' ———‘__— w.— CONSISTENCY THEORY, ACTIVITY THEORY. AND DISENGAGEMENT THEORY: PERSONAL RESOURCES AND ENVIRONMENTAL DEMANDS AS PREDICTORS OF . CONTENTMENT IN THE AGING Dissertation for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY CAROLE OVERMIER BETTINGHAUS 1976 IIIIIIIIIIIIIIIIIIIIIIIOIIIIIIIIIIIIIIIIIIIIIII 0488 4998 This is to certify that the thesis entitled Consistency Theory, Activity Theory and Disengagement Theory: Personal Resources and Environmental Demands as Predictors of Contentment ‘ in the Aging presented by Carole Overmier Bettinghaus has been accepted towards fulfillment of the requirements for Ph.D. I degree“, Psychology_ Major professor [hue February 2, 1977 ' .5. .-.. E“ rib”) / .. '00. V P j - __n \ 147* .. m '- ffwi — ABSTRACT CONSISTENCY THEORY, ACTIVITY THEORY, AND DISENGAGEMENT THEORY: PERSONAL RESOURCES AND ENVIRONMENTAL DEMANDS AS PREDICTORS OF CONTENTMENT IN THE AGING By Carole Overmier Bettinghaus The purpose of this study is to test three sets of hypotheses representing three theories of successful aging: consistency theory, activity theory, and disengagement theory. Successful aging may be defined as growing older while maintaining a relatively high level of general satisfaction with one's life. Historically, the two competing theories have been activity theory and disengagement theory. The former theory states that con- tinuing middle aged activities is related to high satisfaction for the aging. The latter theory holds that disengagement amounts to withdrawal of the aging person and society from each other. In the course of this withdrawal the aging person accepts and, perhaps, desires decreased interaction. Disengagement theory claims that disengagement and satisfaction are positively related. Findings have been equivocal regarding these two theories. Recently a third theoretical position has been advanced by Gubrium, which we have labeled consistency theory. This theory holds that satisfaction of the elderly person will be high when personal Carole Overmier Bettinghaus resources are consistent with environmental demands, and satisfaction will be low when personal resources are inconsistent with environ- mental demands. A The data for this study were gathered in conjunction with the experimental evaluation of a health care service provided in private residences by a team of professionals. All data were obtained from persons who did not receive the experimental service. Data were collected in two interviews, separated by approximately six months. The sample consisted of 280 persons, 45 or older, having chronic physical impairment of various degrees of severity, who resided in a private residence at the time of the first and second interview. Hypotheses of the study were tested on the total sample and on the sample subdivided into a sample of l55 older persons of 65 or more, and a sample of 125 younger persons of 45 to 64. For each theory, hypotheses have been formed following four paradigms of the relationships between the “time" of the independent variables and the criterion. Between the first and second interview the sample size decreased considerably, due to experimental mortality which was largely true mortality. Three of the four paradigms utilize data from the reduced sample. To study the hypothesized effects of the theories we used a scale of contentment as a measure of satisfaction. The independent variables used were (l) age environment, (2) residential location of primary care giver, (3) spouse survival, (4) days of bed disability, (5) independence in activities of daily living, and (6) financial ability. Carole Overmier Bettinghaus . Each of the l2 hypotheses has been evaluated using five dif- ferent two-by-two fixed factor analyses having unequal cell frequen- cies. Both analysis of covariance and analysis of variance were computed using a least squares model. Age of the subject was the covariate for all four paradigms and prior score on contentment was the second covariate when the criterion was contentment measured at the second interview. Results of the study point to activity theory as the strong leader among the three theories. Consistency theory is second to activity theory. We were led to reject disengagement theory. In discussing these results, problems with choice of variables were identified which suggest these results may be artifactual. It is suggested that activity theory is tied to physical health, consis- tency theory is tied to social health, and disengagement theory is tied to psychological health. CONSISTENCY THEORY, ACTIVITY THEORY, AND DISENGAGEMENT THEORY: PERSONAL RESOURCES AND ENVIRONMENTAL DEMANDS AS PREDICTORS OF CONTENTMENT IN THE AGING By Carole Overmier Bettinghaus A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology I976 ACKNOWLEDGMENTS I wish to express my appreciation to the members of my commit- tee, Dr. Eugene Jacobson, Dr. Sidney Katz, Dr. William Mehrens, Dr. Lawrence Messe', and Dr. James Phillips. Dr. Phillips, who served as Chairman of my graduate committee and my dissertation com- mittee, deserves special thanks for many hours of careful, critical reading of early drafts. A special debt of gratitude is owed to the late Dr. Margaret Blenkner for introducing me to the challenge and fascination of study in the field of Gerontology. Other faculty members who have contrib- uted advice useful to the data analysis of the study include Dr. Mary Ellen McSweeney, Dr. William Schmidt, and Dr. Robert Sprafka. Although not members of my graduate committee, Dr. William Crano, Dr. Jeanne Gullahorn, and Dr. John E. Hunter have played instrumental roles in my earlier career as a doctoral candidate. Crucial help in providing software subroutines and facilitating retrieval of the data from the computer has been provided by B. K. Shrivastrava, M.A. Friendly support and constant cooperation were contributed by the immediate members of my family. Often tasks that would have been shared within the household were completely assumed by one or more of my children, Bruce Bettinghaus, Joyce Bettinghaus, or Karen Betting- haus Keating, or by my husband, Dr. Erwin P. Bettinghaus. Thanks are also due to my parents, Irma Overmier and Harry Overmier for the initial impetus toward educational achievement which they provided. ii TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES Chapter I. BACKGROUND AND RATIONALE . Introduction . . The Gerontological Theories The Activity Theory- Disengagement Theory Controversy . Reconciling the Controversy Correlates of Successful Aging Rationale of the Study. Consistency Theory . The Criterion Measure "Contentment" . The First Factor--Age Environment The Second Factor--Ability Resources . Chronological Relationships of Hypotheses-- Paradigms . . . . . . . . The Research Hypotheses . . The Consistency Hypotheses . The Activity Hypotheses . The Disengagement Hypotheses II. METHODOLOGY AND PROCEDURES Context of This Study . . . Identification of the Sample . Characteristics of the Sample Operationalizing the Variables Criterion as Contentment . Social Context as Age Environment Activity Resource as Physical Health Activity Resource as Social Support . . Activity Resource as Financial Ability . Paradigms for the Theories and Hypotheses General Paradigms for the Three Theories Paradigms for the Four Specific Hypotheses Page vi Chapter Page Analysis of the Data . . . . . . . . . . 64 Analysis of Covariance . . . . . . . . . 64 Analysis of Variance . . . . . . . . . . 65 Cell Means and Frequencies . . . . 66 Analysis of Data, Controlling for Age Group . . 66 III. RESULTS . . . . . . . . . . . . . . . 68 Introduction . . . . . . . . . 68 Results for the Total Sample . . . . . . . . 69 Descriptive Data . . . . . . . 69 Analysis of Covariance and Variance . . . . 76 Results for the Older (65 or Older) Sample . . . 85 Descriptive Data . . . . . . . 85 Analyses of Covariance and Variance . . . . 91 Results for the Younger (45 to less than 65). Sample . . . . . . . 99 Descriptive Data . . . . . . . . 99 Analyses of Covariance and Variance . . . . . 105 Summary of the Results . . . . . . . . . . 110 IV. DISCUSSION . . . . . . . . . . . . . . 116 Introduction . . . . . . . . 116 Discussion of the Present Study. . . . . 116 Sample Characteristics at Time #1 and Time #2 . 117 Characteristics of Older Sample and Younger Sample . . . . . . 119 Results for the Consistency Hypotheses . . . . 121 Results for the Activity Hypotheses . .- . . . 121 Results for the Disengagement Hypotheses . . . 122 Comparison of the Results for the Three Theories . . . . 122 Discussion of the Results for the Three Theories . . . 124 The Additional Finding of a Main Effect for Spouse Nonsurvival . . . . 131 Discussion of Suggestions for Future Research . . 132 Suggestions for Future Instrumentation . . . . 133 Suggestions for Data Collection Strategies . . 137 Suggestions for Analysis of Future Research . . 138 Suggestions for Sampling in Future Research . . 139 Future Research Questions . . . . . 141 Suggestions for Future Research Strategy . . . 145 Summary. . . . . . . . . . . . . . . 146 iv Chapter APPENDICES A. Comparison of Several Measures of Satisfaction/ Morale . . . . . B. Codebook and Data Collection Items for Variables Used in This Study . . . . . C. Data Results Not Reported in This Study LIST OF REFERENCES Page 149 159 183 194 Table IO. 11. 12. 13. 14. LIST OF TABLES Replicated Relationships Between Several Variables and Satisfaction . . . Characteristics of Subjects in Total Sample at Time #1 and #2 . . Characteristics of Primary Care Giver and Residence for Total Sample at Time #1 and #2 Characteristics of Subjects in Older Sample at Time #1 and #2 . Characteristics of Primary Care Giver and Residence for Older Sample at Time #1 and #2 Characteristics of Subjects in Younger Sample at Time #1 and #2 . . . . . . . . Characteristics of Primary Care Giver and Residence for Younger Sample at Time #1 and #2 . The f, M, and §Q_of Contentment at Time #1 for Total Sample--(H01C, A, D) . . . . . . . . . The f, M, and §D_of Contentment at Time #2 for Total Sample--(H02C, A, D) . . . . . . . . . . The f, M, and §D_of Contentment at Time #2 for Total Sample--(Ho3C, A, D) . . . . . . . . . The f, M, and SD of Contentment at Time #1 for Total Sample--(Ho4C, A, D) . . . . . . . . . Analysis of Covariance and Variance for Total Sample--(HolC, A, D) . Analysis of Covariance and Variance for Total Sample--(H02C, A, D) . . Analysis of Covariance and Variance for Total Sample--(H03C, A, D) . vi Page 15 38 4O 43 44 45 46 70 71 72 73 77 78 79 Table Page 15. Analysis of Covariance and Variance for Total Sample--(Ho4C, A, D) . . . . . . . . . . . 8O 16. One-Nay Analyses of Covariance-Variance--Age Environment for Total Sample . . . . . . . . . 81 17. Significant Main Effects Found in Total Sample . . . 84 18. The f, M, and SD of Contentment at Time #1 for Older SampTe--(HolC, A, D) . . . . . . . . . 86 19. The f, M, and SD of Contentment at Time #2 for Older SampTe--(H02C, A, D) . . . . 87 20. The f, M, and SD of Contentment at Time #2 for Older SampTe--(H03C, A, D) . . . 88 21. The f, M, and SD of Contentment at Time #1 for Older SampTe--(Ho4C, ,D) . . . 89 22. Analysis of Covariance and Variance for Older Sample--(HolC, A, D) . . . . . . . . . . . 92 23. Analysis of Covariance and Variance for Older Sample--(H02C, A, D) . . . . . . . . . . . 93 24. Analysis of Covariance and Variance for Older Sample--(H03C, A, D) . . . . . . . . . . . 94 25. Analysis of Covariance and Variance for Older Sample--(H04C, A, D) . . . . . . . . . . . 95 26. One-Nay Analyses of Covariance-Variance--Age Environment for Older Sample . . . . . . . . . 96 27. Significant Main Effects Found in Older Sample . . . 98 28. The f, M, and SD of Contentment at Time #1 for Younger TSample--(HolC, A, D). . . . . . . . . 100 29. The f, M, and §Q_of Contentment at Time #2 for Younger Sample--(H02C, A, D) . . . . . . . . . 101 30. The f, M, and SD of Contentment at Time #2 for Younger TSample--(H03C, A, D). . . . . . . . . 102 31. The f, M, and SD of Contentment at Time .#1 for Younger TSample--(H04C, A, D). . . . . . . . 103 vii Table Page 32. Analysis of Covariance and Variance for Younger Sample--(HolC, A, D) . . . . . . . . . . . 106 33. Analysis of Covariance and Variance for Younger Sample--(H02C, A, D) . . . . . . . . . . . 107 34. Analysis of Covariance and Variance for Younger Sample--(H03C, A, D) . . . . . . . . . . . 108 35. Analysis of Covariance and Variance for Younger Sample--(Ho4C, A, D) . . . . . . . . . . . 109 36. One-Nay Analyses of Covariance—Variance--Age Environment for Younger Sample . . . . . . . . 111 37. Significant Main Effects Found in Younger Sample . . 112 38. Summary of Findings Regarding Interaction Effects . 113 39. Summary of Significant Findings Regarding Main Effects . . . . . . . . . . . . . . . . 114 A1. List of Various Measures of Satisfaction/Morale . . 152 A2. Intercorrelations for Several Measures of Satisfaction/Morale . . . . . . . . . . . . 154 C1. Data Results Not Reported for Total Sample-- (Ho3C, A, D) . . . . . . . . . . . 184 C2. Data Results Not Reported for Older Sample-- (HolC, A, D) . . . . . . . . . . . 185 C3. Data Results Not Reported for Older Sample-- (HoZC, A, D) . . . . . . . . . . . 186 C4. Data Results Not Reported for Older Sample-- (Ho3C, A, D) . . . . . . . . . . . 187 c5. Data Results Not Reported for Older Sample-- (Ho4C, A, D) . . . . . . . .. . . . 188 C6. Data Results Not Reported for Younger Sample-- (HolC, A, D) . . . . . . . . . . . . 189 C7. Data Results Not Reported for Younger Sample-- (HoZC, A, D) . . . . . . . . . . . . . .. 190 viii TabIe Page C8. Data Results Not Reported for Younger Sample-- (Ho3C, A, D) . . . . . 191 C9. Data Results Not Reported for Younger Sample-- (Ho4C, A, D) . . . . . . . . . . . . . 192 C10. Data Results Not Reported-~One-Nay Analyses-- Older Sample . . . . . . . . 193 ix Figure V0014) 10. 11. 12. 13. 14. 15. LIST OF FIGURES Types of older persons defined by activity and engagement . . . . Types of older persons defined by social and psychological engagement . . . . Direction of several variables related to satis- faction reported by Adams General paradigm for consistency theory . General paradigm for activity theory . General paradigm for disengagement theory Immediate cross sectional hypothesis--consistency theory paradigm . . . . . . . . . Immediate cross sectional hypothesis-~activity theory paradigm . . . . . . . Immediate cross sectional hypothesis--disengagement theory paradigm . . . . . . . . . . Delayed cross sectional hypothesis--consistency theory paradigm . . . . . . . . . Delayed cross sectional hypothesis--activity theory paradigm . . . . . . . . Delayed cross sectional hypothesis--disengagement theory paradigm . . . . . . . . . Chronological lag hypothesis--consistency theory paradigm . . . . . . . . . . Chronological lag hypothesis--activity theory paradigm . Chronological lag hypothesis--disengagement theory paradigm . . . . . . . . . Page 13 53 54 54 55 56 56 58 58 59 60 60 61 Figure Page 16. Chronological anticipation hypothesis--consistency theory paradigm . . . . . . . . . . . . 62 17. Chronological anticipation hypothesis--activity theory paradigm . . . . . . . . . . . . 62 18. Chronological anticipation hypothesis--disengagement theory paradigm . . . . . . . . . . . . 63 19. Comparison of significant findings for three theories by four hypotheses . . . . . . . . . . . . 123 20. Types of appropriate variables for type of theory by type of health . . . 128 81. Three demographic variables: sex, ethnic group, marital status . . . . . . . . . . . . . 160 B2. Highest occupation variable . . . . . . . . . 161 83. Educational attainment variable . . . . . . . 162 B4. Variable of PCG sex in relation to participant . . 163 B5. Variable of PCG relationship to participant . . . 164 B6. Length of time at residence variable . . . . . . 165 B7. Number of cohabitants variable . . . . . . . . 166 B8. Owner of residence variable . . . . . . . . . 167 89. Head of household variable . . . . . . . . . 168 810. Criterion variable: contentment . . . . . . . 169 811. Factor A: age environment . . . . . . . . . 171 812. Factor 8: residential location of PCG . . . . . 172 B13. Factor C: spouse surviva1--time #1 . . . . . . 173 814. Factor C: spouse survival--time #2 . . . . . . 174 D: B15. Factor days of bed disability . . . . . . . 175 816. Factor E: independence of activities of daily living . . . . . . . . . . . . . . . . 176 xi Figure Page 817. Factor F: financial ability--annua1 family income component . . . . . . . . . . . . . . 180 818. Factor E: financial ability--index of economic dependence component . . . . . . . . . . . 181 xii CHAPTER I BACKGROUND AND RATIONALE Introduction In recent years social scientists have focused their attention on several minorities. Most of these groups have attained prominence as a result of audible demands either for them or by them, rather than on the basis of their numerical growth. One minority group, however, has gained new attention not by speaking louder, but by increasing its membership. This is the group identified as the elderly. The increased membership of older persons in today's society is related to increases in economic surpluses which have made possible the alleviation of body-breaking toil and physical fatigue through labor-saving machinery and the 40 hour week, and the virtual elimi- nation of contagious diseases through the charitable and tax-supported funding of medical research. As the elderly have become more numer- ous, they have gained added attention from business, government, edu- cation, and behavioral science establishments. The field of gerontology has emerged, and has stimulated con- troversy over explanations of "successful" aging. The two major Funding for data collection and data analysis for this study has been provided under the auspices of a National Institute of Health Grant #HS-01059-03 administered through the Department of Health, Edu- cation, and Welfare, Public Health Service, Sidney Katz, M.D., princi- pal investigator. Access to the data has been granted by Dr. Katz, and facilitated by the cooperation and accommodation of Sr. Mary Honora Kroger, Ph.D., current project director, and Dr. Margaret Blenkner, past project director. 