‘-01.;nu' «t-J‘ t H‘A't 'v’ : Variable SE IA HIE l. PDI .160(.839) -.213(-1.124) -.591(-3.528)*** 2. SE --- .059(.307) -.l61(-.844) 3. IA --- .228(l.206) 4. HLC --- H.015: Values of 2 corresponding to r are in parentheses. **#p<.ml 56 (a low, negative relationship between PDI and HLC) was consi- derably stronger than anticipated. Health Locus of Control correlated - . 591 with physical abilities Impact, in contrast to the other two variables , each of which correlated negligibly with Impact (i.e. , SE/PDI-. 160, IA/PDI= -. 213) . In fact none of the bivariate relationships , other than that of Health Locus of Control and Impact, attained statistical significance at alpha . 05 or beyond . W25 Interrelationships among the four variables within the con text of family relationships are exhibited in Table 14 . Hypothesis 3 (a moderate, negative correlation between PDI and IA) received confirmation in accord with the classification sys- tem whereas the relationships predicted in Hypotheses 4 (a low negative correlation between PDI and HLC) and 5b (a low, posi- tive correlation between HLC and IA) were stronger than antici- pated. Table 14 closely resembles the correlation matrix of the across contexts variables (Table 6) . Within the family rela- tionships context, however, Information Adequacy bears a some- what weaker inverse relationship with Impact than was the case at the across contexts level (i.e. , -. 597 vs. -. 709) . Similarly, Information Adequacy ' 8 functional relationship with the variable most associated with Impact, Health Locus of Control , is sub— stantially lower within the context of family relationships , relative to the across contexts level (i.e. , . 460 vs. . 676) . It is likely that for many chronically ill adolescents family relationships are an ongoing challenge, even in families with 57 generally good relationships . The positive, moderate correla- tion between Impact and Sel f-Ef f icacy suggests that adolescents with disease-affected family relationships are relatively more interested in acquiring information about how to improve family relationships . Adolescents who perceived a high level of con- trol over health in general (Health Locus of Control) were much more apt to have family relationships with relatively low levels of disease-related discord. The observed correlation between Impact and Health Locus of Control most likely is the product of many determinants, including pre-diagnosis family functioning. C. Contextual Specificity of Bivariate Correlations Seven of the 28 inter-context differences in correlations between Psychosocial Disease Impact and Self-Efficacy attained statistical significance at alpha . 05 or beyond. 1 The observed differences are particularly noteworthy given the low power associated with N830 . The finding that one of every four tests of inter-context differences is significant provides moderate support for the heterogeneous relationship between Impact and Self-Efficacy specified in Hypothesis 6 . Only 3 of the 28 inter-context differences between Psycho- social Disease Impact and Information Adequacy attained statis- tical significance at alpha . 05 or beyond. Since one or two instances of statistical significance probably resulted from chance alone, the three significant inter-context differences observed do not support the heterogeneous relationship between 1 The author gratefully acknowledges assistance from Hripsime A. Kalaian in conducting these analyses . 58 Impact and Information Adequacy described in Hypothesis 6 . D . Supplemental Analyses WWW Psychosocial Disease Impact at the across contexts level was predicted with various linear combinations of the other three across contexts variables . The correlation and R2 of each set of predictors are exhibited in Table 15 . The amount of variance explained per combination ranges from approximately one-half (SE, IA) to two-thirds (SE, IA, HLC) . The addition of Sel f -Ef f icacy accounts for only two percent more variance than either Information Adequacy or Health Locus of Control alone. WW Appearance Impact was predicted with various linear combinations of the other two appearance-related variables in conjunction with Health Locus of Control . The correlation and R2 of each set of predictors are displayed in Table 16 . The amount of appearance-related Impact explained per combination ranges from 30 percent (SE, HLC) to 51 or 52 percent (all other combinations) . The addition of Self-Efficacy and/or Health Locus of Control to Information Adequacy did not enhance prediction even one percent . All intercorrelations within peer relationships are mod- erate. As is evident from comparisons among the R2 values in Table 17 , the range of explained variance among the four linear combinations of Impact predictors is narrow (i. e. , 27% to 35%) . An examination of Table 18 reveals that the addition of Table 14 01‘2015110' 9190‘ 1- 01 4" {‘e0110 3'; Variable SE HIE 1. PDI .479(2.710)" -.597(-3.578)"* -.778(-5.404)"* 2. SE --- -.394(-2.l64)"' -.623(-3.637)*** 3. IA -- .460(2.584)""" 4. HLC --- 1191;; Values of z correspondin to r are in parentheses. "' p<.05, "p<.01, """"'p<. 1 Table 15 ' ri ' f ' l A Contexts Predictor R R2 Variables 1. SE, IA 724*" .524 2. SE, HLC 783“" .614 3. IA, HIE 811*“ .657 4. SE, IA, HLC .816Iklntt .666 EDIE; "*p<.001 60 Sel f-Ef f icacy or Information Adequacy to Health Locus of Control explains only an additional five or six percent of the variance in Impact within the context of future orientation. As exhibited in Table 19 , the linear combination of all three predictors accounts for only about one-fourth of the variance in sexual functioning Impact . The same set of predic- tors in the other seven psychosocial contexts explains one-third to two-thirds of total Impact variance. As is exhibited in Table 20, only negligible increases in the prediction of f reedom/ independence Impact were achieved by adding Self-Efficacy and/ or Information Adequacy to Health Locus of Control . The Health Locus of Control/ Impact relationship most likely is the product of bidirectional influences, depen- dent on objective disease-related restrictions as well as general predispositions relevant to personal control . A comparison of the R2 values of the second and fourth lin- ear combinations exhibited in Table 21 indicates that Infome- tion Adequacy does not contribute uniquely to the prediction of Impact in the context of school achievement. The R2 values in Table 22 indicate that only an additional 1% of the physical abilities Impact variance is explained by supplementing Health Locus of Control with Sel f-Ef f icacy and Information Adequacy. Without the effect of Health Locus of Control , the predictor set explains very little ( i . e. , 8%) Impact variance . The relative nonassociation among the context- specific, disease-related information variables is particularly enigmatic within the context of physical abilities , given the 61 Table 16 .er-J'quro runs-0‘s 101-1510‘211‘ Predictor R R2 Variables l. SE(AP), IA(AP) .714 .510 2. SE(AP), HLC .547" .297 3. IA(AP), HIE 720*" .518 4. SE(AP), IA(AP), HLC 720*" .518 N012; I""'p<.01, "*p<.001 Table 17 M‘11m{anamtmflmm[amSEQnI;m—3mmfiflmmmic-:11!l'l'CIEE: Predictor R R2 Variables 1. SE(PR), IA(PR) .571“ .326 2. SE(PR), HLC .551M .303 3. IA(PR), HIE .506” .256 4. SE(PR), IA(PR), HLC .591" .349 Note; * p<.05, "p<.01 Table 18 _-'I'l°-...'."0_ o 0 on -o . 05901311031112) Predictor R R2 Variables 1. SE(FO), IA(FO) .588" .346 2. SE(FO), HIE 735*" .540 3. IA(FO), HIE 733*" .538 4. SE(FO), IA(FO), HLC 771*" .594 N912: "p<.01, ***p<.001 Table 19 IL .11‘ ' '91-; f 01‘ - o‘ ifi 10.91. '1--.lF~101niin Predictor R R2 Variables 1. SE(SF), IA(SF) .452* .204 2. SE(SF), HLC .464"' .215 3. IA(SF), HLC .311 .097 4. SE(SF), IA(SF), HLC 507* .257 Note; *p <.05 63 Table 20 11.