1 theoretical positions which have been advanced are referred to as "activity theory" and "disengagement theory." Recent attempts to resolve the controversy over the usefulness of these two theories have led to the advancement of a third type of theory, which we have labeled "consistency theory." Gubrium (1972) has been the major proponent of a consistency approach to the explanation of successful aging. Activity theory is the more familiar of the two major the- ories. Its popularity may, in part, be a consequence of cultural values for progress, industry, and achievement. Many similarly worded cultural expressions provide reinforcement of these values, and the part played by activity in their achievement. In addition, there are popular beliefs that "staying active" will overcome most situations which have negative value in the culture, such as grief, fear, loneliness, worry, conflict, and frustration. In the early years of gerontological research, activity theory was an unstated assumption, and remains pervasive. Disengagement theory, as stated by Cumming, Dean, and Newell (1958) and elaborated by Cumming and Henry (1961), has fewer cultural roots and is less well understood. Disengagement is char- acterized by withdrawal of the aging person and society from each other. In the course of this withdrawal the aging person accepts and, perhaps, desires decreased interaction. Disengagement is classi- fied as a concomitant to successful aging, by these authors. In the ensuing eighteen years since this controversy began, the activity theory has been stated more formally and the two theories have been discussed by a number of authors (Birren, et al., 1972; Blenkner, 1958; Bortner, 1966; Brown, 1974; Carp, 1968; Cumming, Dean, Newell, & McCaffrey, 1960; Cumming, 1963; Havighurst, 1968; Lemon, et al., 1972; Lipman & Smith, 1968; Lowenthal, 1965; Lowenthal & Boler, 1965; Maddox, 1963, 1964, 1965; Martin, 1973; Neugarten, 1972; Prasad, 1964; Rose, 1964; Tallmer & Kutner, 1970; Tissue, 1968; Youmans, 1969; among others). The consistency position has been discussed more recently by a smaller group of authors (Loeb, Pincus & Mueler, 1966; Tissue, 1971; Gubrium, 1972; Martin, 1973). No one has, as yet, attempted to compare all three theories of successful aging. The purpose of this study is to test each of three sets of hypotheses. Each set of hypotheses will be designed to represent one of the theories: consistency theory, activity theory, or disengagement theory. Successful aging which is referenced repeatedly in the disen- gagement-activity controversy has been operationalized in various ways. Typically, however, it is defined as the achievement, by an older person, of a relatively high score on some measure of morale, satisfaction, or happiness. Thus, the criterion for success in aging has been hedonistic. In earlier phases of the activity/disengagement controversy, other criteria were used. These have been listed by Havighurst (1961) as: (1) A way of life thatis socially desirable for this age group (normatively defined), (2) Maintenance of middle- -age activity, (3) A feeling of satisfaction with one' 5 present status and activities (nonindependent of disengagement and/or activity), and (4) A feeling of happiness and satisfaction with one '5 life (in general) (pp. 910). In the present study, the criterion measure for successful aging is a five-item scale of "contentment" which was developed by Blenkner, Bloom, and Weber (1964). It is an instrument for measurement of general happiness and satisfaction. This report is organized into four chapters. The first chap- ter deals with an introduction to the matters and materials of the study; including sections on: (1) the gerontological theories of successful aging, (2) empirical findings related to successful aging, (3) rationale of the study, and (4) hypotheses of the study. The second chapter treats the methodology and procedures of the study, including sections on: (1) the context in which the study was con- ducted, (2) identification of the sample, (3) characteristics of the sample, (4) operationalizing the variables, (5) research paradigms for the hypotheses, and (6) analysis of the data. The third chapter reports the results from (1) the total sample, (2) the older sample, and (3) the younger sample. For each sample, a report is made of the frequencies, and means and standard deviations of the criterion variable, with the results which are reportable for (1) analysis of covariance and (2) analysis of vari- ance. The feurth chapter is a discussion generated by the results; and is subdivided into (1) a discussion of the present study, with a section for each gerontological theory, and (2) a discussion of a variety of suggestions for future research aimed at deciphering the process of successful aging. The Gerontological Theories The Activity Theory-Disengagement Theory Controversy Much of the research conducted has been equivocal, and research purporting to support either activity or disengagement theories can be questioned on the basis of nonrepresentativeness of the sample, questionable research methodology or noncomparability of measures (Birren, et al., 1972; Blenkner, 1958; Bortner, 1966; Maddox, 1963, 1964, 1965; Rose, 1964; Tallmer & Kutner, 1970; Tibbitts & Schmelzer, 1965; & Youmans, 1969). Some authors have ably described activity theory, while others have covered disengagement theory thoroughly. However, the clearest contrast between the two theories has been presented by Havighurst, Neugarten and Tobin (1968), who emphasize that neither theory fully accounts for all the empirical evidence, although each of the two older theories can claim some support from findings. These authors view "activity" as a construct for explaining the later years of life as a direct extension of the active life style of middle age. In opposition to this, they view "disengagement" as a developmental construct for explaining the later years of life as a life style changed from that of middle age. They also make the important distinction that activity theory views the older person as an antagonist to society, struggling to maintain an activity pattern of a middle aged person, in the face of the withdrawal of society from contact with the older person. In the disengagement view they characterize the relationship between the older person and society as mutual and cooperating, with both the older person and society working to change the life style of the elderly person after middle age. Maddox (1964) emphasizes some assumptions of the disengage- ment theory which have been neglected by Havighurst and his collab- orators: "(1) that the process of social and psychological withdrawal is modal for the aging population, (2) that this process is both intrinsic and inevitable, and (3) that the disengagement process is not only a correlate of successful aging but also probably a condition of it" (p. 80). Most authors have assumed that support for disengagement is equivalent to nonsupport for the activity formulation. It is possi- ble that failure to recognize an area where the two constructs do not completely overlap has led to some of the inconclusive research results. The two constructs of activity and disengagement may be treated as two dimensions. The relationship of the two dimensions .may be schematized as orthogonal, and each dimension may be simplified to include only polar values, as in Figure l. Engagement Dimension Activity Dimension Disengaged Persons Engaged Persons Active Active and Active and Persons Disengaged Persons Engaged Persons Inactive Inactive and Inactive and Persons Disengaged Persons Engaged Persons Figure 1. Types of older persons defined by activity and engagement. In her writing Cumming (1963) has called attention to the existence of persons who although disengaged, remain active, while doubting that it is possible to be engaged and inactive at the same time. Maddox (1965), in discussing the problem, identified the group of persons whose activity may decrease, while they remain engaged, as one overlooked by researchers. Failure to recognize that some older persons belong to each of these "mixed" groups, would have led researchers to treat activity and engagement as a single dimension because they saw the two constructs as strongly correlated. Havighurst, Neugarten, and Tobin (1968) point to another dis- tinction regarding disengagement and activity theory which may be related to the types of persons defined in Figure 1. They distinguish between social engagement and psychological engagement, equating social engagement with easily observed social interaction, activity, and participation, while equating psychological engagement with sub- jectively perceived involvement, emotional investment, preoccupation, and commitment. For this discussion, it is convenient to relabel the dimensions of Figure 1, as has been done in Figure 2. PsychOIogical Engagement Dimension Social Engagement Psychologically Psychologically Dimension Disengaged Persons Engaged Persons Socially Socially Engaged, Socially and Engaged Psychologically Psychologically Persons Disengaged Persons Engaged Persons Socially Socially and Socially Disengaged, Disengaged Psychologically Psychologically Persons Disengaged Persons Engaged Persons Figure 2- Types Of older persons defined by social and psychological engagement. They go on to indicate that activity theory assumes that social disengagement can occur without psychological disengagement, and this is consistent with the stand identified above by Maddox. They also indicate that it is assumed in disengagement theory the two types of disengagement accompany each other, or that psychological disengagement precedes social disengagement. This is based upon the developmental aspect of disengagement theory. Such assumptions are consistent with the stand identified above by Cumming. The activity disengagement controversy has thrived in the field of gerontology recently. On the one hand the two theories were defined as opposite in type and in their predictions for the older person and society. On the other hand the researchers involved over- looked the existence of older persons who embodied "mixed" levels of engagement and activity. Both facts led to an antagonistic view and an antagonistic stance in research strategy. Finally, the theoretical picture was confused further by a distinction between social and psy- chological disengagement, with social engagement appearing to be not greatly different from activity. Reconciling the Controversy In discussing his results, Martin (1973) offers what appears to be the most empirically based discussion of the reconciliation of the two theories based upon the relation of the two dimensions of activity and engagement. Martin describes a situation in which a group of older persons have moved, or plan to move, into a retirement . community which will separate them from their family relations (disengagement), and which they all expect will give them a lot of activities with their age peers (activity). This example seems equivalent to maintaining a high level of activity in a constricted number of roles. Gubrium (1972) proposes a method for resolving the conflict between the activity and disengagement theoretical positions. His pr0posal is based partially upon an earlier (1970) study and partially upon an important underlying assumption: . . . let us assume that persons feel most satisfied with themselves and their living conditions when there is con- gruency between what is expected of them by others of sig- nificance and what they may expect of themselves. Any inconsistency between these two bodies of expectations will be said to lead to life dissatisfaction among the aged (pp. 282-3). For Gubrium, who the significant others are depends upon whether or not the person's age environment is concentrated, or non- concentrated, and what they may expect of themselves is dependent upon the person's behavior flexibility, measured in terms of health, solvency, and ongoing social support. Gubrium essentially discusses a two pronged position: (1) When the demands of the environment (of an older person) for flexi- bility of the person, are consistent with the person's ability to be flexible, morale will be high. (2) When there is inconsistency between the environment's demands for flexibility and the person's ability to be flexible, morale will be low. The underlying assumption indicates that Gubrium has adopted a consistency position to explain the occurrence of successful aging. Although he is the first to advance exactly this statement of a 10 consistency model for successful aging, other authors have fore- shadowed his idea. Tissue, (1971) proposed that consistency of life space and "disengagement potential" influenced morale. Loeb, Pincus, and Mueller (1966) speak of consistency between living space/social space and variables such as time perspective, "biography," physiologi- cal status and psychological disposition. These authors also mention that any reorganization of these variables occurs gradually for an individual. The model which Gubrium (1972) presents indicates a construct of the elderly person's environment which includes (1) a social con- text and (2) an individual context. In defining "the social context" of the elderly person's environment, Gubrium refers to "normative out- comes of social homogeneity, residential proximity, and local protec- tiveness" (p. 282). He then outlines the process by which he theorizes these normative outcomes arise: . as the local environments of the aged become concen- trated with old people, it is likely that local activity norms become age-linked, i.e., persons' expectations on each other's behavior become rooted in relatively common rather than diverse experiences. If such age concentrated environments are proximate as well as age-homogeneous and exhibit relative continuity as such, then what Rose calls a "sub-culture of aging" will probably emerge. The behavioral implication of such a subculture is that the activity that is expected of persons, sanctioned, or labeled as deviant, is significantly different from that in age-heterogeneous locales. . . In highly heterogeneous environments the variety of situations that persons are likely to encounter are maximal. This implies that any person must have a sufficient command of himself to "make out" as Goffman states, from one situation to the next. The resources he possesses then must be suffi- ciently endowed so as to allow him to fulfill a variety of expectations. Now what of homogeneous environments? The variety of situations with which persons are confronted here are quite narrow'h1terms of demands on flexibility. Facility in one situation is likely to mean facility in most (p. 282). 11 Gubrium goes on to explain that "the individual context" of the elderly person's environment refers to "activity resources such as health, solvency, and social support that influence behavior flexi- bility." The Gubrium model will be further specified in the sections of this chapter which deal with the rationale and research hypotheses of the present study. Having discussed the theoretical positions regarding the variables related to successful aging, we now turn our attention to the reported empirical findings. Many of these findings have been reported without a theoretical framework. Correlates of Successful Aging Adams (1971) claims that the concepts of satisfaction, per- sonal adjustment, positive self-concept, self-esteem, and morale have been used interchangeably to operationalize successful aging. Because of this, the range of variables which have been correlated with successful aging is quite large. A content analysis of eight articles, completed in conjunction with this study, yielded a set of 59 variables related to successful aging. Of these, 38 were listed only once among the eight articles. Among the 21 variables which were reported in more than one of the eight articles there was poor agreement between authors regard- ing the type of classification to be assigned to each variable (Adams, 1971; Bortner & Hultsch, 1970; Edward & Klemmack, 1973; Lawton, 1972; Maddox, 1963; Palmore & Luikhart, 1972; Spreitzer & Snyder, 1974; Tobin & Neugarten, 1961). 12 In 1971, Adams prepared a review article of the "Correlates of Satisfaction Among the Elderly." Figure 3 is reproduced from that article. It summarizes the complexity of reported findings as a simple + or - sign, indicating the direction of the relationship established by earlier research, between specific variables and the variable of satisfaction. In such a summary approach, some important details are lost, such as the type of statistical test used, the strength of the relation, the type of measurement used for the varia- bles and the number of replications represented. Table l is an update of the review done by Adams, adding data from six additional studies which he had not included in his review (Bortner & Hultsch, 1970; Edwards & Klemmack, 1973; Goodman, Dye, Harel & Bley, 1971; Maddox, 1963; Palmore & Luikhart, 1972; Spreitzer & Snyder, 1974). In Table 1, column A indicates the number of replications of positive (+) or negative (-) findings reported in the Adams review article. Each of the six additional columns report the value of findings reported in one of the six additional studies. The total column represents the sum of replicated positive findings for each variable. A positive finding is a positive rela- tion (not necessarily a correlation) between successful aging and the variable of interest. All the variables from the Adams study for which findings were replicated have been included in Table 1, along with a few selected additional variables. This data on replication of findings gives us an understanding of how confident we may be in each variable as a reliable "correlate" 13 Selected Correlates of Satisfaction, Personal Adjustment, Positive Self-Concept, Self- Esteem, "Morale," or Other Indicators of Psychological Well-Being Biological Correlates: (+) good health (Jeffers & Nichols, 1961; Loeb et al., 1963; Marshall & Eteng, 1970) I-I physical disability (Lowenthal & Boler, 1965) - advancing age (Kutner et al., 1956) (O) advancing age (Maddox & Eisdorfer, 1962; Pihlblad & McNamara, 1965) (-) to age 75 or 80 (O or +) thereafter (Loeb et al., 1963) Psychological Correlates: +) perception of health as "good" (Hansen & Yoshioka, 1962) (-) perception of age as "old" (Hansen & Yoshioka, 1962; Phillips, 1961) (-) perception of life space as contracting (Lipman, 1961; Tobin & Neugarten, 1961) (-) perception of relative deprivation (Phillips, 1961) (-) feeling of inadequacy by males or of rejection by females (Lieberman, 1960) (+) "vocabulary of motives" to justify low status (Gillespie, 1968) (+) favorable pre-retirement attitude (Thompson, 1958) (+) accurate pre-conception of retirement (Thompson, 1958) (+) belief in afterlife (Jeffers & Nichols, 1961) Sociological Correlates: Personal Characteristics (0) rural-urban residence (Hansen & Yoshioka, 1962) (+) high socio-economic status (Kutner et al., 1956) (+) high)education (Hansen & Yoshioka, 1962; Marshall & Eteng, 1970 (+) high income (Hansen & Yoshioka, 1962; Marshall & Eteng, 1970; Thompson, Streib, & Kosa, 1963) (+) income maintenance (Lloyd, 1955; Loeb et al., 1963) (+) income adequacy (Hansen & Yoshioka, 1962; Thompson, 1958) (+) home ownership (Hansen & Yoshioka, 1962) +) supported independence from family (Townsend, 1963) (+) living alone, but not isolated (Loeb et al., 1963) Figure 3. Direction of several variables related to satisfaction reported by Adams. (From "Correlates of Satisfaction Among the Elderly" by D. L. Adams, The Gerontologist, 1971, 11, 66.) 14 'Sociological Correlates: Roles and Role Changes (+) (+ (+) (+) + + + + + +Ilo++ AAAAAA AAA A A AA continuity of life styles (Williams & Wirths, 1965) ) retaining pastpatterns of living (Zborowski & Eyde, 1962) higher role counts (Lipman & Smith, 1968; Tobin & Neugarten, 1961 large social life space (Lipman, 1961; Tobin & Neugarten, 1961 being married (Hansen & Yoshioka, 1962; Kutner et al., 1956) widowhood (Lopata, 1968; Lowenthal, 1965) being employed (Hansen & Yoshioka, 1962; Kutner et al., 1956) retirement (Kutner et al., 1956; Lipman, 1961; Loeb et al., 1963; Lowenthal, 1965; Thompson, 1958) retirement for females (Loeb et al., 1963) retirement for non-whites (Lloyd, 1955) length of retirement to 3-7 years (-), thereafter (Marshall & Eteng, 1970) if made preparations for retirement (Marshall & Eteng, 1970) if retirement is voluntary (Thompson et al., 1963) if retirement is voluntary (Lowenthal, 1965) if reluctant to retire (Thompson, 1958) if retired because of "poor health" (Marshall & Eteng, 1970) if household tasks are assumed, by males, after retirement ‘ (Lipman, 1961) Sociological Correlates: Social Relations and Activities (+) high level of interaction (Anderson, 1967b; Lipman, 1961; Tobin, 1961) high level of engagement (Lipman, 1961) high level of activity (Maddox & Eisdorfer, 1962) high level of social relations (Davis, 1962; Rosow, 1967) high age density of neighborhood (Rosow, 1967) if member of a reference group (Phillips, 1961) if in a "useful contribution climate" (Filer & O'Connell, 1962) high friendship association (Lemon et al., 1969; Lloyd, 1955) loss of friends (Lopata, 1968) inability to make new friends (Lopata, 1968) if satisfied with amount of contact with friends and relatives (Loeb et al., 1963) high interpersonal relations with children, relatives and friends (Kutner et al., 1956) if member of family group (Townsend, 1963) level of association with relatives, neighbors, or formal or solitary activities (Lemon et al., 1969) Figure 3. (Cont'd). 15 TABLE 1: Replicated Relationships Between Several Variables and Satisfaction. A B E G L M P 5' Total Demographic Variables Employed/Retired +7 0 +1 +1 0 +9 Income Level +3 0 +1 0 0 +4 Marital Status +4 0 0 0 +4 Educational Level +2 0 +1 0 +3 Occupation Type 0 +1 +1 Socio-Economic Level +1 +1 Church Attendance +1 +1 Age -2 O O O -1 O -3 Sex (of Self) O O O -l O -1 Family Size -1 O -1 Rural/Urban O O 0 Race (of Self) O 0 Health Variables Actual Health +3 +1 +4 Physical Disability -1 -1 +1 -2+1 Self-Perception Variables Perception of Financial Adequacy +4 +1 +1 +6 Self-Perceived Health +1 +1 +1 +1 +1 +5 Self-Perceived Age +2 +2 Self-Feelings +1 +1 Self-Perceived Activity +1 +1 Psychological Variables Feelings of Usefulness +1 +1 Absence of Clinical Depression +1 +1 Belief in Internal Control +1 +1 Social Variables Interaction Index +3 +3 Friendship Association +2 +1 +3 Activity' +1 +1 +1 +3 Role Count +2 +2 Life Space +2 0 0 +2 Interpersonal Relations +1 +1 Organizational Activity +1 +1 Contact with Others 0 O Note. A=Adams, 1971 B=Bortner & Hultsch, 1970 E=Edwards & Klemmack, 1973 G=Goodman, Dye, Harel, 81 Bley, 1971 L=Lawton, 1966 M=Maddox, 1963 P=Palmore & Luikhart, 1972 S=Spreitzer & Snyder, 1974 16 of successful aging. We can have some confidence in the reliability of: (1) employment status, (2) self-perceived health, (3) self- perception of financial adequacy, (4) actual income level, and (5) marital status as correlates of satisfaction, since each of these has been supported in four or more pieces of independent research. Rationale of the Study Consistency Theory This study proposes to approach the study of successful aging by testing the consistency model suggested by Gubrium (1972), against the disengagement and activity theories of aging which were previously summarized. Gubrium's model suggests four postulates, although he does not state them explicitly. In these postulates age environment is one factor, the second factor is activity resources and satisfac- tion/morale is the criterion. 1. Elderly persons in nonconcentrated, demanding age environments with high activity resources will exhibit high satisfaction/morale (consistent situa- tion). 2. Elderly persons in concentrated, nondemanding age environments with low activity resources will exhibit high satisfaction/morale (consistent situation). 3. Elderly persons in nonconcentrated, demanding age environments with low activity resources will exhibit low satisfaction/morale (inconsistent situation). 4. Elderly persons in concentrated, nondemanding age environments with high activity resources will exhibit low satisfaction/morale (inconsistent situation). ‘ He reasons that inconsistency is related to low satisfaction/ morale in the situation found in postulate three because the older person is unable to perform as a "typical member" of the age l7 subculture. He describes persons referred to in postulate four as "bitter and humiliated." It appears that Gubrium has been strongly influenced by the theoretical position of symbolic interaction espoused by Arnold Rose (1964). His proposed consistency model and attendant postulates are based upon the symbolic interaction of an aging person with his/her environment. Another indication of the influence of a symbolic inter- actionist framework is his warning that under the condition in which interaction is inoperative, consistency may also be inoperative. In reference to his basic assumption, which was stated earlier, he says: It is safe to make such an assumption provided that the situa- tion of self-regard is the same as the situation in which persons experience the expectations of others referring to self. If these situations arerurt the same, then the costs to self-conception of any changes in others' definitions of self may easily be nil. For an old person, what this infers is that if he commits his behavior and orients his mind to others in his locale, then how they conceive of him will influence his action. But, on the other hand, if "he's in it but not with it,"'UHy1thenechanism of congruency may be inoperative (p. 283). In addition to the general theory of symbolic interaction, sup- port for the proposed consistency model Could also be interpreted as giving support to the general behavioral theory of relative deprivation. Sherwood and Nadelson (1968), in testing such a theory as predictive of "despair" in the elderly, found it to be more predictive of despair than a developmental theory. Despair may be conceived of as a polar opposite to contentment. They pointed to three postulates of relative deprivation theory: The basic postulate of relative deprivation theory is that the greater the perceived differences between current cir- cumstances and more favorable circumstances of peers in one's reference group, the greater the feelings of deprivation lead1ng to despair, low morale and misery. . . . The theory 18 of relative deprivation . . . postulates the importance of introspection at almost every point in a person's life . . . it is possible to postualte feelings of relative deprivation . in terms of the perceived difference between one's own present unfavorable circumstances and previous success . . . (and) the greater the disparity between anticipated degree of loss and actual perceived loss, the greater the impact on feelings of deprivation and resultant despair (pp. 414-416). Should the present test of a consistency model for the pre- diction of level of satisfaction/morale prove successful, it will remain impossible to decipher whether the outcome is related to external pressures for consistency such as are treated in symbolic and social interaction, or internal pressure for consistency such as are treated in perception, cognition, and introspection about relative deprivation. In all likelihood, both internal and external processes are necessary to obtain a consistent outcome for an indi- vidual. This study will undertake to compare the applicability of consistency, disengagement, and activity theories of successful aging. To do so, each theory will be tested for ability to predict degree of successful aging, measured as contentment. The Criterion Measure "Contentment" Havighurst (1961) has identified at least four different typologies of criteria of "successful aging." It is the criterion of "a feeling of (general) happiness and satisfaction with one's life" (p. 10) which has become the standard utilitarian criterion for successful aging in the majority of studies. However, adoption of a standard concept of satisfaction/morale as a criterion for successful aging is only the first step toward clarification. No standard data collection instrument for assessment of general satisfaction has been 19 widely accepted. Adams (1971) summarized the confusion about the operational definition of general satisfaction: . . the difficulty of trying to assess "individual well- being" has resulted in a variety of concepts, definitions, and measurements, most of which are related either semantid cally or empirically, but some of which are culturally or subculturally biased. Among these are the concepts of sat- isfaction, happiness, morale, successful aging, adjustment, adaptation and positive self-image. The measurements have ranged from observation of overt behavior to self-reporting indices to thematic apperception interpretations (p. 64). In this study, we have adopted a five item measure of satis- faction/morale, the contentment scale developed by Blenkner, Bloom, and Weber (1964). For a discussion of the many other measures cur- rently in use for measurement of satisfaction/morale, see Appendix A. Blenkner and her collaborators (1964) obtained a concurrent validity of .69 for the contentment scale in relation to a five place observer's rating scale using independent raters. This scale is~ reported by Bloom (1975) to have a correlation of .29 with Powell Lawton's (1972) morale scale. The scale's developers found a reliability of .88 (KR formula #21) for an earlier 19 item version of the contentment scale. Use of the Spearman Brown Prophecy Formula reveals that the expected relia- bility for the five item version of the contentment scale would be .65. In the hypotheses which follow, "contentment" will be operation- alized uSing this scale. The First Factor--Age Environment In this study data on age environment was available only for the residential environment. Two qualities of each person's age 20 environment are the structure and the process which are involved. Structure of the age environment is classified as "concentrated" if all the persons in the age environment are of an age similar to the subject, and classified as "nonconcentrated" if any of the persons in the age environment are of a dissimilar age. Process in the age environment is classified as "demanding" if living in such an environ- ment demands many responsibilities or duties of the subject, and classified as "nondemanding" if living in such an environment demands few responsibilities and duties of the subject. The combination of structure and process creates four poten- tial cross-classifications: (l) concentrated and demanding, (2) con- centrated and nondemanding, (3) nonconcentrated and demanding, and (4) nonconcentrated and nondemanding. We have predicted that there will be a strong negative correlation between structure and process, with an increase in the similarity of age related to a decrease in demands, and a decrease in the similarity of age related to an increase in demands. For that reason, we have combined the first and second cross-classification into a classification of concentrated and nondemanding, and we have combined the third and fourth cross- classification into a classification of nonconcentrated and demanding. Coding into these two classifications is done on the basis of the age structure within the subject's residence, and the process or demand characteristic is assumed to ride on the coat tails of the age struc- ture. 21 The Second Factor-~Ability Resources We have followed Gubrium's suggestion that health, solvency, and social support are the three types of variables which may be considered as ability resources. It was necessary that we work with available measures from an existing data set. Within that data set, two measures have been identified as measures of physical health. One of these measures iS'UKznumber of days within the past two weeks that the subject has spent all or most of the day in bed. This is akin to a measure of "sick days" and few such days should indicate a high level of health, while more such days should indicate a low level of health. The other of these measures is the number of activities of daily living the subject needs the aid of another person to accomplish. The six activities of daily living are bathing, dressing, transferring from bed to chair or from chair to standing, self-care in toileting, continence of bowel and bladder, and self-care in eating. To the extent that the subject needs the aid of another, this is a measure of disability. To the extent that help is not needed, this may be perceived as a measure of health or physical well-being. Within the data set two measures have been identified as measures of social support. Both measures are natural dichotomies. The first of these fiswhether or not the subject has a living spouse. Such a resource is perceived as a physical, psychological, and finan- cial support as well as a social support. The second of these is whether the person who serves as primary care giver to the subject lives in the same building, or in a different building. A primary 22 care giver was defined as the relative or friend giving the most direct help to the subject with matters of personal care and household tasks. This constitutes a measure of how readily accessible the PCG is to the subject. It is believed that a readily accessible PCG gives a greater degree of physical, psychological, and social support to the subject than is the case if the PCG is not readily accessible. Finally, two measures of solvency or financial support were identified in the data set. One was the annual family income for the subject's residential family. The other was an index of economic dependence, which is a composite of three dichotomous items, employment or unemployment, home ownership or no home ownership, and receipt of public aid or no receipt of public aid. The poles of this latter measure must be reversed to use it as a measure of solvency, with low financial dependence indicating high financial support and high financial dependence indicating low financial support. Further details regarding these measures will be found in Chapter 2. Chronological Relationships of Hypotheses--Paradjgms For each of the three theories of successful aging four dif- ferent forms of hypotheses will be developed. Each of the four forms of hypotheses will reflect a different time relationship of the criterion and independent variables. The two times referred to in the hypotheses are time #1, the first wave of data collection, and time #2, the second wave of data collection, approximately six months after time #1. Each form amounts to a pattern of relationship of the 23 independent variables and criterion which will be repeated, with slight alteration, to fit each theory. The forms or patterns will be referred to as “paradigms." In the first paradigm, which we will label "cross-sectional," both the independent variables and the criterion are measured at time #1. This amounts to a fairly straightforward procedure, and is essentially the basic hypothesis and least complex of the four para- digms. In the second paradigm, which we will label ”delayed cross- sectional," both the independent variables and the criterion are measured at time #2. Although this may appear to be a simple repli- cation of the first paradigm on the same sample at a later time, it really amounts to the same basic hypothesis applied to a decreased sample and altered set of data. In the first place, we expect the size of the study sample left alive at time #2 to be noticeably smaller. The selective loss of persons due to death is also expected to change the group values of all the independent variables in a biased manner. While the sampling and selection procedures will be detailed in Chapter II, it is important to know that the sample at time #1 will have been selected for the study through an acute care hospitali- zation, or an outpatient visit. Because of this, it will represent a type of extreme group. We expect the survivors of such an extreme group to change or regress toward the means of the independent and criterion variables by time #2. While the effect of regression to the mean should depress the ability to support the hypothesis, the 24 selective loss due to death may enhance the ability to support the hypothesis. Since it is not entirely clear what effect the two processes will have in combination, the test of the "delayed cross- sectional" paradigm of the basic hypothesis is important. The third paradigm will be labeled "chronological lag." In this pattern, the independent variables are from time #1, and the dependent or criterion variable is from time #2. Many balance theor- ists have remarked upon the gradualness of cognitive reorganization or restructuring. Recognition that such a gradual change may be in effect under the circumstances which we are concerned with has led us to test a hypothesis that the state of independent variables at an earlier time may relate to the state of a dependent variable at a later time. This paradigm can make use of only the sample which sur- vives at the later time. The fourth paradigm also must be tested on the sample sur- viving at the later time. It will be labeled "chronological antici- pation." In this pattern, the independent variables are from time #2, and the dependent or criterion variable is from time #1. Normal humans have the ability to bind time, predict future states of independent variables from their present store of knowledge, and make preparation for the future based on those predictions. Cog- nitive activity preparing for the future may cause real changes in cognitive structure, if an individual strives to maintain a balanced cognitive state. Through such changes in cognitive structure, related to preparation for predicted changes, it can happen that a 25 dependent variable will assume a changed state prior to the accom- plishment of the predicted change in independent variables. Such changes in person's states, which cannot be attributed to present circumstance, but can be attributed to reliably predicted future circumstance, are frequently labeled as ”anticipation." The runner who "jumps the gun" and the "sour grapes" behavior of the person who predicts future self-failure are both types of anticipation. Through testing this fourth paradigm we will test the criterion at an earlier time, in association with the actual occurrence of a change in the independent variables at a later time. We did not have measures available of the subject's predic- tions of future levels of the independent variables. If such meas- ures had been available, the anticipatory effect could be tested with a paradigm which used the criterion from time #1 and the predicted level of the independent variables, which had been measured at time #1. In that situation, the sample for testing the hypothesis would be larger, consisting of all persons in the sample at time #1. The fourth paradigm, which we have used to test for anticipa- tion effect, calls for a test which is imperfect, since it does not include a test of the anticipation experienced by nonsurvivors, and infers the prediction, at an earlier time, of a change in independent variables from evidence that a change has occurred in the independent variables at a later time. The data representing independent variables in this fourth paradigm will be inaccurate if the change observed has not been predicted or ifaapredicted change does not occur. 26 The four paradigms will be discussed further in the next chap- ter. All paradigms and theoretical alterations for the three theor- ies of successful aging will be given as 12 separate hypotheses at the end of this chapter. The number (1-4) in the notation of each of the hypotheses indicates the paradigm number. The Research Hypotheses The Consistency Hypotheses The major purpose of the present paper is to test Gubrium's underlying assumption by testing the four postulates which were formulated from the Gubrium position against hypotheses derived from disengagement and activity theory. Inspection of the language of the four consistency postulates reveals a pattern which is reducible to a single summarizing statement of Gubrium's consistency position: Contentment, among the elderly, is dependent upon an interaction between activity resources and age environ- ment. The form of the interaction is such that con- tentment will be high when either of the following situations exists: A. Activity resources of the individual are high, and age environment is nonconcen- trated and demanding. B. Activity resources of the individual are low and age environment is concentrated and non- demanding. The first two research hypotheses are related to direct test of Gubrium's assumption and its attendant postulates, at two differ- ent times. The two times referred to in the hypotheses are time #1, 27 first wave of data collection, and time #2, second wave of data collection, approximately six months after time #1. For each of the four consistency hypotheses (denoted by the "C" in the notation) the form of the interaction is defined as in the summarizing statement, above. HolC: Contentment, among the elderly, measured at time #1 is dependent upon an interaction between activity resources measured at time #1 and age environment. M929: Contentment, among the elderly, measured at time #2 is dependent upon an interaction between activity resources measured at time #2 and age environment. In addition, it may be that the effect of an interaction between the two variables will not be immediately measurable. Although Gubrium makes no mention of such a possibility, Loeb, Pincus, and Mueller (1966) mention as one of the three characteris- tics of a "congruent reorganization" that: "any contraction occurs gradually, i.e., there is no sudden collapse of any dimension." To test for such a time lag in the hypothesized interaction effect, a third hypothesis is needed. 5939: Contentment, among the elderly, measured at time #2 is dependent upon an interaction between activity resources measured at time #1 and age environment. Finally, it may be that the effect of an interaction between the two factors will be noticable prior to the actual occurrence of the interaction, as though anticipation were at work. That leads to the fourth in this set of consistency hypotheses. uggg; Contentment, among the elderly, measured at time #1 is dependent upon an interaction between activity resources measured at time #2 and age environment. 