-.,1 1 ' ‘11115 0, 01, .1-110.5,-"1011111‘11'11'1; LED Predictor R R2 Variables 1. SE(FI), IA(FI) .343 .118 2. SE(FI), HIE 681*“ .464 3. IA(FI), HIE 695*“ .484 4. SE(FI), IA(FI), HIE 696*" .484 EQLQ: M"'p<-001 Table 21 11. V._._"‘1 01.1 01 .1- 11 100 ' Predictor R R2 Variables 1. SE(SA), IA(SA) .529" .279 2. SE(SA), HLC .605“ .366 3. IA(SA), HLC .483“ .233 4. SE(SA), IA(SA), HIE .605 *"' .366 Note; *p<.05, """p<.01 64 prominent, concrete nature of physical concomitants of chronic illness. A comparison of the R2 values in Table 23 underscores the dominance of Health Locus of Control as a predictor of Impact within family relationships . The addition of Self-Bf f icacy to Health Locus of Control does not yield any discernible increase in prediction. By supplementing Health Locus of Control with Information Adequacy, an additional 7% of Impact variance is explained. W A two factor ANOVA was conducted with each of the four across contexts variables to detect effects of sex, age or an interaction of sex and age. The results of all four ANOVAs are exhibited in Table 24 . The only effect of significance is that the mean of across contexts Sel f-Ef f icacy for females is higher than the mean of the same variable for males . Female subjects , to a greater extent than their male counterparts , perceived themselves as being able to derive benefit from disease-related psychosocial information. Context-level analyses will reveal the domains in which such differences in perception were most apparent . W The effects of age and sex on each of the four appearance variables are exhibited in Table 25 . The difference between males and females in terms of Self-Efficacy is greater with regard to appearance relative to any other psychosocial context . 65 Table 22 Ml'vr"-c.-1-ro. 0 '1- 0' 11l-A1'11.-. P Predictor R R2 Variables l. SE(PA), IA(PA) .274 .075 2. SE(PA), HIE .594“ .353 3. IA(PA), HLC 596* .355 4. SE(PA), IA(PA), HLC .601" .361 Note: "p<.01 Table 23 MliV- . ' . '101' . 9 '1' 0'11 I .0. :F- 1.1 .'l-..n1.. (FR) Predictor R R2 Variables 1. SE(PR), IA(FR) 653*" .427 2. SE(FR), HIE .778*** .605 3. IA(FR), HIE 823*" .678 4. SE(FR), IA(FR), HIE 825*" .680 N919. * * *p<.m1 66 The Sel f—Ef f icacy ratings of females were approximately one and one-half points higher (on the 6 point rating continuum) than the comparable ratings of males . Although not significant at alpha . 05 (p< . 08) , females perceived greater appearance Impact (a difference of 1 . 2 on the rating continuum) which may have heightened their need for disease-related information and , consequently, their Sel f-Ef f icacy ratings . The effects of age and sex on each of the four peer rela- tionships variables are exhibited in Table 26 . The analysis of the variance of Sel f-Ef f icacy ratings revealed that the average rating of females (5 . 31) was significantly higher than the average rating of males (4 . 12) . As was the case with appear- ance , the difference between the Information Adequacy ratings of males and females fell slightly short of significance at alpha . 05 (p< . 06) : males were somewhat more apt to consider their current level of information adequate. However, differences between genders in Information Adequacy depended on age , as indicated by the significance of the interaction effect . As is evident from an examination of the cell means displayed in Table 27 , younger females were significantly less satisfied, relative to their male counter- parts and older adolescents , with their current level of disease-related information about peer relationships . Although experience with disease manifestations generally increases with age and consequently may produce an increasingly sufficient knowledge base, the reasons for the specificity in gender are difficult to discern. Perhaps the groups of younger adolescents 67 Table 24 E l . [IE . E I: Mean Wig—Younger Older F 512 PDI A Sex) 25.59 31.08 1.41 NS. B Age) 28.72 26.83 .17 NS. NS. SE A 31.88 38.54 5.15 .05 B 32.67 37.92 3.33 NS. AxB .19 NS. IA A 38.88 32.38 3.00 NS. B 34.17 38.92 2.23 NS. AxB .72 NS HLC A 73.12 67.31 2.03 NS. B 68.72 73.42 1.09 NS. AxB .09 NS. Table 25 Mean Wet Older F 518 PDI A Sex) 3.47 4.69 2.23 NS B Age) 3.83 4.25 .19 NS 2.63 NS SE A 4.06 5.54 7.52 .01 B 4.61 4.83 .32 NS. AxB .55 NS. ACLI A 4.35 3.00 3.33 NS B 3.67 3.92 .15 NS AxB .30 NS 68 differed on unassessed experience-related variables , such as amount of time since diagnosis. If most of the younger males had been diagnosed several years earlier than the younger fe- males , they may have compiled a greater amount of information relevant to dealing effectively with peers on disease related matters . The interaction is to a large extent a consequence of one or two extreme cases in the young female group. Subject 25 , for example, is the only subject who gave the lowest possible rating to each Information Adequacy item. Within the context of future orientation, the association between internal Locus of Control and relatively low Impact (and vice verse) is consistent across age groups and genders (Table 28) . There were no differences due to age among the remaining variables. Males, however, were again more satisfied than fe- males with their current levels of future-relevant information. The effects of age and sex on each of the variables within the sexual functioning context are exhibited in Table 29 . Although none of the effects was significant with alpha . 05 , the greater mean Sel f-Ef f icacy rating of the older adolescents (p< . 07 ) suggests that the perceived benefit of information re- garding sexual functioning tended to increase with age . As is evident from the values exhibited in Table 30 , none of the effects of sex and age attained statistical significance within the context of freedom/ independence . The effects of sex and age on each of the four school achievement variables are displayed in Table 31 . The analysis of the variance of Information Adequacy revealed that the 69 Table 26 El'tll' I-C '2 131' 1' [BB] Mean Wet Qldcr F 4.18. PDI Sex) 3. 29 3.92 .56 NS. BAge) 3.28 4.00 .85 NS. .56 N .S. SEA 4.12 5.31 4.27 .05 B 4.39 5.00 1.20 NS. AxB .07 NS. IA A 5.18 3.92 3.87 NS. B 4.28 5.17 3.87 NS. AxB 4.32 .05 Table 27 Cllli II 131' !' [PEI Age Younger Older Male 5.20 5.14 Sex Female 3.13 5.20 70 average rating of females (5 . 71) was significantly higher than the average rating of males (4 . 69) . Combined with the (nonsignificant) trend toward greater Sel f -Ef f icacy among females , this observation suggests that the female subjects were more interested in school-related information because they had not yet compiled information repertoires that they would con- sider to be adequate . As is indicated by the values displayed in Table 32 , the effects of sex and age were not statistically significant for any of the variables within the context of physical abilities . As evidenced by the consistently low F values displayed in Table 33 , the effects of age and sex are not statistically significant in the context of family relationships . E . Integration and Conclusions Three of the six hypotheses (1 , 2 , and 5A) describe rela- tionships involving Sel f-Ef f icacy. Correlations compatible with the relationship stated in Hypothesis 5a (a low positive corre- lation between HLC and SE) were observed in two psychosocial contexts, future orientation and sexual functioning. Otherwise, however, none of the hypotheses involving Self-Efficacy were supported. The primary reason for this seems to be that variables in addition to self-efficacy may have influenced ratings on the Self-Efficacy scale. Sel f-ef f icacy is a useful construct for understanding coping behavior (e.g. , Bandura, 1977) which, in the present case, involves the utilization of disease-related psychosocial Table 28 71 Mean er Older F sjg, PDI A Sex) 3.24 4.08 1.11 NS. B Age) 3.56 3.67 .02 NS. .09 NS. SE A 4.76 5.15 .51 NS. B 4.50 5.58 3.61 NS. AxB .00 NS. IA A 5.06 3.69 4.99 .05 B 4.17 4.92 1.83 NS. AxB .45 NS. Table 29 - n 1 Mean Warmer Older F sjg, PDI A Sex) 2.65 2.85 .11 NS. B Age) 2.61 2.92 .27 NS. NS. SE A 3.76 4.46 1.44 NS. B 3.61 4.75 3.62 NS. AxB .00 NS. IA A 4.35 3.92 .18 NS. B 4.17 4.17 .02 NS. 1.01 NS. Table 30 72 El'EIZ' I-C £111 1 [El] A Mean Water Older F 518. PDI A Sex) 3.12 4.08 1.98 NS. 8 Age) 3.39 3.75 .33 NS. .00 NS SE A 4.53 5.00 .45 NS. B 4.61 4.91 .19 NS. AxB 1.09 NS. IA A 4.82 4.69 .01 NS. B 4.39 5.33 3.03 NS. AxB .11 NS. Table 31 51' I!!' II: .5] “yv [5!) Mean Wager Older F Asia PDI A Sex) 3.29 3.31 .02 NS. B Age) 3.11 3.58 .37 NS. .66 NS. SE A 3.06 4.15 2.23 NS. B 3.17 4.08 1.66 NS. AxB .07 NS. IA A 5.71 4.69 4.21 .05 B 4.94 5.75 3.35 NS. AxB 7.52 NS 73 information. Coping, as discussed earlier, is the process of managing taxing circumstances (Lazarus and Folkman, 1984) . For many of the participants in the current study, in various psychosocial contexts, there were no ”taxing demands . " Conse- quently, variables other than sel f-ef f icacy , per se, must have influenced their responses to Sel f-Ef f icacy items . Patterns of interaction among the variables which may partially account for the observed Sel f-Ef f icacy ratings are proposed below. In the present study, low ratings on the Self-Efficacy scale are generally associated with either: (a) relatively high levels of Information Adequacy, independent of actual self- efficacy (i . e . , "I do not need that type information because my illness does not affect that context") : or (b) relatively low levels of Information Adequacy in conjunction with relatively low levels of actual self-efficacy (i. e. , "I do not know much about how to address my problems in that context and I doubt that I could apply new information effectively") . The first type of low Sel f-Ef f icacy subject is likely to have low Impact ratings and high Locus of Control ratings . The second type of low Sel f-Ef f icacy subject is likely to have high Impact ratings and low Health Locus of Control ratings . High ratings on the Sel f-Ef f icacy scale are generally associated with either: (a) relatively high levels of Information Adequacy in conjunction with relatively high levels of actual self-efficacy (i. e. , "I know a lot about how to handle my problems in that context, but I could always make good use of additional information") , or (b) relatively low levels of Table 32 E l' [11' I-C 11.15”” :25: 74 Mean f F i PDI A Sex) 4.12 4.85 1.02 NS. B Age) 4.78 3.92 2.26 NS. .60 NS. SE A 4.06 4.85 2.69 NS B 4.39 4.42 .05 NS AxB .73 NS. IA A 4.53 4.31 .19 NS. B 4.39 4.50 .01 NS. AxB .16 NS. Table 33 El'EIZ' IE 13,131.].3] Mm WW ngig. PDI A Sex) 2.41 3.31 1.08 NS. B Age) 2.89 2.67 .14 NS. .54 NS. SE A 3.53 4.31 .99 NS. B 4.00 3.67 .30 NS. AxB .43 NS. IA A 4.88 4.15 .92 NS. B 4.44 4.75 .39 NS. AxB 1.21 NS. 75 Information Adequacy in conjunction with relatively high levels of actual self-efficacy ( i . e. , ”I do not know much about how to address my problems in that context , but I am willing and prob- ably able to apply relevant new information effectively") . The first type of high Sel f-Ef f icacy subject is likely to have mod- erately low Impact ratings and high Health Locus of Control ratings . The second type of high Sel f-Ef f icacy subject is likely to have moderately high Impact ratings and high Locus of Control ratings. The notion of contextual specificity of correlations among variables received moderate support only in relation to Psycho- social Disease Impact and Sel f-Ef f icacy: the relationship between Impact and Information Adequacy varied relatively little between contexts . Although the relationships in some contexts (such as family relationships) closely reflected the across contexts pattern, relationships in other contexts (such as sexual functioning) were highly unique . As with the bivariate relationships , patterns of multivariate prediction differed among the various psychosocial contexts . The moderate degree of contextual specificity observed in this study lends credence to recent caveats about reliance on broadly based measures of psychosocial functioning . One of the most surprising outcomes of the study was the absence of statistically significant age effects. Of the 28 two-way ANOVAs, only the interaction of age and sex in one context, peer relationships, attained significance. It was anticipated that the advances in social cognition which 76 typically accompany increases in age would be evident in the ratings. The rationale for such expectations included, for example, the finding of Leadbeater and Dionne (1981) who observed a close relationship between identity achievement and formal operational thought. Identity achievement status (vs . identity diffusion) was associated with superior performances when confronted with identity-related issues . Given that the questionnaire of the present study is replete with identity- related issues , the lack of observed differences may be due to the insensitivity of the rating continuum, rather than to homogeneity in social cognitive development. Similar "task artifact" explanations have been offered for other instances in which expected qualitative cognitive differences between age groups were not manifested in performance (e.g. , Neimark, 1980) . Yet other scales similar in format to the questionnaire used in the present study have reflected differences among various adolescent age groups in terms of social cognition (Ford, 1982) and locus of control (Germain, 1985) . Alternatively, perhaps aspects of the nonrandom selection process , such as volunteer- ism, did produce a somewhat homogeneous sample in terms of social cognition. The only significant difference among genders at the across contexts level was in relation to Self-Efficacy. The Self- Efficacy ratings of females were significantly higher than those of their male counterparts in two contexts , appearance and peer relationships . Especially females who were relatively less satisfied with their current levels of information believed that 77 they could make good use of additional information about appear- ance and peer relationships which would take into account the effects of illness (or treatment) . V . SUMMARY AND RECOMMENDATIONS Relationships among four variables relevant to the psycho- social perspectives of chronically ill adolescents were studied in each of eight psychosocial contexts and at the across- contexts level . The four variables were Psychosocial Disease Impact (PDI) , Self-Efficacy (SE), Information Adequacy (IA) , and Health Locus of Control (HLC) . All variables other than Health Locus of Control referred to disease-related psychosocial infor- mation in the following contexts: appearance, peer relation- ships , future orientation, sexual functioning, freedom/ independence, school achievement, physical abilities, and family relationships . Of particular interest was the degree of association between Psychosocial Disease Impact and the other variables (individually and in combination) . Differences due to age and sex were examined. Interpretations of some of the most significant correlations were provided based on the presumed nature of the variables . The observed correlations were con- trasted with the hypothesized relationships . The sample consisted of 30 health-impaired adolescents, ages 13-18 years inclusive. Approximately two-thirds of the subjects had cancer whereas several disorders were represented in the remaining third. Each subject responded to a 42-item questionnaire which contained the four operationalized variables. 78 79 A. Summary of Bivariate Correlations Both the substantive and formal versions of the bivariate hypotheses are restated below. The coefficients that correspond to the adjectives high , moderate and low are as follows: high is _>_ . 60, medium is . 30-. 59, low is s . 29. All coefficients of . 36 or greater are statistically significant (at alpha 5 . 05) , given N=30 . The obtained bivariate correlations in relation to the hypotheses are exhibited in Table 34 . 