28 The Activity Hypotheses In opposition to the four interaction hypotheses based upon Gubrium's consistency approach to successful aging, are several main effect hypotheses which are related to disengagement and activity theories of successful aging. These hypotheses are designed specific to certain factors and directions of the main effects. The following is a summary statement of activity theory assumptions in regard to the types of variables available in this study. Contentment, among the elderly, is dependent upon main effects for A. Independence in activities of daily living (independence associated with higher content- ment, dependence associated with lower con- tentment). Days of bed disability (no days associated with higher contentment, 1-14 days associated with lower contentment). Other main effects which may be indirectly related to activity theory are: Non-concentrated, demanding (vs. concentrated, nondemanding) age environment, which would be expected to demand more kinds and amounts of activity from the older person; Spouse survival (vs. spouse non-survival), which is expected to demand more task-oriented activity and allow more leisure activity; 29 3. Primary care giver residing in a different building (vs. the same building), which is expected to place some burden on the older person to manage self-care matters personally; 4. High financial ability (vs. low financial ability), which is expected to make available the opportunity for more leisure and elective activities, such as travel, education, shopping, etc. The first two research hypotheses in regard to activity theory are related to HolC and H02C. All activity theory hypotheses will be denoted by the appearance of "A" in the notation. The direction of all hypothesized main effects for activity theory will be as defined earlier, in the summary statement. HolA: Contentment, among the elderly, measured at time #1 is dependent upon main effects for: a. Independence in activities of daily living measured at time #1. b. Days of bed disability measured at time #1. HoZA: Contentment, among the elderly, measured at time #2 is dependent upon main effects for: a. Independence in activities of daily living measured at time #2. b. Days of bed disability measured at time #2. The next two hypotheses for activity theory are related to Ho3C and Ho4C in a parallel fashion. H03A: Contentment, among the elderly, measured at time #2 is dependent upon main effects for: a. Independence in activities of daily living measured at time #1. b. Days of bed disability measured at time #1. 3O Ho4A: Contentment, among the elderly, measured at time #1 is dependent upon main effects for: a. Independence in activities of daily living measured at time #2. b. Days of bed disability measured at time #2. The Disengagement Hypotheses The following is a summary statement of disengagement theory assumptions which are available for testing in this study: Contentment, among the elderly, is dependent upon a main effect for age environment. (Concentrated, nondemanding associated with higher contentment, non-concentrated, demanding associated with lower contentment.) Other main effects which may be indirectly related to dis- engagement theory are: 1. Dependence (vs. independence) in activities of daily living, which is expected to allow the person to disengage from responsibility for performance of AOL; 2. Some days (vs. no days) of bed disability, which is expected to allow the person to disengage from inter- action outside a single room; 3. Spouse non-survival (vs. spouse survival), which is expected to demand that there be disengagement from the spouse, and allow disengagement from married-set friends; 4. Primary care giver residing at the same building (vs. a different building), which is expected to allow the disengagement from persons and places outside of the residential building; 5. Low financial ability (vs. high financial ability), which is expected to demand disengagement from an expensive life style, and activities of a middle aged person. 31 The disengagement theory hypotheses for this study will all be notated with a "D" and the four disengagement hypotheses are related to the four consistency hypotheses, HolC to H040, stated earlier. The direction of the main effect for each hypothesis is as defined earlier, in the summary statement. HolD: Contentment, among the elderly, measured at time #1 1s dependent upon a main effect for age environment measured at time #1, when crossed with factors meas- ured at time #1. H020: Contentment, among theelderly, measured at time #2 1s dependent upon a main effect for age environment measured at time #1, when crossed with factors meas- ured at time #2. H030: Contentment, among the elderly, measured at time #2 is dependent upon a main effect for age environment measured at time #1, when crossed with factors meas- ured at time #1. H040: Contentment, among the elderly, measured at time #1 is dependent upon a main effect for age environment measured at time #1, when crossed with factors meas- ured at time #2. Although the main effects referred to in the first and fourth disengagement hypotheses and in the second and third disengagement hypotheses are essentially the same main effects, we have stated four hypotheses, rather than two, to continue the parallelism between dis- engagement hypotheses and the four hypotheses of the other two theories. However, we may expect slightly different results for the main effects in the first and fourth hypotheses, due to the difference in the sample sizes on which the analyses will be performed. Such a difference in results is not expected between the main effect results for the second and third hypotheses, since they are tested on rela- tively equal samples. 32 In Chapter II the differences in these samples will be described. That chapter also contains a description of the operation- alization of the variables, diagrams of paradigms for each hypotheses, and the details of the analyses used to test the hypotheses. CHAPTER II METHODOLOGY AND PROCEDURES Context of This Study This study has been conducted in conjunction with research evaluating a demonstration of home health care service. The demon- stration project was implemented at two urban and three rural sites in the state of Michigan between August, 1973, and June, 1976, by staff of the Office of Health Services, Education and Research, College of Human Medicine, Michigan State University. The home care service being demonstrated and field-tested consisted of the care of chronically impaired individuals in their residences, rather than in an institutional setting. This service was rendered by teams consist- ing of: (1) a part-time physician, (2) a half-time supervising nurse or social worker, and (3) two fall-time health assistants. The services provided to patients included health assessment, health edu- cation, therapy, social, psychological, and financial counseling of patient and family, referral to other agencies, transportation, limited homemaking, and limited nursing services. In the demonstration project, data was gathered from a group of subjects who served to pretest the evaluation procedures, and who were not included in the evaluation of the demonstrated service. This pretest group of subjects was chosen by the same criteria and from the same screening sources as the subjects who participated in 33 34 the evaluation. The criteria for choosing the subjects were that they be: (1) 45 years of age or older, (2) chronically impaired, (3) in need of the types of services which could be provided by the home care service, and (4) not in need of 24-hour supervision or skilled nursing care. Subjects were screened for these entry criteria either while they were inpatients about to be discharged from a hospital, or as outpatients in ambulatory health care clinics. Those subjects who were judged eligible to be included in the evaluation study were randomly assigned into two approximately equal-sized groups. The experimental group was offered the demonstration home care service. The control group was not offered the demonstration service, but was not restrained from access to any of the health care services normally available to such persons in each community. Informed consent was obtained from each subject before any interviews were conducted. The interviewers were trained research workers, not involved in any way in the provision of the demonstrated home health care service. Subjects in the experimental group and control group were interviewed three times. Approximately six months separated succeeding interviews. These interviews will be referred to as time #1, time #2, and time #3. Subjects in the pretest group were interviewed at time #1 and time #2 only. Further interviews were not necessary in order to pretest the procedures for the follow- up interviews. 35 Identification of the Sample Since the sample for the present study was chosen from among the subjects in the home health care demonstration described above, it consists basically of persons 45 or older, who have been identi- fied as having chronic physical impairment of various degrees of severity. For purposes of this study two added restrictions were employed, that each subject have at least one cohabitant and that the subjects be living in a residential setting at the time of all the interviews utilized in the study. In an attempt to obtain nearly equal numbers of subjects from rural and urban settings, the sample was restricted to one of the urban sites and two of the rural sites. At those three sites, the sample was further restricted to include only persons in the pretest group and the control group, none of whom were offered any of the experimental home health care service. The study sample identified in this manner included 280 sub- jects. At time #1 this sample was decreased, due to 20 interviews with missing data, to a size of 260. At time #2 the sample was decreased still more, due to experimental and/or actual mortality of 79 subjects during the six-month interval and 12 interviews with missing data, to a size of 189. No data was reported for time #3, since the pretest group included in the sample for this study was not interviewed at time #3. The sample identified in this manner was not intended to be representative of the entire population of elderly persons. There was no intention to estimate population parameters from the results of this study. Neither should this sample be mistaken for a sample 36 representative of all elderly persons who are "sick." This sample was deficient as a representative sample of the sick and elderly for the following reasons: (1) it included pre-elderly persons (45 to less than 65), (2) it did not include the institutionalized elderly, (3) it did not include the elderly in need of skilled nursing care and/or 24-hour-a-day supervision, (4) it excluded the sick elderly newly hospitalized for an episode of acute illness, (5) it excluded the sick elderly not availing themselves of health care facilities, and (6) it included those persons recently discharged from a hospital who may have been in a state of relatively good health. Although the basic characteristics of this sample were tailored for use in the home health care demonstration project, the sample also had characteristics which satisfied several important needs in testing the hypotheses of this study. Limitation of the sample to persons who were judged to be chronically impaired was expected to have the effect of providing more nearly equal groups of subjects with high and low activity resources than if the sample had been chosen as representative of the general population. In the general population, we would expect to obtain a much larger frequency of subjects in the high group than in the low group. This is an important consideration for data analysis used to test the study's hypotheses. A second consideration in choosing a sample limited in this manner was that the resource abilities of such a sample were char- acterized as dynamic, subject to fairly rapid shifts through a relatively wide range of levels. This allowed one to expect changes 37 in activity resources large enough to be observable, and observable during relatively short time periods. The major focus of the four hypotheses was the effect of certain variables as they operate for the elderly. However, persons in their late middle years (over 45 and under 65 years of age) were included in the sample so that the effects could be compared at ages defined as elderly and middle aged. Characteristics of the Sample Table 2 presents the data obtained for type of site, age, sex, ethic group, marital status, highest occupation and educational attainment.1 The data are tabled by frequency and percentage for the 280 subjects in the total sample at time #1 and for the 201 sub- jects in the total sample at time #2. All the subjects described for time #2 were included in the description for time #1. The third column of the table accounts for the group of 79 subjects lost between time #1 and time #2. For a few of the characteristics, there are small amounts of missing data. Data is complete for all samples for the characteristics of type of site, age, and sex. The characteris- tic of highest occupation has the most missing data, with eight missing cases at time #1 and five missing cases at time #2. The age reported in this table is the age of the participant at the time of screening for the study. The differential percent- ages by sex reported for this sample are not extraordinary, since 1See Appendix B for items and code book. 38 Table 2: Characteristics of Subjects in Total Sample at Time #1 and #2. Sample Lost (at Time .2) Due to Experi- Time 41 Time #2 mental and/or Sample Sample Actual Mortality (N 8 280) (N = 201) (N = 79) Characteristics of Subjects N z N t N : Type of Urban 136 48.6 96 47.3 40 50.6 Site Rural 144 51.4 105 52.2 39 49.4 it 66.5 66.2 Ma Age 45 to <65 125 44.6 92 45.8 33 41.8 265 155 55.4 109 54.2 46 58.2 Sex Male 119 42.5 86 42.8 33 41.8 Female 161 57.5 115 57.2 46 58.2 Ethnic White 240 86.0 172 86.0 68 86.0 Group Black 39 14.0 28 14.0 11 14.0 Marital Married 186 66.4 138 68.7 48 60.8 Status Divorced/Separated 22 7.9 15 7.5 7 8.9 Widowed 65 23.2 44 21.9 21 26.6 Never Married 7 2.5 4 2.0 3 3.8 Highest Higher Executives and Occupation Higher Professionals 2 .7 1 .5 l 1.3 Business Managers and Lesser Professionals 9 3.2 8 4.1 l 1.3 Administrative Personnel 36 13.2 24 12.2 12 15.8 Clerical, Sales and Technical 49 18.0 36 18.4 13 17.1 Skilled Manual Labor 42 15.4 31 15.8 11 14.5 Semi-Skilled Manual Labor 66 24.3 47 24.0 19 25.0 Unskilled Manual Labor 53 19.5 36 18.4 17 22.4 Housewife 15 5.5 13 6.6 2 2.6 Educational Graduate Degree 2 .7 2 2.6 Attainment Bachelor's Degree 6 2.2 4 2.0 2 2.6 Some College 19 6.9 16 8.1 3 3.8 High School Grad. 31 11.2 24 12.1 7 9.0 10th-11th Grade 45 16.3 29 14.6 16 20.5 7th-9th Grade _ 118 42.8 85 42.9 33 42.3 Less than 7th Grade 55 19.9 40 20.2 15 19.2 Note. N and t of all non-missing data has been reported for each characteristic. Small amounts of missing data reduce the sample size for some variables. .NA = datum not available. 39 females survive longer than males1 and are known to utilize health care facilities in larger numbers than males.2 Table 3 presents the data for the characteristics of the primary care giver (PCG) as to: (1) sex of the subject and sex 1"The mortality sex ration (age-specific death rate for the specified year for the male population divided by the corresponding age-specific death rate for the female population), for the (United States) population as a whole, did not change appreciably over the study years: for 1950, 1960, and 1969 the death rates for the male population were, respectively, 1.34, 1.36, and 1.35 times the corre- sponding rates for the female population." Furthermore, the mortality sex ratio was greater than unity for each 5-year age-specific sub- population during these three report years. In summary, more males die than females, at all ages, in recent U. S. population history; and over all ages, death rates are at a ratio of approximately four (for males) to three (for females). From: U. S. Department of Health, Education, and Welfare. Mortality Trends: ,Age, Color, and Sex, United States-~1950-69. Vital and Health Statistics Data from the National Vital Statistics System, Series 20--Number 15, DHEW Publication No. (HRA) 74-1852. Public Health Service, Health Resources Administration, Rockville, Maryland, 1973, pp. 20 and 30. 2The numbers of discharges from short-stay hospitals for the United States population over 45, in 1972, was: 6,480,000 for males and 7,619,000 for females. From: U. S. Department of Health, Education, and Welfare. Utilization of Short-Stangospitals: Summary of Nonmedical Sta- tistics: United States--1972, Vital and Health Statistics Data from the National HeaTth Survey, Series l3--Number l9, DHEW Publi- cation No. (HRA) 75-1770. Public Health Service, Health Resources Administration, Rockville, Maryland, 1975, p. 3. The numbersrfiiphysician visits for the United States popu- lation over 45, in 1969, was: 123,489,000 for males and 181,748,000 for females. For persons 45 to 64 years of age, males made an aver- age of 4.1 physician visits per year, and females made an average of 5.2 such visits per year. For persons 65 years of age and older, males made an average of 5.5 physician visits per year, and females made an average of 6.6 such visits per year. From: U. S. Department of Health, Education, and Welfare. Age Patterns in Medical Care, Illness, and Disability: United States, 1968—1969, Vital and Health Statistica Data from the National Health Survey, Series lO--Number 70, DHEW Publication No. (HSM) 72-1026. Public Health Service, Health Services and Mental Health Administration, Rockville, Maryland, 1972, p. 30. 40 Table 3: Characteristics of Primary Care Giver and Residence for Total Sample at Time 01 and 02 Sample Lost (at Time 42) Due to Experi- glgg1zi 52$§152 mental and/or Actual Mortality (H I 280) (N a 201) (N = 79) Characteristics of Primary Care Giver N z N S N 1 Sex Subjectsfemale S PCG=male 83 29.6 63 31.3 20 25.3 Relation Subject=female a PCG=female 78 27.9 52 25.9 26 32.9 Subject=ma1e a PCG=fema1e 115 41.1 82 40.8 33 41.8 Subject=male S PCG=male 4 1.4 4 2.0 -- -- Relation- Spouse 152 54.3 113 56.2 39 49.4 ship Child 68 24.3 50 24.9 18 22.8 Other Relative 37 13.3 23 11.5 14 17.7 Other Nonrelative 23 8.2 15 7.5 8 10.1 Characteristics of Subject's Residence Length Less than 1 Year 44 15.8 26 13.0 18 23.1 of Time at 1 Year ‘- 5 Years 62 22.3 45 22.5 17 21.8 Residence 5 Years- 10 Years 32 11.5 29 14.5 3 3.8 10 Years or more 140 50.4 100 50.0 40 51.3 Number of One Cohabitant 155 55.8 117 58.5 38 48.7 Cohabitants Two Cohabitants 53 19.1 39 19.5 14 17.9 Three Cohabitants 19 6.8 9 4.5 10 12.8 Four Cohabitants 21 7.6 16 8.0 5 6.4 Five or more Cohabitants 30 10.7 19 9.5 11 14.1 Owner of Landlord 47 17.3 30 15.3 17 22.7 Residence Other Resident 53 19.6 33 16.8 20 26.7 Self or Spouse 171 63.1 133 67.9 38 50.7 Head of Self 130 48.5 96 50.3 34 44.2 Household Spouse 83 31.0 62 32.5 21 27.3 Other Person 55 20.5 33 17.3 22 28.6 Note. N and X of all non-missing data has been reported for each characteristic. Small amounts of missing data reduce the sanple size for some variables. 41 of the PCG in combination, and (2) family relationship of the Pop.1 Table 3 also presents residential characteristics for subjects in the total sample at the two different times and for those subjects lost from the sample at time #2. These characteristics are: (1) length of time the subject has lived at the residence, (2) number of co-habitants of the subject at the residence, (3) type of person who owns the residence, and (4) type of person who is head of the house- hold.2 The PCG mentioned in Table 3 was defined as the person (either relative or friend) who gives the subject the most direct help with both personal care and household tasks. It is not sur- prising that 79% of all PCG's are female. The helping role has been a traditionally female role. This tradition is reflected in this age-defined sample. Furthermore, more females than males are left available to act as PCG. It is noteworthy that only 17% lived in rental housing. In Table 3 data is complete for all samples for the two characteristics of the PCG. The characteristic, "head of household" has the most missing data, with 12 missing cases at time #1 and 10 missing cases at time #2. In reading the data in Table 2 and 3, it should be remembered that all data reported remains constant from time #1 to time #2. Marital status, for example, may actually change during the six- month period, but the data reported will not change. The differences 1 2 See Appendix 8 for items and codebook. See Appendix 8 for items and codebook. 