1. Subjects who have high Sel f-Ef f icacy (believe that they could make good use of disease-related information) are likely to have used information effectively in the past , thereby lowering their Psychosocial Disease Impact. Subjects with low Sel f - Efficacy are likely to have high Impact as a conse- quence of not using psychosocial information in the past. This relationship will be expressed as a high negative correlation between Sel f-Ef f icacy and Psychosocial Disease Impact. Hypothesis 1: A significant, negative correlation will exist between Psychosocial Disease Impact and Self-Efficacy. The relationship between Self-Efficacy and Psycho- social Disease Impact stated in Hypothesis 1 will be stronger than any other hypothesized relationship . Hypothesis 2: The magnitude of the negative correlation between 80 Psychosocial Disease Impact and Sel f-Ef f icacy will be greater than the correlation between Psychosocial Disease Impact and any other variable . Subjects who have adequate levels of disease-related information are likely to have benefited from such information in terms of Psychosocial Disease Impact. Subjects with very inadequate levels of disease- related information have had little information to bring to bear on the psychosocially relevant effects of their diseases. This relationship will be expressed as a moderate, negative correlation between Information Adequacy and Psychosocial Disease Impact . Hypothesis 3 : A significant , negative correlation will exist between Psychosocial Disease Impact and Information Adequacy. Subjects who perceive a great deal of control over their health in general will tend to have relatively low Psychosocial Disease Impact whereas subjects who perceive little control over their health will tend to have relatively high levels of Impact. This relationship will be expressed as a low, negative correlation between Health Locus of Control and Psychosocial Disease Impact. Hypothesis 4: A nonsignificant negative correlation will exist between Psychosocial Disease Impact and Health Locus of Control . 81 5 . Subjects who have high Sel f-Ef f icacy will tend to have high (internal) Health Locus of Control (and vice versa) . Subjects who have high levels of Information Adequacy will also tend to show high levels of Health Locus of Control . These relationships will be expressed as low, positive correlations . Hypothesis 5 : A nonsignificant positive correlation will exist between: (a) Health Locus of Control and Sel f- Efficacy: (b) Health Locus of Control and Information Adequacy . Table 34 S EDI’ICI"BI’ 3.. W Psychosocial Hypothesis Context 1 2 3 4 5a 5b Across OODICXIS the as so Appearance "1 are 41111 Peer relationships u we Future orientation .. u Sexual functioning "1 s s s Freedom/independence u es School achievement n u Physical abilities “ Family relationships " liillllll lllllllll {IIOQ *‘ill Note: - a hypothesis not supported "' a hypothesis supported " =- 0orrelation stronger than anticipated (e.g., high rather than moderate) 82 Correlations between Psychosocial Disease Impact and Self- Efficacy generally were moderate and positive , ranging from . 187 to . 515 across contexts . Correlations between Psychosocial Disease Impact and Information Adequacy were negative and varied considerably in magnitude across contexts , ranging from - . 213 to - . 778 . Correlations between Psychosocial Disease Impact and Health Locus of Control were negative and tended to be moderately high, ranging from -. 301 to - . 77 8 . Most correlations between Sel f -Ef f icacy and Information Adequacy or Health Locus of Control were negative and moderately low, but there was considerable variability in these relation- ships among contexts . Correlations between Information Adequacy and Health Locus of Control were positive and tended to be moderately high . Interpretations of the observed correlations were based on the extent to which each scale was believed to represent its underlying variable, as well as on the nature of the variables themselves . The proposed relationships are particularly tenuous, given the small size of the sample. Issues were raised regarding the match between the Sel f - Efficacy scale and the sel f-ef f icacy construct on which it was based. Self-efficacy may have been only one of a set of variables that determined ratings on the Self-Efficacy scale . For example, Sel f-Ef f icacy ratings may have been the product of very different considerations for persons who were confronted with significant disease-related psychosocial challenges , relative to individuals who were not . 83 B. Summary of Contextual Specificity The question of whether or not the bivariate correlations involving Psychosocial Disease Impact would vary significantly among psychosocial contexts was addressed in Hypothesis 6 . Both the substantive and formal versions of Hypothesis 6 are restated below. 6. The relationship between Sel f-Ef f icacy and Psychosocial Disease Impact will vary significantly from one psychosocial context to another. The same will be true for the relationship between Psychosocial Disease Impact and the other context-specific variable, Information Adequacy. This variability will be evidenced by significant differences between the eight psychosocial contexts in terms of the correlations which involve Psychosocial Disease Impact. Hypothesis 6: There will be significant variability between psycho- social contexts with regard to bivariate correlations which involve Psychosocial Disease Impact. One of every four inter-context differences in correlations between Impact and Sel f-Ef f icacy attained statistical significance. This observation provides moderate support for the contextual heterogeneity of the relationship between Impact and Self-Efficacy specified in Hypothesis 6 . Inter-context 84 differences in correlations between Impact and Information Adequacy were nonsignificant in all but three comparisons , and therefore are not consistent with the context-specific relationship described in Hypothesis 6 . C. Summary of Multivariate Prediction Despite the high correlations that existed among pairs of individual predictors of Psychosocial Disease Impact in various psychosocial contexts , linear combinations were often productive in claiming additional Impact variance. The addition of Infor- mation Adequacy to Health Locus of Control generally enhanced prediction of Impact substantially whereas the addition of Self- Efficacy did not . D. Summary of Age and Gender Effects Age did not have a significant main effect on any of the variables, either across or within psychosocial contexts . Within the context of peer relationships , however, younger females rated their current level of information as relatively inadequate . Males perceived greater Information Adequacy than females with regard to future orientation and school achieve- ment . Females produced higher Sel f-Ef f icacy ratings with regard to appearance and peer relationships: this difference held true at the across-contexts level as well . 85 E . Recommendat ions W Given the strength of many of the correlations in this sur- vey, further exploration in this area is warranted. What is needed first are replications with larger samples of randomly selected representatives of the population of chronically ill adolescents . Variables should continue to be framed within nar- row psychosocial contexts . The effects of secondary variables such as length of time since diagnosis, IQ, objective disease status, and prognosis should be assessed and removed (as covariates) so as to increase the salience of the relationships among the variables of primary interest. Causal mechanisms thought to underlie reliably observed functional relationships can then be tested . Such an endeavor might involve providing specific disease-related psychosocial information to one member of an Impact-matched pair and nonspecific attention to the other: observed reductions in Impact could then be said to be the result of increased adequacy of information. The relationship between various measures of social cogni- tion and behavior in interpersonal situations has generally been only modest (Kendall & Braswell, 1985) . Concordance would likely be improved if the variables that are to be compared are measured in reference to more narrowly defined social contexts . The present project provides a prototype for diverse avenues of context based psychosocial research . 86 :1 . l 1 E J l I . Clinical uses of the scales employed in this study would necessitate careful analyses of client ratings within each con- text . Similarities among the across context means of various subjects in the present study obscured important individual differences . For example, Subjects 1 and 21 had the same mean Impact score (4 . 25) yet had very different standard deviations (2 . 19 and . 46, respectively) . A contextual analysis reveals that Subject 1 perceived extreme impact in half of the contexts and relatively little in the others . In contrast, Subject 21 ' 8 ratings were consistently within the upper end of the moderate range. Although large samples of chronically ill adolescents need to be surveyed before commonalities can be stated with certainly, there is no need to defer clinical applications until such results are available . Measures such as those employed in the present study may be an efficient means of developing an understanding of idiosyncratic client perspectives . APPENDICES APPENDIX A MICHIGAN STATE UNIVERSITY mammmmm ”WOWMOdflu-I“ ANDDMNOIMGMDUAHW November 23, 1987 Mr. Randall J. Verge 8416 wonderland Blvd. Bedding, California 96003 Dear Mr. Verge: Subject: Proposal Entitled, "Interrelationships Among Disease Impact, Self-Efficacy. and Information in Psychosocial Contexts of Adolescence" UCRIHS' review of the above referenced project has now been completed. I am pleased to advise that since the reviewers' comments have been satisfactorily addressed, the conditional approval given by the Committee at its November 2, 1987 meeting has now been changed to full approval. You are reminded that UCRIHS approval is valid for one calendar year. If you plan to continue this project beyond one year, please make provisions for obtaining appropriate UCRIHS approval prior to November 2, 1988. Any changes in procedures involving human subjects must be reviewed by the UCRIHS prior to initiation of the change. UCRIES must also be notified promptly of any problems (unexpected side effects, complaints. etc.) involving human subjects during the course of the work. Thank you for bringing this project to our attention. If we can be of any future help, please do not hesitate to let us know. Aincarely , 251..“ 2. 6.156“ r” Benry E. redeck, Ph.D. Chairman, UCRIHS HEB/jms cc: Dr. E. Clarizio MSU is an Alfie-leave Action/Equal Opportunity lulu-mos 88 APPENDIX B Request for Study Participants Name Address City, State, Zip Dear Dr. This is a request for your assistance in securing subjects for my Ph. D. dissertation (Michigan State University) . I am looking for chronically ill adolescents, ages 13-18 years inclusive, excluding those suspected of being mentally retarded . My interest is not with particular diseases , but rather with any chronic illness which would likely impact psychosocial development in one or more of the following contexts: appearance, peer relationships, future orientation, sexual functioning, freedom/ independence, school achievement, physical abilities, and family relationships. The analysis will focus on the extent to which psychosocial impact is mediated by certain factors such as perceptions of "self-control" and the adequacy of the patient ' s knowledge of how to cope with psychosocial aspects of the illness (e.g. , hair loss, weight fluctuations, increased dependency, limited physical abilities, etc.) . The results of this study will be a helpful step in determining how to attenuate the psychosocial impact of chronic illness on adolescent development . However, personal benefit as a direct result of participation in this "basic research" project is unlikely. I would be very grateful if you were willing and able to provide access to one or two potential subjects . Please be assured that the rights and welfare of your patients are not likely to be compromised in any way. My prospectus has been approved by MSU ' 8 very stringent University Committee on Research Involving Human Subjects . Here ' s how to help: Inform the adolescent patient and his/her parent or guardian that you are assisting in a psychological study of chronic illness during the teen-age years . Indicate that subject participation consists only of responding to a 42 item questionnaire. Mention that I will send 89 9 0 Page 2 a $10 . 00 check to the participant upon receipt of the signed consent forms and completed questionnaire . Ask if they would allow you to release their names and phone number to me . Assure them that I will explain the research and that they will not be pressured to participate . Notify me, by phone or mail , of these candidates . (University- imposed time constraints require that the data collection phase be completed very soon. . . ) Thank you very much for considering my request . I will contact you by phone in approximately 10 days to answer any questions you may have and to determine the availability of potential subjects . Sincerely, Randall J. Varga School Psychologist 8416 Wonderland Blvd . Redding, CA 96003 (916) 244-4600 extension 261 RJV/kmh APPEN DI X C H69, TEQNSL’ Bordon Verge, 0 psuch- ologist, writing hls doctoral thesis For michigcrl State. is 0.3-“, looking for teens with cancer CZ (5:) who want to make more. \7 -4) He is gathering research on carter patients mtmen the cares of l3 and 18 unclusnve. If oou. wuth the permission of unur parents. fill out o 42-item questionnaire, he will pa; 001.; $10. H? has sent me ten Mstionr'oires to distribute, so the first 10 of 001.1 teens to call me can make a fast $10. Call me. Perm Hreirberg, For information at 281-9966. 91 APPENDIX D Teenager/Patient Questionnaire Teen-agerlPatient Questionnaire Age: Date: This questionnaire deals with some of the personal and social aspects of having a chronic illness during the teen-age years. Please rate each item by circling one of the six numbers below it, based on how you feel about the item. This information is confidential and will be shared only with your doctor(s). 1. My illness (or treatment) has a negative effect on my social life. l 2 3 4 5 6 No, Not sure, 7 Yes, not at all maybe definitely 2. I currently have adequate information about increasing my freedom and independence, which takes into account the effects of my illness (or treatment). 1 2 3 4 5 6 No, Not sure, Yes, not at all maybe definitely 3. I currently have adequate information about increasing my school achievement, which takes into account the effects of my illness (or treatment). 1 2 3 4 5 6 No, Not sure, Yes, not at all maybe definitely 4. I could make good use of information about planning for the future, despite the effects of my illness (or treatment). l 2 3 4 5 6 No. Not sure, Yes, not at all maybe definitely 17 92 10. 11. 12. 923 No matter what I do. if I am going to get sick. I will get sick. l 2 3 4 S 6 No. Not sure, Yes. not at all maybe definitely When I get sick I as to blame. 1 2 3 4 5 6 No. Not sure, Yes, not at all maybe definitely No matter what I do, I am likely to get sick. 1 2 3 4 5 6 No, Not sure, Yes, not at all maybe definitely If I take care of myself. I can avoid illness. 1 2 3 4 5 6 No, Not sure. Yes. not at all maybe definitely Regarding my health. I can only do what my doctor tells me to do. 1 2 3 4 5 6 No. Not sure, Yes. not at all maybe definitely I currently have adequate information about improving my sexual functioning, which takes into account the effects of my illness (or treatment). l 2 3 4 5 6 No. Not sure, Yes. not at all maybe definitely I could make good use of information about how to minimize the effect that my illness (or treatment) has had or will have on my sexual functioning. 