42 observed at the two times between frequencies for.c1asses of each characteristic are due only to the 79 cases lost to experimental and/or actual mortality. Data from the total sample and from the two age subsamples will be analyzed to test the hypotheses. In Tables 2 and 3 data describing the total sample was reported. Tables 4 and 5 report data for the sample of persons 65 years of age and older (the older sample). Tables 6 and 7 repeat the descriptive report for the sample of persons 45 to less than 65 years of age (the younger sample). A larger percentage of the older sample was white, located at a rural site, from the top four occupation classes, and widowed than the younger sample (see Table 4 and 6). Further comparison of the data in these two tables indicates that for the older sample, compared to the younger sample, a smaller proportion was divorced or separated and was at least a high school graduate. None of these differences is surprising, given the effects of social history as well as the effect of the differential placement in the life cycle of these two cohort groups. To a great extent the sample lost between time #1 and time #2 was lost due to death, rather than to refusal or inability to locate the participant at the second interview. For the older sample, this lost group was more frequently rural, white, and widowed. For the younger sample, the lost group was more frequently urban, black, and divorced or separated. The overall effect of this differential loss to the two age groups,fbr type of site, ethnic group and marital 43 Table 4: Characteristics of Subjects in Older Sample at Time #1 and +2. Sample Lost (at line ''2) Time p] Time ,2 Due to Experi- mental and/or samp'e samp'e Actual Mortality (N = 155) (N = 109) . (N = 46) Characteristics of Subjects N G N i N Type of Urban 69 44.5 51 46.8 18 39.1 Site Rural 86 55.4 58 53.2 28 60.9 X 75.4 75.2 Age 45 to <65 -- -- -- -- -- ~- 265 155 100.4 109 100.0 46 100.0 Sex Male 65 41.9 46 42.2 19 41.3 Female 90 58.1 63 57.8 27 58.7 Ethnic White 138 89.1 95 87.2 43 93.5 Group Black 17 1049 14 12.8 3 6.5 Marital Married 94 60.6 68 62.4 26 56.5 Status Divorced/Separated 5 3.2 5 4.6 -- —- Widowed 53 34.2 35 32.1 18 39.1 Never Married 3 1.9 l .9 2 4.3 Highest Higher Executives and Occupation Higher Professionals 2 3 l 1.0 1 2.3 Business Managers and Lesser Professionals 8 5.4 7 6.7 1 2.3 Administrative Personnel 22 14.8 15 14.3 7 15.9 Clerical, Sales and Technical 30 20 l 21 20.0 9 20.5 Skilled Manual Labor 23 15 4 17 16.2 6 13.6 Semiskilled Manual Labor 29 19.5 22 21.0 7 15.9 Unskilled Manual Labor 24 16.1 13 12.4 11 25.0 Housewife 11 7.4 9 8.6 2 4.5 Educational Graduate Degree 2 1.3 -- -- 2 4.3 Attainment Bachelor's Degree 4 2.6 2 1.9 2 4.3 Some College 9 5.8 7 6.5 2 4.3 High School Grad. 10 6.5 8 7.4 2 4.3 lOth-llth Grade 22 14.3 15 13.9 7 15.2 7th-9th Grade 74 48.1 52 48.1 22 47.8 Less than 7th Grade 33 21.4 24 22.2 9 19.6 Note. N and 1 of all non-missing data has been reported for each characteristic. Small amounts of missing data reduce the sample size for some variables. 44 Table 5: Characteristics of Primary Care Giver and Residence for Older Sample at Time #1 and 02. Sample Lost (at Time =2) Due to Experi- Time #1 Time #2 mental and/or Sample Sample Actual Mortality (N = 155) (N = 109) (N = 46) Characteristics of Primary Care Giver N t N 1 N 3 Sex Subject=female 5 PCG=male 39 25.2 30 27.5 9 19.6 Relation Subject=female 6 PCG=female 51 32.9 33 30.3 18 39.1 Subject=ma1e 6 PCG=female 62 40.0 43 39.4 19 41.3 Subject=male S PCG=male 3 1.9 3 2.8 -- -- Relation- Spouse 75 48.4 56 51.4 19 41.3 ship Child 42 27.1 31 28.4 11 23.9 Other Relative 26 16.7 13 11.9 13 28.3 Other Nonrelative 12 7.8 9 8.2 3 6.5 Characteristics of Subject's Residence Length of Less than 1 Year 23 14.9 12 11.1 11 23.9 Time at 1 Year - 5 Years 33 21.4 26 24.1 7 15.2 Residence 5 Years -10 Years 15 9.7 15 13.9 -- -- 10 Years or more 83 53.9 55 50.9 28 60.9 Number of One Cohabitant 101 65.6 78 72.2 23 50.0 Cohabi- Two Cohabitants 24 15.6 15 13.9 9 19.6 tants Three Cohabitants 7 4.5 4 3.7 3 6.5 Four Cohabitants 8 5.2 4 3.7 4 8.7 Five or more Cohabitants 14 9.0 7 6.5 7 15.2 Owner of Landlord 17 11.3 13 12.3 4 9.1 Residence Other Resident 40 26.7 22 20.8 18 40.9 Self or Spouse 93 62.0 71 67.0 22 50.0 Head of Self 70 46.7 53 50.5 17 37.8 Household Spouse 41 27.3 31 29.5 10 22.2 Other Person 39 26.0 21 20.0 18 40.0 Note. N and Z of all non-missing data has been reported for each characteristic. Small amounts of missing data reduce the sample size for sore variables. 45 ' Table 6: Characteristics of Subjects in Younger Sanple at Time 01 and #2. Sample Lost (at Time *2) . . . Due to Experi- Time #1 Tine =2 Sanple Sample nental a"°/°F Actual Mortality (N = 125) (N = 92) (N = 33) Characteristics of Subjects N N l N % Type of Urban 67 53.6 45 48.9 22 66.7 Site Rural 58 46.4 47 51.1 11 33.3 'i 55.5 55.5 Age 45 to <65 125 100.0 92 100.0 33 100.0 265 -- —- -- -- -- -- Sex Male 54 43.2 40 43.5 14 42.4 Female 71 56.8 52 56.5 19 57.6 Ethnic White 102 82.2 77 84.6 25 75.8 Group Black 22 17.7 14 15.4 8 24.2 Marital Married 92 73.6 70 76.1 22 66.7 Status Divorced/Separated 17 13.6 10 10.9 7 21.2 Hidowed 12 9.6 9 9.8 3 9.1 Never Married 4 3.2 3 3.3 1 3.0 Highest Higher Executives and Occupation Higher Professionals -- -- -- -- -- -- Business Managers and Lesser Professionals 1 .8 1 1.1 -— —- Administrative Personnel 14 11.4 9 9.9 5 15.6 Clerical, Sales and Technical 19 15.4 15 16.5 4 12.5 Skilled Manual Labor 19 15.4 14 15.4 5 15.6 Semiskilled Manual Labor 37 30.1 25 27.5 12 37.5 Unskilled Manual Labor 29 23.6 23 25.3 6 18.8 Housewife 4 3.3 4 4.3 -- -- Educational Graduate Degree -- -- -- —- -- -- Attainment Bachelor‘s Degree 2 1.6 2 2.2 -— -- Sone College 10 8.2 9 10.0 1 3.1 Hign School Grad. 21 17.2 16 17.8 5 15.6 lOth-llth Grade 23 18.9 14 15.6 9 28.1 7th-9th Grade 44 36.1 33 36.7 11 34.4 Less than 7th Grade 22 18.0 16 17.8 6 18.8 Note. N and Z of all non-missing data has been reported for each characteristic. Small amounts of missing data reduce the sample size for some variables. 46 Table 7: Characteristics of Primary Care Giver and Residence for Younger Sample at Time #1 and 92. Sample Lost -(at Time *2) Due to Experi- Tine a] Time 82 mental and/or Sample Sample Actual Mortality (N = 125) (N = 92) (N = 33) Characteristics of Primary Care Giver N z N % N % Sex Subject=fema1e & PCG=ma1e 44 35.2 33 35.9 11 33.3 Relation Subject=fenale & PCG=female 27 21.6 19 20.7 8 24.2 Subject=ma1e 8 PCG=female 53 42.4 39 42.4 14 42.4 Subject=male & PCG=male l .8 l 1.1 -- -- Relation- Spouse 77 61.6 57 62.0 20 60.6 ship Child 26 20.8 19 20.7 7 21.2 Other Relative 11 8.8 10 10.9 1 3.0 Other Nonrelative 11 8.8 6 6.6 5 15.2 Characteristics of Subject's Residence Length of Less than 1 Year 21 16.8 14 15.3 7 21.9 Tine at 1 Year '- 5 Years 29 23.4 19 20.7 10 31.2 Residence 5 Years- 10 Years 17 13.7 14 15.2 3 9.4 10 Years or more 57 46.0 45 48.9 12 37.5 Number of One Cohabitant 54 43.5 39 42.4 15 46.9 Cohabi- Two Cohabitants 29 23.4 24 26.1 5 15.6 tants Three Cohabitants 12 9.7 5 5.4 7 21.9 Four Cohabitants 13 10.5 12 13.0 1 3.1 Five or more Cohabitants 16 12.8 12 13.0 4 12.5 Owner of Landlord 30 24.8 17 18.9 13 41.9 Residence Other Resident 13 10.7 11 12.2 2 6.5 Self or Spouse 78 64.5 62 68.9 16 51.6 Head of Self 60 50.8 43 50.0 17 53.1 Household Spouse 42 35.6 31 36.0 11 34.4 Other Person 16 13.6 12 14.0 4 12.5 Note. N and 3 of all non-missing data has been reported for each characteristic. Small amounts of missing data reduce the sample size for some variables. 47 status, was to make the two age group samples more comparable at time #2 than at time #1. A similar assessment may be made of the two age samples by comparing Table 5 with Table 7. In the older sample, compared to the younger sample, females more frequently had a female PCG, while all participants more frequently had a child or other relative as a PCG, lived in their home for 10 years or more, had only one cohabi- tant, lived where the owner is an “other resident," and where the head-of-the-household was an "other person." At the same time, comparison of the two age samples showed that the older sample was less frequently cared for by a PCG who was a spouse, and was less frequently a renter. The heavier loss of persons from the sample who had female PCGs, noted for the total sample, may be almost entirely accounted fOr within the older sample. Persons whose PCG was a relative other than a spouse or child were more frequently lost from the older sample, and were less frequently lost from the younger sample. This differential loss from the two age samples affected the sample pr0portions at time #2, making the two samples more comparable for this characteristic. A similar effect of the loss of subjects from the two samples may be noted on other variables reported in Tables 5 and 7. The overall result is to move the two samples to greater comparability at time #2 on the characteristic of residential tenure of 10 years or longer, and the type of owner of the residence. 48 In general, the bias of sample loss produces greater simi- larity in the two age samples at time #2 than at time #1. Only three of the 12 hypotheses are tested on the sample at time #1. Because of the change in the nature of the age samples from time #1 to time #2, it is important to note this change in the nature of the sample when discussing the findings. Qperationalizing the Variables1 Criterion as Contentment In this study the criterion variable of satisfaction/morale was operationalized using the five-item contentment scale developed by Blenkner, Bloom, and Weber (1964). This scale has a concurrent validity of .69 using independent data and a test-retest reliability of .65. Several of its items appear to have been borrowed from the Lawton 24-item morale scale. The contentment measure includes items on: (1) self-perceived health, (2) general satisfaction, (3) satis- faction with arrangements for housework and maintenance, (4) general worry, and (5) "things get worse as I get older." Individual items are dichotomous. This measure has been used as a six-place scale (R=0-5) with a high score of 5. The score is obtained by aggregating the number of items which were given "contented" responses by each subject. 1For the form of the items used in interviewing the partici- pants, the codebook, and method for constructing each of the follow- ing variables, see Appendix B. 49 Social Context as .Age Environment The social context factor was operationalized as cohabitant age environment. The two levels of cohabitant age environment are: (l) concentrated, nondemanding cohabitant age environment, and (2) nonconcentrated, demanding cohabitant age environment. A non- concentrated age environment was conceived of as one in which the subject lives with persons of a variety of ages. In such a setting, it was expected that the subject would be subjected to many demands regarding the subject's roles relative to each generation, as well as to heavy demands of the wider range of activities in a household with a wider age range. These demands take the form of implicit or explicit requests for accommodation of the subject's activity to the activities of the cohabitants. 0n the other hand, the concentrated age environment was conceived of as one in which the subject lives with persons of nearly the same age or older. In the concentrated age environment it was expected that the demands described above would be considerably lessened. Lessened demand on the subject was labeled "nondemanding." A concentrated, nondemanding age environment was defined as one in which the subject lives with others, all of whom are either 45 or older, or younger than 45 but less than 10 years younger than the subject. A nonconcentrated, demanding age environment was defined as one in which the subject lives with others, at least one of whom is younger than 45 and 10 years or more younger than the subject. 50 Activity Resource as Physical Health The three variables which comprised the activity resources factor are physical health, social support, and financial ability. The first of these variables, physical health, was operationalized using two separate measures. The first measure was the number of days spent in bed for all or most of the day, during the past 14 days. This “h; a measure of morbidity adopted by the National Health Survey. The NHS definition of "Days of Bed Disability" specifies that days spent in hospitals, nursing homes, or other health care institutions be considered as bed disability days, whether the per- son is in bed most of the day or not. The range of this measure was originally 0-14 days with a score of 0 days indicating best health. For purposes of analysis in this study, it was dichotomized into: (1) zero days of bed disability (favorable), and (2) any number of days (1-14) of bed disability (unfavorable). The second measure of physical health which has been used is Katz's independence of activities of daily living (ADL) (Katz, Ford, et al., 1963). This scale is formed by a count of the number of activities of daily living which the person performs independently.‘ The six activities are: (l) bathing, (2) dressing, (3) toileting, (4) transferring (from bed to chair, etc.), (5) feeding self, and (6) control of elimination. The range as originally measured was 0-6 with a score of 6 indicating best health. For purposes of analysis in this study, ADL was dichotomized into: (1) independent in all six actities (favorable), and (2) dependent in any number (1-6) of the six activities (unfavorable). 51 Activity Resource as Social Support Social support, the second variable in the activity resources factor, was operationalized using two measures. The first measure was dichotomous for survival of spouse vs. nonsurvival of spouse. Persons who were never married were treated as having a nonsurviving spouse. The second measure was dichotomous for persons having pri- mary care givers (PCGs) living in the same building vs. living in a different building. A PCG was defined as the person (either relative or friend) giving the subject the most direct help with both personal care and household tasks: things like bathing, dressing, laundry, cooking, shopping, cleaning, etc. PCG proximity was a two-place scale; either a PCG was identified who did not live with the subject, or a PCG was identified who lived in the same building with the sub- ject. Activity Resource as Financial Ability The third variable, financial ability, was measured in two ways, first as a six-place classification of annual family income, with a high score of 6; and second in terms of the index of economic dependence (IED), developed by Katz, Ford, et al. (1972). The IE0 is a four-place scale based upon: (1) employment, (2) home ownership, and (3) independence of assistance from private or public agencies. A score of 1 indicates greatest solvency or independence. Because higher levels of missing data were obtained on both these two measures than on any of the other measures, and because the 52 missing data for these two measures was frequently found to be non- overlapping, a summary index of financial ability was constructed from a weighted average of these two measures. The weights used in this data reduction were the correlations of each measure with an unweighted average of the two measures. The unweighted measure was obtained after substitution of the median value for missing data in either measure when valid data was present for the complement measure. The correlation of the two measures was computed to confirm the likelihood that they were measuring the same concept. The correlation obtained for the two measures was .40 at the initial interview (time #1), .34 at the six- month interview (time #2), and .37 at the initial interview (time #1) for subjects having a six-month interview (time #2). This summary measure of financial ability was the only measure used for testing hypotheses involving the financial ability aspect of the activities resources factor. For purposes of analysis, it was dichotomized at: (1) equal to or greater than its median, and (2) less than its median. Paradigms for the Theories and Hypotheses General Paradigms for the Three Theories The general paradigm for consistency theory used in testing the four consistency hypotheses is schematized in Figure 4. In this model the criterion variable of satisfaction/morale, Operationalized as contentment, is dependent upon an interaction between social con- text and individual context. The direction of the interaction is 53 SOCIAL CONTEXT INDIVIDUAL CONTEXT Local Environment Activity Resources Age Environment Physical Health CONCENTRATED, NONCONCENTRATED, Social Support NONDEMANDING DEMANDING F1030C161 Solvency Age Environment Age Environment HIGH SATSIFACTION/ SATISFACTION/ Activity MORALE 4: Resources Contentment Contentment LON SATISFACTION/ L SATISFACTION/ Activity MORALE MORALE Resources Contentment Contentment Figure 4. General paradigm for consistency theory. identified by the signs (> or <). The predictions for this para- digm were tested by examining the magnitude and direction of the appropriate interaction term for each of the four consistency hypothe- ses. The paradigm for activity theory is shown in Figure 5. The activity and consistency paradigms differ in two ways. For the activ- ity paradigm, the variable of individual context includes only the two measures which comprise the physical health portion of individual context in the consistency paradigm. Further, in the activity para- digm individual context is conceived of as "activity," rather than "activity resources." The direction of the hypothesized main effect on the variable of individual context is indicated by the signs (> or <). Figure 6 shows the general paradigm for disengagement theory to the extent we were able to operationalize it within this study. 54 INDIVIDUAL CONTEXT Individual Activity - Independence of AOL Days Bed Disability SOCIAL CONTEXT Local Environment Age Environment CONCENTRATED, NONDEMANDING Age Environment NONCONCENTRATED, DEMANDING Age Environment HIGH SATISFACTION/ SATISFACTION/ Activit MORALE MORALE y Contentment Contentment LON SATISFACTION/ SATISFACTION/ A tivit MORALE MORALE c y Contentment Contentment Figure 5. General paradigm for activity theory. INDIVIDUAL CONTEXT Activity Resources SOCIAL CONTEXT Local Environment Age Environment Physical Health CONCENTRATED, NONCONCENTRATED. Social Support NONDEMANDING DEMANDING Financial Solvency Age Environment Age Environment HIGH . SATISFACTION/ SATISFACTION/ Act1v1ty MORALE a MORALE Resources Contentment Contentment LON. . SATISFACTION/ ) SATISFACTION/ Act1v1ty MORALE MORALE Resources Contentment Contentment Figure 6. General paradigm for disengagement theory. 55 It should be recognized that neither activity nor disengagement the- ories were completely tested in this study. The loss of theoretic detail is;especially noticeable in disengagement theory. The hypothe- sized main effect for higher contentment for persons in a concen- trated, nondemanding (disengaged) age environment is in opposition to the interaction effect of social and individual context which marks the consistency theory approach. The hypothesized effect has been indicated in Figure 6 by the direction of the signs (> or <). Paradigms for the Four Specific Hypotheses Immediate Cross Sectional Paradigm and Hypotheses. This paradigm (Figure 7-9) utilized cross sectional data from 260 of the 280 initial interviews (time #1) for which complete data were avail- able. For the analyses of covariance, age of the subject at time of screening into the study was used as the only covariate. CONCENTRATED, NONCONCENTRATED. NONDEMANDING DEMANDING Age Environment Age Environment HIGH Activity Contentment , . Contentment Resources At Time #1 4* At Time #1 At Time #1 LOW Activity Contentment :1 Contentment Resources At Time #1 At Time #1 At Time #1 Figure 7. Immediate cross sectional hypothesis-—consistency theory paradigm. 56 Figure 9. CONCENTRATED, NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment Nig?vity Contentment Contentment At Time #1 At Time #1 At Time #1 Aggivity Contentment Contentment At Time #1 At Time #1 At Time #1 8. Immediate cross sectional hypothesis--activity theory paradigm. CONCENTRATED, NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment HIGH Activity Contentment Contentment Resources At Time #1 i At Time #1 At Time #1 LOW Activity Contentment Contentment Resources At Time #1 At Time #1 At Time #1 Immediate cross sectional hypothesis--disengagement theory paradigm. 57 Hypotheses for each of the three theories were tested within this paradigm. The specific paradigm for consistency is shown in Figure 7, for activity in Figure 8, and for disengagement in Figure 9. The wording of the three hypotheses is repeated below for convenience. Within each paradigm the signs (> or <) indicate the direction of the hypothesized effect. HolC: (Consistency Theory,see Figure 7.) Contentment, among the elderly, measured at time #1 is dependent upon an interaction between activity resources measured at time #1 and age environment. Hol : (Activity Theory, see Figure 8.) Contentment, annng the elderly, measured at time #1 is depend- ent upon main effects for: a. Independence in activities of daily living measured at time #1 b. Days of bed disability measured at time #1. HolD:. (Disengagement Theory, see Figure 9.) Contentment, among the elderly, measured at time #1 is dependent upon a main effect for age environment measured at time #1 when crossed with factors measured at time #1. Delayed Cross Sectional Paradigm and Hypotheses. This para- digm (Figure 10-12) utilized cross sectional data from 196 of the 201 six-month interviews (time #2) for which complete data were available. For the analyses of covariance, the two covariates were: (1) age of the subject at the time of screening into the study, and (2) prior score (time #1) of the subject on the criterion variable contentment. Hypotheses for each of the three theories were tested within this paradigm. The specific paradigm for consistency is shown in Figure 10, for activity in Figure 11, and for disengage- ment in Figure 12. The wording of the hypotheses is given below. 58 CONCENTRATED, NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment HIGH Activity Contentment l Contentment Resources At Time #2 i» At Time #2 At Time #2 LOW : Activity Contentment I Contentment Resources At Time #2 At Time #2 At Time #2 Figure 10. Delayed Cross sectional hypotheses--consistency theory paradigm. CONCENTRATED, NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment Nig?vity Contentment Contentment At Time #2 At Time #2 At Time #2 Aggivity Contentment Contentment At Time #2 At Time #2 At Time #2 Figure 11. Delayed cross sectional hypotheses--activity theory paradigm. 59 CONCENTRATED. NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment HIGH Activity Contentment Contentment Resources At Time #2 j, At Time #2 At Time #2 LOW ) Activity Contentment Contentment Resources At Time #2 At Time #2 At Time #2 Figure 12. Delayed cross Sectional hypotheses--disengagement theory paradigm. H02C: (Consistency Theory, see Figure 10.) Contentment, among the elderly, measured at time #2, is dependent upon an interaction between activity resources meas- ured at time #2 and age environment. H02A: (Activity Theory, see Figure 11.) Contentment, among the elderly, measured at time #2 is dependent upon main effects for: a. Independence in activities of daily living measured at time #2 b. Days of bed disability measured at time #2. H020: (Disengagement Theory, see Figure 12.) Contentment, among the elderly, measured at time #2 is dependent upon a main effect for age environment measured at time #1, when crossed with factors measured at time #2. Chronological Lag Paradigm and Hypotheses. This paradigm (Figure 13-15) utilizes data from the initial interview (time #1) and the six-month interview (time #2) in combination. The number of such cases for which complete data were available was 196. For the analyses of covariance, the two covariates were: 60 CONCENTRATED, NONCONCENTRATED. NONDEMANDING DEMANDING Age Environment Age Environment HIGH Activity Contentment Contentment Resources At Time #2 k At Time #2 At Time #1 Low, . l. fict1v1ty Contentment Contentment esoirces At Time #2 At Time #2 At Time #1 Figure 13. Chronological lag hypothesis--consistency theory paradigm. CONCENTRATED, NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment Nig?vity Contentment Contentment At Time #1 At Time #2 At Time #2 Aggivity Contentment Contentment At Time #1 At Time #2 At Time #2 Figure 14. Chronological lag hypothesis--activity theory paradigm. 61 CONCENTRATED, NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment HIGH Activity Contentment Contentment Resources At Time #2 m At Time #2 At Time #1 LOW > Activity Contentment l Contentment Resources At Time #2 At Time #2 At Time #1 Figure 15. Chronological lag hypothesis--disengagement theory paradigm. (1) age of the subject at the time of screening into the study, and (2) prior score (time #1) of the subject on the criterion variable, contentment. Hypotheses for each of the three theories were tested within this paradigm. The specific paradigm for consistency is shown in Figure 13, for activity in Figure 14, and for disengagement in Figure 15. The wording of the three hypotheses is given below. H03C: (Consistency Theory, see Figure 13.) Contentment, among the elderly, measured at time #2 is dependent upon an interaction between activity resources measured at time #1 and age environment. H03A: (Activity Theory, see Figure 14.) Contentment, among the elderly, measured at time #2 is depend- ent upon main effects for; a. Independence in activities of daily living measured at time #1 b. Days of bed disability measured at time #1. H030: (Disengagement Theory, see Figure 15.) Contentment, among the elderly, measured at time #2 is dependent upon a main effect for age environment measured at time #1, when crossed with factors measured at time #1. 62 Chronological Anticipation Paradigm and Hypotheses. This paradigm (Figure 16-18) utilizes data from the initial interview (time #1) and the six—month interview (time #2) in combination. The number of such cases for which complete data were available was 192. For the analyses of covariance, age of the subject at time of screen- ing into the study was used as the only covariate. CONCENTRATED, NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment HIGH Activity Contentment Contentment Resources At Time #1 a At Time #1 At Time #2 LOW # Activity Contentment Contentment Resources At Time #1 At Time #1 At Time #2 Figure 16. Chronological anticipation hypothesis--consistency theory paradigm. CONCENTRATED, NONCONCENTRATED. NONDEMANDING DEMANDING Age Environment Age Environment Nig?vity Contentment Contentment At Time #2 At Time #1 At Time #1 Aggivity Contentment Contentment At Time #2 At Time #1 At Time #1 Figure 17. Chronological anticipation hypotheses--activity theory paradigm. 63 CONCENTRATED, NONCONCENTRATED, NONDEMANDING DEMANDING Age Environment Age Environment HIGH Acti Reso At T vity Contentment Contentment urces At Time #1 h At Time #1 ime #2 LOW Acti Reso At T L vity Contentment ) Contentment urces At Time #1 At Time #1 ime #2 Figure 18. “YD paradigm. ure 16, for ure 18. Th 595g; Ho4A: H040: Chronological anticipation hypotheses-~disengagement theory paradigm. otheses for all three theories were tested within this The specific paradigm for consistency is shown in Fig: activity in Figure 17, and for disengagement in Fig- e wording of the hypotheses is given below. (Consistency Theory, see Figure 16.) Contentment, among the elderly, measured at time #1 is dependent upon an interaction between activity resources measured at time #2 and age environment. (Activity Theory, see Figure 17.) Contentment, among the elderly, measured at time #1 is dependent upon main effects for: a. Independence in activities of daily living measured at time #2 b. Days of bed disability measured at time #2. (Disengagement Theory, see Figure 18.) Contentment, among the elderly, measured at time #1 is dependent upon a main effect for age environment measured at time #1, when crossed with factors measured at time #2. 64 Analysis of the Data All data analysis reported for this study made use of the Statistical Package for the Social Sciences (SPSS) routines available on the Control Data Corporation 6500 computer at Michigan State Uni- versity. For each of the two types of analysis discussed below, a set of analyses was run. The analysis set consisted of five differ- ent iterations of each of the four paradigms shown in Figures 7 through 18. The factor "age environment" was used in each of the five iterations. The second factor was defined by the following variables, in turn, for the five iterations: 1. Social support - residential location of the primary care giver (PCG) Social support - spouse survival Physical health days of bed disability Physical health activities of daily living 01-th Financial ability (score summarized from annual family income and index of economic dependence). Since each of the factors in every iteration was dichotomous, the analyses amounted to two-by-two fixed factor least squares model analyses with unequal cell sizes. The alternative of a single four- way analysis was rejected, due to small cell frequencies. The total analysis set is discussed below. To obtain the overall main effect for age environment, additional one—way analyses were performed for each paradigm. Analysis of Covariance The first level analysis of the data was accomplished using two-by-two analyses of covariance. All the analyses made use of the 65 variable of "age at time of screening into the study“ as a covariate, in order to control for the negative relation which age of subjects was expected to have on the dependent variable of contentment. This relation was reported in Table l as having been replicated in three studies. For the analyses for which the criterion variable was "con- tentment measured at time #2," an additional covariate was intro- duced. This second covariate, "contentment at prior measurement (time #1)," was necessary since the level of contentment at prior measurement was expected to effect the level of the criterion variable "contentment measured at time #2." Past studies have obtained a test-retest reliability of .65 for contentment. Analysis of covariance assumes that there is no interaction between the covariates and the factors. The test for significant interactions of this type was run using a regression procedure, and the level of interaction was nonsignificant in all cases for the total sample. Analysis of Variance The multiple classification analysis (MCA) output produced in conjunction with the analysis of covariance table indicated that the effectcfiithe covariates in several of the analyses was so great that, after adjustment had been made for the covariates, the sign of the deviations reversed for the two levels of age environment. Because itseemed likely that the covariate of subject's age is corre- lated with age environment, a decision was made to run analyses of variance, in addition to the analyses of covariance. 66 Cell Means and Frequencies In order to interpret the analysis of variance and analysis of covariance tables, the cell means, standard deviations, and number of cases were produced for each analysis. This data was identical for the two types of analysis, since the covariates were not involved in the cross-tabulated data. Analysis of Data, Controlling for Age Group, The above analyses were conducted on the total sample. The sample included persons nearing the status of elderly (45 to less than 65) so that we could compare the hypotheses for the late middle aged as opposed to the elderly. In order to accomplish this compari- son, the sample was divided into these two age groups (45 to less than 65, and 65 and over) and all the analyses described above were run again on each of the two groups. In some cases the division of the sample into two smaller groups of subjects caused analyses to be run with such a small frequency (10 or less) in one or more cells that neither the analysis of variance or covariance was reported. However, the test results were reported whenever possible. Altogether, 24 analyses of covariance, 24 analyses of vari- ance, 20 means, standard deviations, and frequency distributions were run for each of the three samples: 1. The total sample 2. The sample of older persons (aged 65 and older) 3. The sample of younger persons (aged 45 to less than 65) 67 Therefore, the complete data analysis package included 72 analyses of covariance, 72 analyses of variance, 60 means, standard deviations, and frequencies, and 18 tests for interaction of the covariates with the factors (four age covariates and two prior contentment score covariates for each of the three analysis samples). In some cases, the interaction of covariate and factor was found to be significant for the age samples, and, in such a circumstance, the analysis of covariance was not reported. “The results of the usable analyses will be reported in the following chapter. CHAPTER III RESULTS Introduction The results of this study are reported separately for three discrete samples. The first sample is the total sample. The other two samples for which results are reported are mutually exclusive and exhaustive subsamples of this total sample. The two subsamples are: ' (1) all persons in the total sample who are aged 65 or older, and (2) all persons in the total sample who are aged 45 but less than 65 years of age. The analyses done for the two subsamples explore the differential applicability of the study's 12 hypotheses to those who are unquestionably elderly compared to those who are considered "less" elderly from a comparative viewpoint or "pre-" elderly from a developmental viewpoint. For each sample, the results are organized so that first a report appears of the frequency of cases (subjects) within each cell of the analysis, along with means and standard deviations of the cri- terion variable, contentment. Secondly, the results of the analyses of covariance and variance are presented. All tables are presented in order of the four paradigms set forth in Chapter I and II. For each table, the order of presentation will be: 1. Factor A, age enviornment crossed with factor 8, residential location of primary care giver (PCG), 68 69 2. Factor A, age environment crossed with factor C, spouse survival, 3. Factor A, age environment crossed with factor 0, days of bed disability, 4. Factor A, age environment crossed with factor E, independence in activities of daily living (AOL), and 5. Factor A, age environment, crossed with factor F, financial ability. The first two analyses involve crossing age environment with social support factors, the third and fourth analyses cross age environment with physical health factors, and the last is an analysis of age environment by financial ability. All factors were defined in Chap- ter II. A final table for each sample gives the results of the one- way analyses for age environment main effect. Results for the Total Sample Descriptive Data It is apparent from the cross-tabulated frequencies, the first number presented in each cell in Table 8 to 11, that the dis- tribution of subjects over cells presents a difficulty in each analy— sis in which factor A, age environment, is crossed with factor 8, residential location of the PCG. It could be argued that analyses based on distributions with such small cell frequencies are suspect and should not be considered. Only those analyses having a cell frequency of 10 or more are reported. When the total sample is reduced further in size for the two subsamples, it will not be even minimally tenable to report the results from the analyses of these two factors. 7() Table 8. The j, M, and SD of Contentment at Time #1 for Total Sample--(H01C, A. D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Total FACTOR B--Time #1 Residential Location of Primaryygare Giver wfives in Sane 132a .100 232 Building as the 3.053b 3.050 3.052 Participant 1.421c 1.290 1.363 Lives Separate 17 11 28 from the 3.176 2.909 3.071 Participant .809 1.758 1.245 FACTOR C--Time #1 Spouse Survival Spouse 117 59 176 does 2.906 3.119 2.977 Survive 1.383 1.340 1.369 Spouse 32 52 84 does not 3.656 2.942 3.214 Survive 1.125 1.335 1.299 FACTOR D--Time #1 Days of Bed Disability No Days 73 56 129 of Bed 3.164 3.339 3.240 Disability 1.385 1.339 1.362 1 to 14 Days 76 55 131 of Bed 2.974 2.727 2.870 Disability 1.346 1.269 1.315 FACTOR E--Time #1 Independence in Activities of Daily Living Independent in 46 53 99 all 6 3.413 3.340 3.374 Activities 1.309 1.300 1.298 Dependent in 103 58 161 1 to 6 2.913 2.759 2.857 Activities 1.366 1.315 1.346 TOTALd 149 111 260d 3.067 3.036 3.054 1.364 1.334 1.349 FACTOR F--Time #1 Financial Ability Above the 95 52 147 Median of 3.053 3.173 3.095 Financial Ability 1.363 1.451 1.391 Below the 49 55 104 Median of 3.143 2.836 2.981 Financial Ability 1.369 1.214 1.292 TOTAL 144 107 251 3.083 3.000 3.048 1.361 1.339 1.350 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. Table 9: The I, M, and §Q of Contentment at Time #2 for Total Sample--(H02C, A, D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Total FACTOR B--Time #2 Residential Location of Prima:y_Care Giver Tfilives in Same 93a 74 167 Building as the 3.462b 3.365 3.419 Participant 1.403c 1.540 1.462 Lives Separate 18 ' 11 29 from the 3.556 2.364 3.103 Participant 1.542 1.433 1.589 FACTOR C--Time #2 5 use Survival gpouse 86 49 135 does 3.384 3.224 3.326 Survive 1.440 1.571 1.485 Spouse 25 36 61 does not 3.800 3.250 3.475 Survive 1.323 1.556 1.479 FACTOR D--Time #2 Days of Bed Disability No Days 83 61 144 of Bed 3.687 3.541 3.625 Disability 1.325 1.478 1.389 1 to 14 Days 28 24 52 of Bed 2.857 2.458 2.673 Disability 1.533 1.503 1.517 FACTOR E--Time #2 Independence in Activities eipfleillLLixjns Independent in 65 59 124 all 6 3.754 3.610 3.685 Activities 1.323 1.474 1.393 Dependent in 46 26 72 l to 6 3.087 2.385 2.833 Activities 1.473 1.416 1.482 TOTALd 111 as 1964 3.477 3.235 3.372 1.420 1.556 1.481 FACTOR F-—Time #2 Financial Ability ——Above the 43 41 » 84 Median of 3.419 3.439 3.429 Financial Ability 1.484 1.550 1.507 Below the 67 44 111 Median of 3.522 3.045 3.333 Financial Ability 1.397 1.555 1.473 TOTAL 110 85 195 ' 3.482 3.235 3.374 1.425 1.556 1.485 aFrequency within the cell. 6 Mean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. 72 Table 10: The 1, M, and §Q_of Contentment at Time #2 for Total Sample--(Ho3C. A, D). FACTOR A-—Age Environment Concentrated Nonconcentrated Nondemanding Demanding Total FACTOR B--Time #1 Residential Location of Primary Care Giver '_Lives in Same 97a 77 174 Building as the 3.5st 3.260 3.402 Participant 1.370c 1.542 1.450 Lives Separate 14 B 22 from the 3.214 3.000 3.136 Participant 1.762 1.773 1.726 FACTOR C--Time #1 Spouse Survival Spouse 87 48 135 does 3.333 3.188 3.281 Survive 1.444 1.566 1.485 Spouse 24 37 61 does not 4.000 3.297 3.574 Survive 1.216 1.561 1.466 FACTOR D--Time #1 Days of Bed Disability No Days 56 48 104 of Bed 3.554 3.292 3.433 Disability 1.400 1.611 1.499 1 to 14 Days 55 37 92 of Bed 3.400 3.162 3.304 Disability 1.448 1.500 1.466 FACTOR E--Time #1 Independence in Activities of Daily Living Independent in 37 46 83 all 6 3.595 3.261 3.410 Activities 1.462 1.612 1.546 Dependent in 74 39 113 1 to 6 3.419 3.205 3.345 Activities 1.405 1.508 1.438 TOTALd 111 as 195‘I 3.477 3.235 3.372 1.420 1.556 1.481 FACTOR F--Time #1 Financial Ability Above the 62 42 104 Median of 3.484 3.476 3.481 Financial Ability 1.376 1.581 1.455 Below the 49 42 91 Median of 3.469 3.000 3.253 Financial Ability 1.487 1.530 1.517 TOTAL 111 84 195 3.477 3.238 3.374 1.420 1.565 1.485 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. 