1 2 3 4 5 6 No. Not sure, Yes. not at all maybe definitely I could make good use of information about how to improve my appearance, despite the effects of my illness (or treatment). 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely 13. 14. 15. 16. 17. 18. 19. 20. £94 I could make good use of information about how to improve my social life. despite the effects of my illness (or treatment). 1 2 3 4 S 6 . No. Not sure. Yes. not at all maybe definitely I currently have adequate information about becoming more physically active. which takes into account the effects of my illness (or treatment). 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely If I get sick. it is my own behavior which determines how soon I get well again. 1 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely Health professionals control my health. 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely Whenever I do not feel well. I should consult a medically trained professional. 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely Hy illness (or treatment) has affected or will affect my sexual functioning. l 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely I could make good use of information about how to improve my family relationships. despite the effects of my illness (or treatment). 1 2 3 4 5 6 No. Not sure, Yes. not at all maybe definitely The main thing which affects my health is what I myself do. 1 2 3 4 S 6 No. Not sure, Yes. not at all maybe definitely 22. 23. 24. 25. 26. 27. 28. 915 I currently have adequate information about improving my appearance. which takes into account the effects of my illness (or treatment). 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely If it's meant to be. I will stay healthy. 1 2 3 4 5 6 No. Not sure. Yes. not at all - maybe definitely My illness (or treatment) limits my physical activity. 1 2 3 4 5 _g No. Not sure. Yes. not at all maybe definitely Luck plays a big part in determining how soon I will recover from an illness. l 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely I could make good use of information about becoming more physically active. despite the effects of my illness (or treatment). 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely I currently have adequate information about improving my family relationships. which takes into account the effects of my illness (or treatment). l 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely My illness (or treatment) has changed my plans for the future. 1 2 3 4 S 6 No. Not sure, Yes. not at all maybe definitely Host things that affect my health happen to me by accident. 1 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely 29. 30. 31. 32. 33. 34. 35. 36. 965 My illness (or treatment) has a negative effect on the amount of freedom and independence that I have. 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely I could make good use of information about improving my school achievement despite the effects of my illness (or treatment). 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely when I recover from an illness. it's usually because other people (for example. doctors. nurses. family. friends) have been taking good care of me. 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely Hy illness (or treatment) has a negative effect on my family relationships. 1 2 3 4 S 6 No, Not sure. Yes. not at all maybe definitely I currently have adequate information about improving my social life. which takes into account the effects of my illness (or treatment). 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely My family has a lot to do with my becoming sick or staying healthy. 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely Hy illness (or treatment) has a negative effect on my school achievement. 1 2 3 4 5 6 No. Not sure, Yes. not at all maybe definitely Hy illness (or treatment) has a negative effect on my appearance. 1 2 3 4 5 6 No. Not sure. Yes. not at all maybe definitely 37. 38. 39. 40. 41. 42. 97 If I take the right actions. I can stay healthy. l 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely I currently have adequate information about planning for my future. which takes into account the effects of my illness (or treatment). 1 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely I could make good use of information about how to increase my freedom and independence. despite the effects of my illness (or treatment). 1 2 3 z. ' 5 6 No. Not sure. Yes. not at all maybe definitely I am in control of my health. 1 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely Having regular contact with my physician is the best way for me to avoid illness. 1 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely My good health is largely a matter of good fortune. l 2 3 4 S 6 No. Not sure. Yes. not at all maybe definitely Thank you for participating in this study. If you would like to discuss any of the questionnaire items. please indicate this in the space below or mention it to the researcher or his assistant. APPENDIX E Subject Consent Form Consent to Participate in a Psychological Study of Chronic Illness in adolescence Researcher: Patient: Randall J. Varga Name: Michigan State University Age: Department of Counseling. Educational Sex: Psychology 8 Special Education Illness: East Lansing. MI 48824 Physician(s): This study examines the impact of chronic illness on several psychological and social aspects of adolescent development. These aspects are appearance. peer relationships. future orientation. sexual functioning. freedom/independence. school achievement. physical abilities. and family relationships. We are interested in determining the extent to which psychosocial disease impact is influenced by certain factors such as feelings of "self-control" and the adequacy of the patient's knowledge about how to deal with psychosocial aspects of the illness (e.g.. hair loss. weight fluctuations. increased dependency. limited physical abilities. etc.). The results of this study will be a first step in determining how to reduce the psychosocial impact of chronic illness. Participation consists of responding to a forty-two item questionnaire. Because some of the questionnaire items deal with potentially sensitive topics. there is a slight risk that responding to such items will produce a sense of personal distress. l. I freely consent to participate in a scientific study being conducted by the researcher identified above. 2. The study has been explained to me and I understand what my participation will involve. 3. I understand that the questionnaire deals with very personal issues such as sexual functioning. family relationships. self-esteem and appearance. 9!! be 959 I understand that the results of the study are confidential and that I will remain anonymous. The exchange of information between my physician(s) (Dr. ) and the researcher (Randall J. Varga) and/or his assistant (Jill Hadsen) will be the only disclosure of information in which I will be identified. I understand that participation in the study will have no effect on the medical care that I receive. I understand that participation in the study will not result in any additional service nor will it guarantee any benefit to me. I understand that I may cancel this consent to participate at any time. I understand that I will receive a sumary of the study when it is complete. Participant Today's Date 1&FH?EHWE[I)(I? Parent/Guardian Consent Form Consent to Participate in a Psychological Study of Chronic Illness in Adolescence Researcher: Patient: Randall J. Varga Name: Michigan State University Age: Department of Counseling. Educational Sex: Psychology 8 Special Education Illness: East Lansing. HI 48824 Physician(s): This study examines the impact of chronic illness on several psychological and social aspects of adolescent development. These aspects are appearance. peer relationships. future orientation. sexual functioning. freedom/independence. school achievement. physical abilities, and family relationships. We are interested in determining the extent to which psychosocial disease impact is influenced by certain factors such as feelings of "self-control" and the adequacy of the patient's knowledge about how to deal with psychosocial aspects of the illness (e.g.. hair loss. weight fluctuations, increased dependency. limited physical abilities. etc.). The results of this study will be a first step in determining how to reduce the psychosocial impact of chronic illness. Participation consists of responding to a forty-two item questionnaire. Because some of the questionnaire items deal with potentially sensitive topics. there is a slight risk that responding to such items will produce a sense of personal distress. l. I freely consent to participation by my son/daughter in a scientific study being conducted by the researcher identified above. 