73 Table 11: The I, M, and S9 of Contentment at Time #1 for Total Sample--(H04C. A, D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Total FACTOR B--Time #2 Residential Location of Rrimaiy Care Giver a _“[ives in Same 91b 71 162 Building as the 3.110 3.225 3.160 Participant 1.41:3c 1.301 1.365 Lives Separate 19 11 30 from the 3.316 2.364 2.967 Participant 1.293 1.502 1.426 FACTOR C--Time #2 S use Survival gpouse 87 46 133 does 2.977 3.065 ' 3.008 Survive 1.414 1.421 1.412 Spouse 23 36 59 does not 3.783 3.167 3.407 Survive 1.126 1.276 1.247 FACTOR D--Time #2 Days of Bed Disability - N0 Days 83 58 141 of Bed 3.157 3.310 3.220 Disability 1.357 1.353 1.353 1 to 14 Days 27 24 51 of Bed 3.111 2.625 2.882 Disability 1.528 1.245 1.409 FACTOR E--Time #2 Independence in Activities of Dail Living Independent in 65 56 121 all 6 3.415 3.339 3.380 Activities 1.211 1.379 1.286 Dependent in 45 26 71 1 to 6’ 2.756 2.615 2.704 Activities 1.554 1.169 1.418 TOTALd 110 82 192d 3.145 3.110 3.130 1.394 1.352 1.372 FACTOR F--Time #2 Financial Ability_ _’Ab0ve the 43 39 82 Median of 3.070 3.359 3.207 Financial Ability 1.438 1.460 1.446 Below the 66 43 109 Median of 3.167 2.884 3.055 Financial Ability 1.365 1.219 1.311 TOTAL 109 82 191 3.128 . 3.110 3.120 1.388 1.352 1.369 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. 74 In Table 8 and Table 11 the criterion variable is contentment measured at time #1, and in Table 9 and Table 10 the criterion varia- ble contentment is measured at time #2. Furthermore, the total fre- quency in Table 8 is considerably greater than in the other three tables, since this table reports on variables obtained only from time #1 data before the experimental and actual mortality experi- enced at time #2 (six months after time #1). Comparison of appro- priate tables for the same factors or criterion allows one to make an assessment of changes in these variables from time #1 to time #2. Further inspection of tables 8 through 11 indicates that factor 8, residential location of primary care giver, and factor C, spouse survival, remain fairly stable across the set of four tables, while factor 0, days of bed disability; factor E, independence in activities of daily living (AOL); and factor F, financial ability, fluctuate across the four tables. Three factors show change from time #1 (in Tables 8 and 10) to time #2 (in Tables 9 and 11). Some limited conclusions may be drawn regarding these changes. For factor 0, days of bed disability, either those who were initially bed-disabled improve, or they are disproportionately lost to experimental mortality. For factor E, independence in AOL, either those who were initially dependent improve, or they are disproportionately lost from the sample. For factor F, financial ability, either those who were initially above the median of financial ability have lessened financial ability by time #2, or they have been disproportionately lost from the sample. In addition, com- parison of means and standard deviations for contentment in Table 8 75 and 11 with statistics in Tables 9 and 10 indicates that the amount of variation and mean level of contentment increased from time #1 to time #2. Comparison of the "high resource" level with the "low resource" level of factor 8, residential location of PCG; factor C, spouse survival; factor 0, days of bed disability; factor E, activi- ties of daily living; and factor F, financial ability; shows that contentment level is either higher or nearly equal for "high resource" as compared to "low resource." There is one outstanding exception to this general observation. In each of the tests of hypotheses, it is the classification "spouse does not survive" which has a higher mean value for contentment. Since the spouse's survival was viewed as an activity resource which could be characterized as a "high social support resource," this is a surprising finding.. In summary, we have identified factor A, age environment, related to factor B, residential location of PCG, as an analysis which has small cell frequencies. Factor B, residential location of PCG, and factor C, spouse survival, were identified as having marginal frequencies which remain stable from time #1 to time #2. The marginal frequencies of factor 0, days of bed disability; factor E, ADL; factor F, financial ability; and the criterion variable of contentment fluctuate from time #1 to time #2. Spouse survival is consistently associated with lower mean contentment level than is spouse non- survival. 76 Analysis of Covariance and Variance For the total sample, it is necessary to forego the reporting of the analyses of variance and covariance for factor A, age environ- ment, by factor 8, residential location of PCG, in the testing of H03, due to small cell frequency. All data from this study not reported in this chapter may be found in Appendix C. It is apparent, from the fact that significant f_values were obtained for the covariates of age and prior (time #1) score in every analysis of covariance, that the choice of these two covariates was validated. These E_values are significant at p_< .01. However, because factor A, age environment, which appears in each analysis, could be highly correlated with the covariate of age, which was used in every analysis of covariance, we have also run the corresponding analyses of variance for all factors and hypotheses. In Tables 12 to 15, analyses of covariance are reported in the first column and analyses of variance are reported in the second column. Thus it is relatively convenient to compare the result of these two analytic approaches. Of greatest interest in such compari- sons are situations in which the two analyses produce results which are significant for one analysis and not significant for the other analysis. These types of results are observed in the following analyses: 1. (Table 12, HolA.)--the analysis of factor A, age environment, by factor 0, days of bed disability. The significant main effect for factor 0, in ANOVA is nonsignificant in ANCOVA. 77 Table 12: Analysis of Covariance and Variance for Total Sample--(H01C. A. D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Age 1 20.010 11.359“r A=Age Environment 1 1.954 1.109 1 .060 .033 B=Residentia1 Location of Primary Care Giver 1 .000 .000 1 .009 .005 A X 8 Interaction 1 .086 .049 1 .418 .227 Residual 255 1.762 256 1.839 Total 259 1.819 259 1.819 Covariate=Age 1 20.010 11.679** AsAge Environment 1 1.262 .737 1 .570 .320 CBSpouse Survival 1 .246 .144 1 3.703 2.078 A X C Interaction 1 12.154 7_094e* 1 11.301 5.342* Residual 255 1.713 256 1.782 Total 259 1.819 259 1.819 Covariate=Age 1 20.010 '11 .588“ A=Age Environment 1 1.538 .891 1 .085 .047 D=Days of Bed Disability_ 1 5.303 3.071 1 8.925 4.973* A X 0 Interaction 1 3.661 2.120 1 2.822 1.573 Residual 255 1.727 256 1.795 Total 259 1.819 259 1.819 Covariate=Age 1 20.010 11.870** A=Age Environment 1 1.563 .334 1 .917 .517 E=Independence in ACth° ities of Dailnyiving, 1 19.403 11.510** 1 17.215 9.710** A X E Interaction 1 .031 .018 1 .096 .054 Residual 255 1.686 256 1.773 Total 259 1.819 259 1.819 Covariate=Age 1 20.627 11.854‘” A=Age Environment 1 2.192 1.260 1 .255 .139 F=Financia1 Ability 1 3.493 2.007 1 .627 .343 A X F Interaction 1 2.110 1.213 1 2.667 1.458 Residual 246 1.740 247 1.829 Total 250 1.822 250 1.822 *p_< .05 *‘p < .01 78 Table 13: Analysis of Covariance and Variance for Total Sample-~(H02C, A. D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Prior Score 1 57.437 33.506** Covariate=Age 1 20.105 11.728** A=Age Environment 1 .157 .091 1 3.076 1.422 B=Residentia1 Locati0n of Primagy Care GivEr 1 1.869 1.090 1 2.716 1.256 A X 8 Interaction 1 3.047 1.778 1 7.015 3.244 Residual 183 1.714 192 2.163 Total 188 2.198 195 2.194 Covariate=Prior Score 1 57.437 33.235** Covariate=Age 1 20.105 ll.634** A=Age Environment 1 .034 .019 1 3.725 1.696 C=Sp0use Survival 1 .283 .164 1 1.840 .838 A X C Interaction 1 2.081 1.204 1 1.530 .697 Residual 183 1.728 192 2.198 Tetal‘ 188 2.198 195 2.194 Covariate=Prior Score 1 57.437 35.218** Covariate=Age 1 20.105 12.327** A=Age Environment 1 .018 .011 1 2.197 1.081 D=Days of Bed Disability 1 20.079 12.311** 1 33.993 16.718** A’X 0 Interaction 1 .077 .047 1 .605 .298 Residual 183 1.631 192 2.033 Total 188 2.198 195 25194 Covariate=Prior Score 1 57.437 35.804** Covariate=Age 1 20.105 12.533** A=Age Environment . l .308 .192 1 5.460 2.716 E=Independence in Activ- ~ ities of Dolly Living 1 22.344 13.929“r 1 35.713 17.769** A X E Interaction 1 2.710 1.689 1 3.373 1.678 Residual 183 1.604 192 2.010 Total 188 2.198 195 2.194 Covariate=Prior Score 1 59.068 34.125** Covariate=Age 1 21.111 12.198** A=Age Environment 1 .196 .113 1 3.150 1.428 F=Financia1 Abiligy 1 .868 .501 l .670 .304 A X F Interaction l .085 .049 1 2.900 1.315 Residual 182 1.731 191 2.205 Total 187 2.209 194 2.204 *p < .05 “p < .01 79 Table 14: Analysis of Covariance and Variance for Total Sample-—(H03C, A. D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Prior Score Covariate=Age A=Age Environment Data Not Reported Data Not Reported B=Residentia1 Location Cell Sizes = 97 77 Cell Sizes = 97 77 of PrimaryyCare Giver 14 8 14 8 A X 8 Interaction Residual Total Covariate=Prior Score 1 57.437 33.312“r Covariate=Age 1 20.105 11.660** A=Age Environment 1 .122 .071 1 4.776 2.203 CESpQuse_Surxixal_ 1 .005 .003 1 5.543 2.556 A x c Interaction 1 3.085 1.789 1 3.070 1 415 Residual 183 1.724 192 2.169 Total 188 2.198 195 2.194 Covariate=Prior Score 1 57.437 33.042** Covariate=Age 1 20.105 11.566** A=Age Environment 1 .109 .063 1 3.017 1 366 1 1 1 D=Days of Bed Disability .407 .234 .998 .452 A X 0 Interaction .103 .059 l -007 .003 Residual 183 1.738 192 2 208 Total . 188 2.198 195 2 194 Covariate=Prior Score 1 57.437 33.057** Covariate=Age 1 20.105 11.571** A=Age Environment 1 .112 .064 1 3.288 1.488 E=Independence in Activ- ities of Daily Living 1 .090 .052 1 .663 .300 A X E Interaction l .570 .328 1 .164 .074 Residual 183 1.738 192 2 209 Total 188 2.198 195 2.194 Covariate=Prior Score 1 57.798 33.363** Covariate=Age 1 19.906 11.49l** A=Age Environment 1 .020 .012 1 2.451 1.114 F=Financial Ability 1 1.806 1.042 1 2.235 1.016 A X F Interaction 1 1.127 .650 1 2.533 1.151 Residual 182 1.732 191 2.200 Total 187 2.209 194 2.204 t p_< .05 it p_< .01 8C) Table 15: Analysis of Covariance and Variance for Total Sample-~(H04C. A, D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Age 1 13.780 7.670“ . A=Age Environnent 1 1.167 .650 1 .088 .047 8=Residentia1 Location of PrimaryyCare Giver 1 1.476 .822 1 .979 .523 A X 8 Interaction 1 7.265 4.044* 1 6.760 3.611 Residual 187 1.797 188 1.872 Total 191 1.883 191 1.883 Covariate=Age 1 13.780 7.661** A=Age Environnent l .285 .158 l .813 .440 C=Spouse Survival 1 1.935 1.076 1 7.268 3.930* A X C Interaction 1 6.408 3.562 1 4.745 2.566 Residual 187 1.799 188 1.849 Total 191 1.883 191 1.883 Covariate=Age 1 13.780 7.644** A=Age Environment 1 1.474 .818 l .018 .010 D=Days of Bed Disability 1 4.172 2.314 1 4.225 2.259 A X 0 Interaction 1 3.451 1.914 1 3.791 2.027 Residual 187 1.803 188 1.871 Total 191 1.883 191 1.883 Covariate=Age 1 13.780 8.074** A=Age Environment 1 .761 .446 1 .454 .252 E=Independence in Activ- ities of Daily Living 1 25.566 l4.981** 1 20.838 11.563** A X E Interaction 1 .015 .009 1 .044 .024 Residual 187 1.707 188 1.802 Total 191 1.883 191 1.883 Covariate=Age 1 12.602 6.957** A=Age Environment 1 1.368 .755 1 .045 .024 F=Financial Ability 1 3.849 2.125 1 1.114 .593 A X F Interaction 1 1.505 .831 1 3.750 1.996 Residual 186 1.811 187 1.879 Total 190 1.875 190 1.875 *p < .05 *‘p < .01 81 mo. v.me _.o. v mt. mmm._ Fm_ mmm._ _mp _eeoe mam.P amp eNm._ mm, _e=e_nom Nmo. oeo. . P Mme. ee~._ _ eeoEeoce>em om< n < 266682 eemmm.e oma.m_ F oma n oeeece>ou Noeozv em_.~ mm_ mmP.N wm_ _eeoe Pm_.m amp NNN.F mm_ .eeeemom mmm._ mmm.~ _ Nae. amp. P ecoEeocw>em 66< n < coeoea ensue... mo_.om P om< n oea_ce>ou enmem.mm ame.am _ ocoom coecanooeece>ou Ammo: 6:6 among mpm.F mmm mpm._ mmm _eeoe emm._ wmm wee._ New _e=e_nom emo. _eo. _ m__._ emm._ _ oeoEeoce>em owe n < coeoau eneae.__ o_o.om _ om< n oeewce>ou ~o_ozv mspo> wcmzcm Eonmmcm m3~m> mcmzcm Eoummcm cowpmecw> we muczom g com: mo mmmcmmo m com: eo mmmcmmo mocowcm> mo umAch< mocmwcm>oo mo mwmeec< .mFaEmm Pouch com newscocw>cm mmimocmwem>ou no mwmapmc< Amzimco "up epoch 82 2. (Table 15, H04A)--analysis of factor A, age environment, by factor C, spouse survival. The nonsignificant main effect for factor C, spouse survival, in ANCOVA is significant in ANOVA. 3. (Table 15, H04C.)--analysis of factor A, age environ- ment, bv factor 8, residential location of PCG. The significant interaction effect for these two factors in ANCOVA is nonsignificant in ANOVA. While there were these important differences in the findings from the two analyses, four main effects and one interaction effect were found to be significant, regardless of whether the analysis used was ANCOVA or ANOVA. Consistency Hypotheses. Significant interaction effects are the results needed to support the consistency hypotheses (HolC. to H04C.) of this study. The results for the total sample yield two such results out of the 20 sets of analyses. We may state these two find- ings as follows: 1. There is a significant interaction (ANCOVA and ANOVA) between factor A, age environment, and factor C, spouse survival at time #1, when the criterion is contentment at time #1 (see Table 12). Contentment is higher for persons in nonconcentrated demanding age environment with a surviving spouse and for persons in a concen- trated, nondemanding age environment without a surviv- ing spouse (see Table 8). 2. There is a significant interaction (ANCOVA only) between factor A, age environment, and factor 8, residential location of PCG at time #2, when the criterion is contentment at time #1 (see Table 15). Contentment is higher for persons in a nonconcentrated, demanding age environment who have a PCG living in the same building (at a later time), and for persons in a concentrated, nondemanding age environment who have a PCG living in a different building (at a later time) (see Table 11). Comparison of the results in these two tables (8 and 11) with the general consistency paradigm (Figure 4) indicates that both 83 findings support the consistency hypotheses. No other effects support- ing the consistency hypotheses were observed in these analyses. Disengagement Hypotheses. The obtained significant main effects for the total sample have been summarized in Table 17. These main effects of significance may be divided into those which support activity theory hypotheses, those which support disengagement theory hypotheses.and those for which no hypotheses exist. In this sample, none of the significant main effects are in a direction which would negate the activity or disengagement hypotheses. The disengagement hypotheses are not supported by any significant main effects. The one-way analyses for age environment, reported in Table 16, show no significant main effects. Activity Hypotheses. The activity hypotheses are supported by two significant main effects for factor 0, days of bed disability, (H02A.); and three significant main effects for factor E, activities of daily living, (HolA., H02A, and Ho4A.). Thus, in the total sample, three of the four activity hypotheses have been supported by five of the eight sets of tests of main effects mentioned in these hypotheses. All these significant findings were in the hypothesized direction; that is, contentment was higher for persons with no days of bed dis- ability (vs some days of bed disability) and with independence in AOL (vs dependence in AOL). One of the main effects which is significant is unrelated to either the activity or disengagement hypotheses. This finding is for a main effect for factor C, spouse survival. The direction of the 84 p0. v .mr« mo. v mr Ne meme .oewe coee_ e ea mew>e_ zpeee to _e cave on oo< ee m_ o_eee mmmugzum xwm :m 5 pcmncwamvs 3.x «5.x pm ucflscmucou mucmvcmamus m . Ne meek .oee>e_ »_eee to we cave on oo< ew m_ opeae mmwpgzuw xwm _._.m E. pcmvcwamug 33.x 14.x um “cw—Scmpcou mucmvcmamncH m _e oe_e .mee>ee »F_eo to Fe weep ea ea oo< e_ we o_eee mmwpw>wpom xwm PPM cw ucwucmqmncH wax «ex pcmsucmucou mocmucmqmvcH m we deep on .eooeoecoaxo Ne meek ea xee_weeneo m_ o_eee ace xee_eeem_o eom co mean oz 4.x 4.x eeoEeeoeeou com co mean a fie deep eooeoecoaxo _e deep on ee xee__een_o NP o_eee eta xew_weemeo eom co memo oz ex eeoEeeoeeou eom co meme 0 Ne meek _e cave on on Fe>e>c=m m_ opeee .mewp cmpm_ e um m>w>czm we: moon mmaoam «x “cospcmucou mmzoam u Heme: cases: We eeoEeeoeeou <>oz< <>ooz< "nH "coo coeoeu eoeeoeecu “once“ are: "we eon: ewe: eeeoeceemem mwm>ch< can: .oFQEem _eeoe e2 eased meooccm ewe: eeeoeceemem ”up wpnmh 85 effect is for high contentment when the spouse does not survive. This effect had been referred to in Chapter I, as a possible mediating effect in support of disengagement theory. Since it was not deemed directly related to disengagement phenomena, no hypotheses were formed regarding this effect. This finding of strong support for this main effect will lead to further discussion in the following chapter. In summary, for the total sample, little support is shown for the consistency hypotheses of this study. At best, such findings are of two significant interaction effects from among 20 sets of tests for interaction. No support is found for the disengagement hypotheses within the total sample. Activity theory presents a different picture. Five of eight sets of tests of the activity hypotheses yielded sig- nificant main effects. In addition, one set of tests yielded signifi- cant main effects for nonsurvival of spouse, to which no formal hypotheses were linked. Results for the Older (65 or Oldey) Sample Descriptive Data The unequal frequency distribution of cases within cells which was noted in the total sample for factor A, age environment, by factor 8, residential location of PCG, is even more apparent in this sub- sample, effecting these two factors in all four tests of the hypothe- ses (see Tables 18 to 21). In this older sample, two additional sets of analyses are affected by small cell frequencies. The analyses affected in this way are factor A, age environment, by factor C, spouse survival (all three hypotheses tested on time #2 sample), and 236 Table 18: The f, M, and §Q_0f Contentment at Time 41 for Older Sample--(H01C. A, D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Total FACTOR B--Time #1 Residential Location of Winery Care Giver ' Lives in Same 88; 41 129 Building as the 3.125 3.561 3.264 Participant 1.346c 1.050 1.272 Lives Separate 13 2 15 from the 3.154 3.000 3.133 Participant .801 2.828 1.060 FACTOR C--Time #1 Spouse Survival Spouse 76 12 88 does 2.921 3.833 3.045 Survive 1.283 .937 1.277 Spouse 25 31 56 does not 3.760 3.419 3.571 Survive 1.091 1.177 1.142 FACTOR D--Time #1 Days of Bed Disability N0 Days 58 24 82 of Bed 3.155 3.792 3.341 Disability 1.322 1.141 1.298 1 to 14 Days 43 19 62 of Bed 3.093 3.211 3.129 Disability 1.250 1.032 1.180 FACTOR E--Time #1 Independence in Activities of Dail Livin Independent in 30 19 49 all 6 3.400 3.842 3.571 Activities 1.303 1.167 1.258 Dependent in 71 24 95 1 to 6 3.014 3.292 3.084 Activities 1.270 1.042 1.217 TOTALd 101 43 144d 3.129 3.535 3.250 1.286 1.120 1.249 FACTOR F--Time #1 Financial Ability Above the 58 15 73 Median of 3.069 3.733 3.205 Financial Ability 1.269 1.163 1.269 Below the 41 26 67 Median of 3.293 3.346 3.313 Financial Ability 1.289 1.093 1.209 TOTAL 99 41 140 3.162 3.488 3.257 1.275 1.121 1.237 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. 87 Table 19. The f, M, and 59 of Contentment at Time #2 for Older Sample-~(H02C. A. D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Tota1 FACTOR Bo-Time #2 Residential Location of FFTmary Care Giver ——[1ves in Same 64g 25 89 Building as the 3.766c 3.880 3.798 Participant 1.178 1.364 1.226 Lives Separate 14 5 19 from the 3.571 3.000 3.421 Participant 1.555 1.414 1.502 FACTOR C-—Time #2 Spouse Survival Spouse 6O 7 67 does 3.633 3.857 3.657 Survive 1.248 1.345 1.250 Spouse 18 23 41 does not 4.056 3.690 3.854 Survive \ 1.211 1.428 1.333 FACTOR D--Time #2 Days of Bed Disability No Days 58 21 79 of Bed 3.914 3.905 3.911 Disability 1.144 1.261 1.168 1 to 14 Days 20 9 29 of Bed 3.200 3.333 3.241 Disability 1.399 1.658 1.455 FACTOR E--Time #2 Independence in Activities 0‘ 0411mm Independent in 41 19 60 all 6 4.098 4.211 4.133 Activities 1.020 1.228 1.081 Dependent in 37 11 48 1 to 6 3.324 .909 3.229 Activities 1.355 1.300 1.341 TOTALd 78 30 108d 3.731 3.733 3.731 1.245 1.388 1.280 FACTOR F-—Time #2 Financial Ability "Above the 21 14 35 Median of 3.857 4.000 3.914 Financial Ability 1.236 1.177 1.197 Below the 56 16 72 Median of 3.696 3.500 3.653 Financial Ability 1.264 1.549 1.323 TOTAL 77 30 107 3.740 3.733 3.738 1.250 1.388 1.284 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. 