2. The study has been explained to me and I understand what participation by my son/daughter will involve. 3. I have reviewed the "Teen-ager/Patient Questionnaire" and am aware that it deals with very personal issues such as sexual functioning. family 1()0 11)]. relationships. self-esteem and appearance. I understand that the results of the study will be confidential and that my son/daughter will remain anonymous. The exchange of information between the physician(s) (Dr. ) who treats my son/daughter and the researcher (Randall J. Varga) and/or his assistant (Jill Hadsen) will be the only disclosure of information in which my son/daughter is identified. I understand that participation in the study will have no effect on the medical care that my son/daughter receives. I understand that participation in the study will not result in any additional service nor will it guarantee any benefit to my son/daughter. I understand that I may cancel this consent to participate at any time. I understand that I will receive a summary of the study when it is complete. Parent/Guardian Today's Date APPENDIX G Study Summary for Participants Remember the questionnaire that you filled out a few months ago for a study of psychosocial aspects of having a chronic illness? Well . the results are in and. as promised. I will attempt to summarize some of them for you without too much jargon or too many words (the full-length version is over 100 pages). The purpose of the study was to determine how strongly cer- tain variables influence the psychosocial impact of having a serious chronic illness. The participants were 30 chronically ill adolescents (most had cancer) , ages 13-18 years inclusive. The questionnaire that you completed was actually several relatively short scales blended together: each scale represented a variable. I ' ll discuss the variables in plain English rather than in terms of their technical names and definitions . I looked at the relationships among the variables in each of eight psychosocial contexts that are often important in adolescent development: appearance. peer relationships, future orientation. sexual functioning, freedom/ independence. school achievement, physical abilities and family relationships. Here ' s some of what I found: 1 . Overall . there was a strong inverse relationship (i . e. . as one variable goes up the other comes down) 102 103 between the level of perceived ”self-control" over psychosocially relevant aspects of the illness (e.g. . weight fluctuations, increased dependency. limited physical abilities. etc. ) and level of perceived psychosocial impact. This relationship was particu- larly strong in regard to family relationships . future orientation and f reedom/ independence . It was weakest in the context of sexual functioning. There were no significant differences between males and females or between adolescents of different ages. The adequacy of a person ' 3 current level of informa- tion about coping with psychosocial ly relevant aspects of illness (or treatment) was a good predictor of degree of psychosocial impact . This was another inverse relationship: high levels of information adequacy were associated with low levels of impact and vice versa. Perceptions of information adequacy did not differ according to sex or age . Although greater need for psychosocial information that takes into account the effects of disease tended to be higher for persons with low levels of psycho- social information. there were many exceptions . Some people who already knew a lot wanted to know more . Some people who had relatively little information doubted that they could make good use of additional information. Some people had little need for disease— related psychosocial information in certain contexts 104 simply because their diseases did not have implications for their functioning in those contexts . Males perceived a greater level of information adequacy in the contexts of future orientation and school achievement : there were no sex differences in terms of information adequacy in the other psycho- social contexts . Younger females were significantly less satisfied with their current levels of informa- tion in the context of peer relationships . In the contexts of appearance and peer relationships. females were more interested than males in additional psycho- social information that would take into account the effects of their illnesses or treatments . Research is kind of like a funnel . This research project was near the wide part; it attempted to answer broad questions . Associations (correlations) among variables do not tell us 5111! those relationships exist. Whenever I use words like "because" to describe how a particular variable got to be high or low. I 'm just speculating. If the associations observed in this study turn out to be reliable (show up in other samples) the next step would be to take a close look at how changes in one variable actually cause changes in one or more of the other variables . Until then. we ' re left with the classic chicken and egg dilemma . For example. did increases in perceived self-control over psychosocially relevant aspects of the disease result from increased knowledge about how to deal with those aspects or did increased knowledge produce increased self-control? They 105 probably influenced each other . but how much influence did each variable have on the other? The goal that drives all of this academic stuff is to help the people who attempt to help you through the tough times . It helps them know what they should be sensitive to and how to be most effective in addressing your concerns . Thanks again for your help with this project . 1. 2. 3. 4. 5. 6. 7. 8. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. PDI TRS SETRS AI TRS HLC TRS PDI (AP) SE (AP) AI (AP) PDI (PR) SE (PR) AI (PR) PDI (F0) SE (F0) AI (F0) PDI (SF) SE (SF) AI (SF) PDI (FI) SE (FI) AI (FI) PDI (SA) SE (SA) AI (SA) PDI (PA) SE (PA) AI (PA) PDI (FR) SE (FR) AI (FR) APPENDIX a W 106 107 APPENDIX H (Continued) 29 . SEX 3 0 . AGE Subject: 6 6 6 Subject: 6 1 6 Subject: 6 1 1 Subject: 4 6 2 Subject: 4 S 5 Subject: 1 6 2 Subject: 2 6 5 Subject: 4 3 5 Subject: 1 6 2 Subject: 3 S 5 Subject: 6 S 5 Subject: 5 6 6 Subject: 1 6 5 Subject: 1 6 5 Subject: 4 6 4 Subject: 6 6 5 Subject: 4 5 5 Subject: 3 6 3 Subject: 5 5 2 Subject: 5 2 6 Subject: 5 S 4 Subject: 1 S 6 Subject: 1 6 3 Subject: 5 6 5 Subject: 6 l 6 Subject: 6 6 6 Subject: 3 6 6 Subject: 6 6 6 Subject: 5 3 6 Subject: 1 6 S \O 00 NOV! § UMP C) O on NOU‘ fi UNH U38bNMN§NUNMNUNHN§NUNbNO‘HU‘HMHOHU‘HNHUHUHOH#HNOO‘OU‘NU‘OSPUUUIS‘GUUNOH 45 Data File 83 6 3 4 6 6 S 1 40 92 l 3 3 1 48 90 1 2 6 1 L\ 108 “ Id 4 S l 6 6 2 6 1 5 6 f») 6 6 4 6 4 1 4 6 1 S 6 h) H 4 6 4 6 [J List of References Achterberg, J. . Matthews-Simonton. S. . & Simonton. O.C. (1977) . Psychology of the exceptional cancer patient: A description of patients who outlive predicted life expectancies . WM. 1.4.. 415-422- American Psychological Association. (1982) . W W (2nd ed- ) . Washington, D.C. : Author. Anastasi, A.A. (1968) . W (3rd ed.) . New York: Macmillan. Anderson. F.J. . Asher. ILA. , Clark. 6.14. Orrick. J.M. . & Quiason, E. P. (1979) . Adjustment of adolescents to chronic disability. WWW, 15(4) . 177-185. Bandura. A. (1977) . Self-efficacy: Toward a unifying theory of behavioral change- W. 5.1. 191-215- Bedell, J.R. , Giordani. B. . Amour. J.C. , Tavormina, J. , & Boll, T. (1977) . Life stress and the psychological and medical adjustment of chronically ill children. M W. 2.1. 237-242- Bibace, R. . & Walsh. N.E. (1980) . Development of children's concepts of illness. Pediatrics. fl. 912-917. 109 110 Blommers. P-J- . 8: Forsyth. R-A- (1977) - WW1 W (2nd ed.) . New York: Houghten Mi f fl in . Blotcky. A.D. . Cohen. D.G. . Conatser, C. . & Klopovich, P. (1985) . Psychosocial characteristics of adolescents who refuse cancer treatment. W. 5.3.. 729-731. Blumberg, B.D. . Lewis. J.M. & Susman, E.J. (1984) . Adolescence: A time of transition. In 11.6. Eisenberg. L.C. Sutkin, & M.A. Jansen (Eds-) . WWW 1.1m (pp. 133-149) . New York: Springer. Campbell . J. D. (1975) . Illness is a point of view: The development of children's concepts of illness. Wm. Ag, 92- 100. Cohen. F. (1980) . Coping with surgery: Information, psychological preparation, and recovery. In L.W. Poon (Ed.) . W W- (PP- 375-382) - WaShington. D. C. : American Psychological Association. Cohen, F. . & Lazarus, R.S. (1973) . Active coping processes, coping dispositions. and recovery from surgery. W W. 3.5, 375-389. Ellis. D.W. . & Davis. L.T. (1982) . The development of self- concept boundaries across the adolescent years . W, 11. 695-710. Erikson. E. (1959) . Identity and the life cycle. Psychological Issues. Monograph 1. New York: International Universities Press e 111 Erikson. E. (1963) . W (2nd ed.) . New York: Norton. Felton, B.J. . & Revenson, T.A. (1984) . Coping with chronic illness: A study of illness controllability and the influence of coping strategies on psychological adjustment. WW Mom, 51, 343-353. Ford. M. E. (1982) Social cognition and social competence in adolescence. Wm, 1g, 323-340. Fisher. R.A. (1915) . Frequency distribution of the values of the correlation coefficient in samples from an indefinitely large population. We. 19. 507-521. Francis, V. , Korsch. B.M. & Morris. M.J. (1969) . Gaps in doctor- patient communication: Patients ' response to medical advice . WW. 2.3.9: 535'540- Fry. E. (1968) . A readability formula that saves time. W W. 11. 513-516. 575-578. Germain. R.D. (1985) . Beyond the internal - external continuum: The development of formal operational reasoning about control of reinforcements. W, z_Q, 939-947. Havighurst, R.J. (1951. W. New York: Longman' s Green. Hofmann. A.D. . Becker, R.D. . & Gabriel, H.P. (1976) . The Wt. London: Free Press, Collier- Macmillan . 112 Katz, E.R. . Kellerman. J. . & Siegel. S.E. (1980) . Behavioral distress in children with cancer undergoing medical procedures: Developmental considerations. W W. as. 356-365- Kellerman. J. . & Katz, E.R. (1977) . Adolescents with cancer: Theoretical . clinical and research issues. W W. 2. 127-131. Kellerman, J. , Zeltzer. L. . Ellenberg, L. . Dash. J. . & Rigler, D. (1980) . Psychological effects of illness in adolescence. I. Anxiety. self-esteem. and perception of control . W Pediatrics. 21(1) . 126-131. Kendall, P.C. . & Braswell, L. (1985) . mm W- New York: Guilford. Kendall, P.C. , & Korgeski. G.P. (1979) . Assessment and cognitive- behavioral interventions. W. .2. 1-21. Kimmel. D.C. . &Weiner. LB. (1985) . W W. Hillsdale. NJ: Lawrence Erlbaum Associates. Lazarus, R.S. . s. Folkman. S. (1984) . Coping and adaptation. In w.o. Gentry (Ed-) . Wiggins. (Pp- 282-325) . New York: Gilford. Leadbeater. B.J. . & Dionne. J.P. (1981) . The adolescent's use of formal operational thinking in solving problems related to identity resolution. AW, 1g, 111-121. 113 Levenson. P.M. . Copeland. D.R. . Morrow. J.R. . Pfefferbaum. B. , & Silberberg, Y. (1983) . Disparities in disease-related perceptions of adolescent cancer patients and their parents . MW. 5.. 33-45- Maddux. J.E. . Roberts. M.C. . Sledden, E.A. . &Wright. L. (1986). Developmental issues in health psychology. American Wilt. 51(1) . 25’34- McKinnely. J.P. . Fitzgerald. H.E. . & Strommen. E.A. (1982) q; ; ._..._;. i .; ,. m. . .; ,._. ;; -. ,,._ . ,. (Rev. ed>) . Homewood, IL: Dorsey Press. Moos. R.H. . a Tsu, V.D. (1977) . The crisis of physical illness: An overview. In R.H. Moos (Ed.) . WW (pp. 3-21) . New York: Plenum Press. Nagy. V.T. . & Wolfe. G.R. (1983) . Chronic illness and health locus of control beliefs. W W. 1(1) . 53-55- National Institutes of Health. (1984) . W teenagerexitheaneenmaerlsguide). (NIH Publication No. 84- 2460) . Washington. DC: 0.8. Government Printing Office. Neimark, E. D. (1980) . Intellectual development in the exceptional adolescent as viewed within a Piagetian framework. W. 1(2). 47-55- Olkin. I. (1967) . Correlations revisited. In J.C. Stanley (Ed.) , uo . .. :m- '1‘! :,!‘-1°! c.1- e . s .. ._...__ re (pp. 102-128) . Chicago: Rand McNalley. 114 Parcel. G.S. . & Meyer, M.P. (1978) . Development of an instrument to measure children ' 3 health locus of control . Health W. .§(2) . 149'159- Pfefferbaum. B. , 8. Levenson. P.M. (1982) . Adolescent cancer patient and physician responses to a questionnaire on Patient concerns. Wm. 1.32. 348-351 . Pfefferbaum, B. , Levenson, P.M. . 8. van Eys. J. (1982) . Comparison of physician and patient perceptions of communication issues . WW. 1:. 1080-1083. Potter. P.C. . & Roberts. M.C. (1984) . Children's perceptions of chronic illness: The roles of disease symptoms, cognitive development. and information. W W. 2. 13-27- WW- (1983) - East Lansing, MI: University Center for International Rehabilitation. Michigan State University . Rotter. J-B- (1954) - Winem- Englewood Cliffs. NJ: Prentice-Hall. Rotter. J.E. . Chance, J. . & Phares,E.J. . (Eds.) . (1972) . New York: Holt. Rinehart. & Winston. Salvia. J- . & Ysseldyke. «LE. (1985) . WW WIDE (3rd ed.) . Boston: Houghten Mifflin. Selman. 12.1» (1980) . W W New York: Academic Press . 115 Simeonsson, R.J. , Buckley. L. . 8. Monson, L. (1979) . Conceptions of illness causality in hospitalized children. M W291. 1. 77'84- Spinetta, J.J. . Elliott, J.S. . Hennessey. J.S. . Knapp, V.S. . Sheposh. J.P. . Sparta. S.N. . 8. Sprigle, R.P. (1982) . The pediatric psychologist ' s role in catastrophic illness: Research and clinical issues. In J.M. Tuma (Ed.) . W WW (1910- 165-237) . New York: Wiley. Strickland. B.R. (1978) . Internal-external expectancies and health related behaviors. W W. 5.5.. 1192-1211. Susman, B.J. . Hersh. S.P. . Nannis. E.D. , Strope, B.E. . Woodruff, P.J. . Pizzo. P.A. . & Levine. A.S. (1982) . Conceptions of cancer: The perspectives of child and adolescent patients and their families. WW, 1, 253-261. Susman, B.J. , Hollenbeck. A.R. , Nannis, E.D. . 8. Strope, B.M. (1980) . A developmental perspective on the psychosocial aspects of childhood cancer. In J. L. Schulman 8. J. Kupst (Ede-) . MW (pP- 129-142) - Springfield. IL: Charles C. Thomas. Tavormina. J. . Kastner. L.S. , Slater. P.M. . &Watt, S.C. (1976) . Chronically ill children: A psychologically and emotionally deviant population? WW. :1. 99-110. 116 Wallston. R.A. . Wallston. B.S. . & DeVellis. R. (1978) . Development of the Multidimensional Health Locus of Control (MHLC) scales. WW. 6(2) . 160-170- Willis. D.J. . Elliott. C.H. . aJay. S.M. (1982) . Psychological effects of physical illness and its concomitants. In J .M. Tuna (Edd . WWW (pp. 28-66) . New York: John Wiley & Sons. Zeltzer, L. . Kellerman. J. . Ellenberg, L. , Dash, J. , & Rigler. D. (1980) . Psychological effects of illness in adolescence. II . Impact of illness in adolescents-critical issues and coping styles. W. 21(1) . 132-138-