88 Table 20: The I, M, and §Q_of Contentment at Time #2 for Older Sample--(H03C. A. D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Tota1 FACTOR 8--Time #1 Residential Location of Primary Care Giver Lives in Same 67g 28 95 Building as the 3.806 3.679 3.768 Participant 1.131c 1.416 1.216 Lives Separate 11 2 13 from the 3.273 4.500 3.462 Participant 1.794 .707 1.713 FACTOR C--Time #1 Spouse Survival Spouse 61 7 68 does 3.557 3.857 3.588 Survive 1.272 1.345 1.272 Spouse 17 23 40 does not 4.353 3.696 3.975 Survive .931 1.428 1.271 FACTOR D--Time #1 Days of Bed Disability No Days 45 19 64 of Bed 3.756 3.789 3.766 Disability 1.246 1.512 1.318 1 to 14 Days . 33 ll 44 of Bed 3.697 3.636 3.682 Disability. 1.262 1.206 1.235 FACTOR E--Time #1 Independence in Activities of Daily Living Independent in 27 17 44 all 6 3.815 4.000 3.886 Activities 1.302 1.369 1.316 Dependent in 51 13 64 1 to 6 3.686 3.385 3.625 Activities 1.225 ' 1.387 1.254 FACTOR F--Time #1 Financial Ability Above the 37 13 50 Median of 3.865 4.231 3.960 Financial Ability 1.004 1.092 1.029 Below the 41 17 58 Median of 3.610 3.353 3.534 Financial Ability 1.430 1.498 1.441 TOTAL 78 30 108 3.731 3.733 3.731 1.245 1.388 1.280 aFrequency within the cell. bMean of the criterion variable within the cell. ‘Standard Deviation of the criterion variable within the cell. 139 Table 21: The 1. fl, and S9 of Contentment at Time #1 for Older Sample--(Ho4c. A. D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Total FACTOR 8--Time #2 Residential Location of Primary Care Giver "tives in Same ' 62: 25 37 Building as the 3.161 3.680 3.310 Participant 1.333c .988 1.260 Lives Separate 13 5 18 from the 3.308 2.800 3.167 Participant 1:377 1.789 1.465 FACTOR C--Time #2 Spouse Survival Spouse 58 7 65 does 2.983 3.857 3.077 Survive 1.331 1.069 1.327 Spouse 17 23 40 does not 3.882 3.435 3.625 Survive 1.111 1.199 1.170 FACTOR D--Time #2 Days of Bed Disability No Days 56 21 77 of Bed 3.196 3.571 3.299 Disability 1.313 1.207 1.288 1 to 14 Days 19 9 28 of Bed 3.158 3.444 3.250 Disability 1.425 1.130 1.323 FACTOR E--Time #2 Independence in Activities of Dail Livin Independent in 40 19 59 all 6 3.475 3.632 3.525 Activities 1.154 1.300 1.194 Dependent in 35 11 46 1 to 6 2.857 3.364 2.978 Activities 1.458 .924 1.358 d d 75 30 105 TOTAL 3.187 3.533 3.286 1.332 1.167 1.291 FACTOR F--Time #2 Financial Ability Above the 20 14 34 Median of 2.900 3.714 3.235 Financial Ability 1.210 1.139 1.232 Below the 54 16 70 Median of 3.259 3.375 3.286 Financial Ability 1.362 1.204 1.320 TOTAL 74 - 30 104 3.162 3.533 3.269 1.324 1.167 1.286 °Frequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. 9O factor A, age environment, by factor D, days of bed disability (two of the three hypotheses tested on the time #2 sample). I In general, the marginal frequencies for all factors behave much the same fbr this subsample as they do for the total sample, factor 8, residential location of PCG, and factor C, spouse survival, are stable, and factor D, days of bed disability, factor E, indepen- dence of AOL, and factor F, financial ability, fluctuate from time #1 to time #2 as described for the total sample. The finding of higher mean contentment at time #2 than at time #1 is also apparent for this older sample. The finding that mean contentment is higher for participants whose spouse does not survive is also upheld for this subsample. Contentment is higher and less variable for the older sample than for the total sample. Although this latter finding appears to be at odds with the negative relation of age and contentment cited earlier, it is a find- ing consistent with the results obtained by Loeb and his colleagues (1963, see Figure 3). It is quite likely that the relation of age and contentment is curvilinear, being negatively related at ages from 45 to 65 and being unrelated or positively related to contentment at ages beyond 65, as has been shown by this research. In summary, we nay say that all findings from the descriptive data for the total sample are upheld by the findings from the descriptive data for the older sample. In addition, drastically small cell frequencies have been observed in the older sample for factor A, age environment, by factor C, spouse survival, and factor A, age environment, by factor 0, days of bed disability. 91 Analyses of Covariance and Variance For this older sample, data have not been reported due to small cell frequency for factor A,—age environment, by factor 8, residential locations of PCG (all four tests of the hypotheses), for factor A, age environment, by factor C, spouse survival (tests of H02., H03., and H04.), and for factor A, age environment, by factor 0, days of bed disability (tests of H02. and H04.). In addition, data from several analyses of covariance have not been reported because significant interactions were found between factors and covariates, in violation of an underlying assumption of the analysis of covariance. In Hol. (Table 22 and Table 26), there is a significant interaction of the covariate age with factor A, age environment, making all analyses of covariance for that hypothesis invalid. In H02. (Table 23), the covariate age interacts significantly with factor 0, days of bed disability. All results of data analysis not reported in this chapter may be found in Appendix C. There are no differential findings of significant E_values between the analysis of covariance and the analysis of variance for this older sample. At the same time, there is evidence that age is not an appropriate covariate for this subsample of persons 65 years of age or older, because the_§ values for the age covariate were nonsignificant in every analysis. Since the covariate of age had little effect in this sample, it is not surprising that no differ- ential results between the two types of analyses have been found. Table 22: 92 Analysis of Covariance and Variance for Older Sample-~(HolC. A, D). Source of Variation Analysis of Covariance Analysis of Variance Mean F Square Value Degrees of Freedom Mean F Square Value Degrees of Freedom Covariate=Age A=Age Environment B=Residential Location Data Not Reported Cell Sizes = 83 41 Data Not Reported Cell Sizes = 88 41 of Primary Care Giver 13 2 l3 2 A x 8 Interaction Interaction of Factor A Residual with Age Total p.= .017 Covariate=Age A=Age Environment Data Not Reported 1 .928 .639 C=Spouse Survival Interaction of Factor A 1 5.420 3.733 A x C Interaction with Age 1 9.303 6.406* Residual p = .017 140 1.452 Total " 143 1.559 Covariate=Age A=Age Environment Data Not Reported 1 5.061 3.306 D=Days of Bed Disability, Interaction of Factor A 1 1.679 1.096 A x 0 Interaction with Age 1 1.998 1.305 Residual p_= .017 140 1.531 Total 143 1.559 Covariate=Age A=Age Environment Data Not Reported 1 3,455 2,292 E=Independence in Activ- Interaction of Factor A ities of Daily Living, with Age 1 6,163 4,075* A X E Interaction 2.: .017 1 .191 .126 ’ Residual 140 1.512 Total 143 1.559 Covariate=Age A=Age Environment Data Not Reported 1 2,753 1.815 F=Financial Abiliqi Interaction of Factor A 1 .085 .055 A X F Interaction with Age 1 2,543 1 570 Residual p_= .017 136 1.522 Total 139 1.531 *p_< .05 **p < .01 Table 23: 923 Analysis of Covariance and Variance for Older Sample--(H02C. A. D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Prior Score Covariate=Age A=Age Environment . B=Residential Location Data Not Reported Data Not Reported of Primary_Care Giver Cell Sizes = 64 25 Cell Sizes = 64 25 A x 8 Interaction 14 5 14 5 Residual Total Covariate=Prior Score Covariate=Age A=Age Environment Data Not Reported Data Not Reported §3§pouse Survival Cell Sizes = 60 7 Cell Sizes = 60 7 A X C Interaction 18 23 18 23 Residual Total Covariate=Prior Score Covariate=Age Data Not Reported Data Not Reported A=Age Environment Cell Sizes = 58 21 Cell Sizes = 58 21 D=Days of Bed Disability 20 9 20 9 A—Y'D Interaction Interaction of Factor 0 Residual with Age Total p_= .043 Covariate=Prior Score 1 20-353 ‘5-553** Covariate=Age l -015 ~01? =Age Environnent l .919 .702 l .197 .135 E=Independence in Activ- ities of Daily Living» 1 13.331 10.185** 1 21.997 15.072“r A X E Interaction 1 2.290 1.749 1 1.431 .981 Residual 99 1.309 104 1.459 Total 104 1.601 107 1.638 Covariate=Prior Score 1 21.615 15.034** Covariate=Age ‘ 060 .041 ' A=Age Environment 1 1.304 .907 1 .074 .044 F=Financia1 Abiligé ‘ 2.149 1.495 1 1.684 1.006 A x F interaction 1 .039 .027 1 .577 .345 Residual 98 1.438 103 1.674 Total ‘03 1.611 106 1.648 * 2,< .05 it p_< .01 94 Table 24: Analysis of Covariance and Variance for Older Sample--(H03C. A. D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Prior Score Covariate=Age A=Age Environment Data Not Reported Data Not Reported B=Residential Location Cell Sizes = 67 28 Cell Sizes = 67 28 of PrimaryyCare Giver 11 2 11 2 A X 8 Interaction Residual Tota1 Covariate=Prior Score Covariate=Age A=Age Environment Data Not Reported Data Not Reported C=Spouse Survjval Cell Sizes = 61 7 Cell Sizes = 61 7 A X C Interaction 17 23 17 23 Residual Total Covariate=Prior Score 1 .20.:52 13.324** “m. i . D=Days of Bed Disability 1 -452 ~309 1 "83 -‘°9 A7X 0 Interaction l -0]3 '009 I ’046 '027 . 1,452 104 1.683 Re51dua1 99 - Covariate=Prior Score 1 20.358 13.985** Covariate=Age 1 .016 .011 A=Age Environment 1 .797 .548 l .068 .041 E=Independence in Activ- ities of Daily Living 1 .424 .292 1 1.849 1.117 A X E Interaction 1 .671 .461 1 1.232 .744 Residual 99 1.456 104 1.655 Total 104 1.601 107 1.638 Covariate=Prior Score 1 20,353 14_422** Covariate=Age l .016 .011 A=Age Environment 1 .485 .344 1 .009 .005 F=Financial Abilitx 1 3.867 2.739 1 4.870 3.010 71"x‘F interaction 1 1.593 1.128 1 2.072 1.281 Residual 99 1.412 104 1.618 Total 104 1.601 107 1.638 . p_< .05 it 2 < .01 95 Table 25: Analysis of Covariance and Variance for Older Sample-~(H04C, A, O). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Age A=Age Environment ' B=Residentia1 Location Data Not Reported Data Not Reported of Primary Care Giver Cell Sizes = 62 25 Cell Sizes = 62 25 A X 8 Interaction l3 5 13 5 Residual Total Covariate=Age A=Age Environment Data Not Reported Data Not Reported C=Spouse Survival Cell Sizes = 58 7 Cell Sizes = 58 7 A X C Interaction 17 23 17 23 Residual Total Covariate=Age A=Age Environment Data Not Reported Data Not Reported D=Days of Bed Disability Cell Sizes = 56 21 Cell Sizes = 56 21 A X 0 Interaction l9 9 l9 9 Residual Total Covarf t :A e 1 1.615 1.011 A=Age En?” rgnment 1 2.469 1.546 1 1.842 1.140 = n nd nce in Activ- E 1323801780311), Living 1 8.199 5.134* 1 7.005 4.334* A x E Interaction 1 .709 .444 1 .621.384 Residual 100 1.597 101 1.616 Total 104 1.668 104 1.668 Covariate=Age 1 1,230 _745 A=Age Environment 1. 3.680 2.228 1 3.216 "95‘ F=Financial Ability 1 .190 .115 1 .333 '20? A x F Interaction 1 2.034 1.231 1 2.410 "“63 Residual 99 1.652 100 1.648 Total 103 1.655 103 1.655 *p < .05 *t g < .01 mo. v d... 96 Po. v .m.. mom.P woo woo._ woF pouch moo.P MOP mmm.~ No— _ooowmmm Nmm.F mom.m P omm.~ wo_.m P peasaoew>cm mm< u < copuoo mom. mpo.— F mm< n muo_eo>ou Nooozw mmo.p so? Foo._ cop _opop mmm.~ oo— wm¢.F Poo _oonwmmm ooo. ooo. _ Noe. moo. _ peacooce>oo moo . < coooao __o. o_o. F oo< . ooeoce>oo eeoo_.eF mom.om F mcoom Lo_co n muowco>ou Nome: ooa omooo ooo._ moo “Po. 4 o. _eooe mmm.~ mop mo< ope: _oou?mmm Pom.m ouo.o _ < cooooo eo cowooocoooo oooEooLP>om mo< n < eooooo ompcoomm poz opoo mm< n moowco>ou 2:2; w:_o> mcoocm soummcm mzpo> mcoscm soommcu cowpowco> we muczom u now: eo mmmcmmo m com: mo mmmcmmo mucoweo> we moma—oc< mucowco>ou mo m_m>~oc< .mPQEom cmcFo so; ucmscocw>cm mmimucoweo>oo we mwmeoc< oozimco new mono» 97 Consistency Hypotheses. Turning to a consideration of find- ings which support the consistency hypotheses (HolC. to Ho4c.), one of the eleven reported interactions proved to be significant. We may state this finding as follows: 1. There is a significant interaction (ANOVA reported only) between factor A, age environment, and factor C, spouse survival at time #1, when the criterion is contentment at time #1 (see Table 22). Contentment is higher for persons in a nonconcentrated, demanding age environment, with a surviving spouse and for persons in a concen- trated, nondemanding age environment without a sur- viving spouse (see Table 18). This interaction is upheld for both the older sample and the total sample. No other effects supporting the hypotheses were observed in these analyses. Disengagement Hypotheses. In this older sample, three sig- nificant main effects are found. These are reported in Table 27. None of the main effects related to the disengagement hypotheses. The one-way analyses for age environment reported in Table 26 show no significant main effects. Activity Hypotheses. All three main effects support activity hypotheses. There are significant main effects for factor E, independence of AOL, when HolA., H02A., and H04A. are under test. In this older sample, three of the four activity hypotheses were supported in three of the eight sets of tests of those hypotheses. To summarize, the findings for the older sample do not give strong support to the consistency hypotheses of this study. Only one finding of 11 sets reported was a significant interaction. The 98 _o. v_o.. mo. v mr we we?» .oeo>op o_wao to _e os_e em on oo< ow om o_oa» mmwpw>wpoo xwm -o co Homocmomuco c..x ex u:m§w:mu:ou moomucmomocH m we move .oow>e_ »_woo to No meek oo oo oo< ow No o_oee mmwpw>wuuo xwm Poo cw “cavemamucfi «ex t..x HomEpcmpcou mocmucmomncH m _e meek .oo.>.~ »_ooo to .4 deep on em oo< oe Pm o_ooe mmwpw>wpuo xwm Foo cw ucmucmomocH rx pomEpcmucou mocmocmamucH m Home: cocoo: no oeoEoooeooo <>oz< <>ooz< “no "coo cooooo coweoe_co oooeeo owes ”no oomo eoez oeeo_eeoo.m mwmxpwc< cmcz .oPQEem eoo_o ow oeooa moooeto o_ez ooeootwooom "mm mpnmh 99 disengagement hypotheses fared worse, with no significant findings. Activity hypotheses yielded significant findings for three of six sets of main effects which were reported. A basic finding for this sample is that the covariate of age does not produce significant 5 values. Thus, the utility of the analysis of covariance is questionable, for this older sample. This matter will be discussed further in the following chapter. Results for the Younger (45 to less than 65) Sample Descriptive Data In this younger sample the total number of cases is smaller (125) than for the older sample (155). Because of this, we expected to be able to report fewer analyses, due to small cell frequencies. This expectation is shown to be accurate in Tables 28 to 31. None of the cell distributions for factor A, age environment; by factor 8, location of PCG; or factor A, age environment; by factor C, spouse survival, are such that analyses may be reported. For factor A, age environment, by factor 0, days of bed disability, the analyses for H02. (Table 33) and H04. (Table 35) are not reportable. For factor A, age environment, by factor E, ADL, the analyses for H02. (Table 33) are not reportable. For factor A, age environment, by factor F, financial ability, the analyses for H01. (Table 30) and H03. (Table 32) are not reportable. All results not reported here may be found in Appendix C. In general, the marginal frequencies for factor B, location of PCG; factor C, spouse survival; factor 0, days of bed disability; 100 Table 28. The 1, fl, and SD of Contentment at Time #1 for Younger Sample-~(H01C, A, D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Total FACTOR B--Time #1 Residential Location of Prima_ry_ Care Giver V7Lives in Same 44g ' 59 103 Building as the 2.909c 2.695 2.786 Participant 1.567 1.329 1.432 Lives Separate 4 9 13 from the 3.250 2.889 3.000 Participant .957 1.691 1.472 FACTOR C--Time #1 Spguse Survival Spouse 41 47 88 does 2.878 2.936 2.909 Survive 1.568 1.374 1.459 Spouse 7 21 28 does not 3.286 2.238 2.500 Survive 1.254 1.261 1.319 FACTOR D--Time #1 Days of Bed Disabilipy "no Days 15 32 47 ' of Bed 3.200 3.000 3.064 Disability 1.656 1.391 1.466 1 to 14 Days 33 36 69 of Bed 2.818 2.472 2.638 Disability 1.467 1.320 1.393 FACTOR E--Time #1 Independence in Activities of Dail Livin ndependent in 16 34 50 all 6 3.438 3.059 3.180 Activities 1.365 1.301 1.320 Dependent in 32 34 66 1 to 6 , 2.688 2.382 2.530 Activities 1.554 1.371 1.459 10141.d 48 68 116d 2.938 2.721 2.810 1.522 1.370 1.432 FACTOR Fo-Time #1 Financial Ability ——Above the 37 37 74 Median of 3.027 2.946 2.986 Financial Ability 1.518 1.508 1.503 Below the 8 29 37 Median of 2.375 2.379 2.378 Financial Ability 1.598 1.147 1.233 TOTAL 45 66 111 2.911 2.697 2.784 1.535 1.381 1.442 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. 101 Table 29.. The I, M, and SD of Contentment at Time #2 for Younger Sample--(H02C. A, D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Tota1 FACTOR B--Time #2 Residential Location of Primary Care Giver Lives in Sane 29; 49 78 Building as the 2.793c 3.102 2.987 Participant 1.634 1.571 1.591 Lives Separate 4 6 10 from the 3.500 1.833 2.500 Participant 1.732 1.329 1.650 FACTOR C--Time #2 Spouse Survival Spouse 26 42 68 does 2.808 3.119 3.000 Survive 1.698 1.596 1.630 Spouse 7 13 20 does not 3.143 2.462 2.700 Survive 1.464 1.506 1.490 FACTOR D--Time #2 Days of Bed Disability N6_Days 25 4O 65 of Bed 3.160 3.350 3.277 Disability 1.573 1.562 1.556 1 to 14 Days 8 15 23 of Bed 2.000 1.933 1.957 Disability 1.604 1.163 1.296 FACTOR E--Time #2 Independence in Activities Of Daily Livi_ng Independent in 24 4D 64 all 6 3.167 3.325 3.266 Activities 1.579 1.509 1.525 Dependent in 9 15 24 1 to 6 2.111 2.000 2.042 Activities 1.616 1.414 1.459 FACTOR F--Time #2 Financial Ability ' Above the 22 27 49 Median of 3.000 3.148 3.082 Financial Ability 1.604 1.657 1.618 Below the 11 28 39 Median of 2.636 2.766 2.744 Financial Ability 1.748 1.524 1.568 TOTAL 33 55 88 2.879 2.964 2.932 1.635 1.587 1.596 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. l()2 Table 30: The I, M, and S9 of Contentment at Time #2 f0r Younger Sample-~(H03C. A. D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Tota1 FACTOR B--Time #1 Residential Location of Primarprare Giver "Eives in Same 30: 49 79 Building as the 2.867 3.020 2.962 Participant 1.634c 1.574 1.589 Lives Separate 3 6 9 from the 3.000 2.500 2.667 Participant 2.000 1.761 1.732 FACTOR C--Time #1 Sp0use Survival Spouse 26 41 67 does 2.808 3.073 2.970 Survive 1.698 1.587 1.623 Spouse 7 14 21 does not 3.143 2.643 2.810 Survive 1.464 1.598 1.337 FACTOR D--Time #1 Days of Bed Disability No Days 11 29 40 of Bed 2.727 2.966 2.900 Disability 1.737 1.614 1.630 1 to 14 Days 22 26 48 of Bed 2.955 2.962 2.958 Disability 1.618 1.587 1.584 FACTOR E--Time #1 Independence in Activities QITQeilxceixine ndependent in 10 29 39 all 6 3.000 2.828 2.872 Activities 1.764 1.605 1.625 Dependent in 23 26 49 1 to 6 2.826 3.115 2.980 Activities 1.614 1.583 1.588 10m.d 33 55 88d 2.879 2.964 2.932 1.635 1.587 1.696 FACTOR F--Time #1 Financial Abilit Above the 25 29 54 Median of 2.920 3.138 3.037 Financial Ability 1.656 1.663 1.648 Below the B 25 33 Median of 2.750 2.760 2.758 Financial Ability 1.669 1.535 1.542 TOTAL 33 54 87 2.879 2.963 2.931 1.635 1.601 1.605 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. dThis row presents the totals for Factor A in the four tables above. p 1(13 Table 31: The f, M, and §_D_ of Contentment at Time #1 for Younger Sample-~(H04C, A, D). FACTOR A--Age Environment Concentrated Nonconcentrated Nondemanding Demanding Tota1 FACTOR B--Time #2 Residential Location of Primary Care Giver —lives in Same 29; 46 75 Building as the 3.000 2.978 2.987 Participant 1.604c 1.390 1.466 Lives Separate 6 g 6 12 from the 3.333 2.000 2.667 Participant 1.211 1.265 1.371 FACTOR C--Time #2 Spouse Survival Spouse 29 39 68 does 2.966 2.923 2.941 Survive 1.592 1.440 1.495 Spouse 6 13 19 does not 3.500 2.692 2.947 Survive 1.225 1.316 1.311 FACTOR D--Time #2 Days of Bed Disability N0 Days " 27 37 64 of Bed 3.074 3.162 3.125 Disability 1.466 1.424 1.431 1 to 14 Days 8 15 23 of Bed 3.000 2.133 2.435 Disability 1.852 1.060 1.409 FACTOR Eo-Time #2 Independence in Activities of Daily Living Independent in 25 37 62 all 6 3.320 3.189 3.242 Activites 1.314 1.411 1.363 Dependent in 10 15 25 1 to 6 2.400 2.067 2.200 Activities 1.897 1.033 1.414 FACTOR F--Time =2 Financial Ability Above the 23 25 48 deian of 3.217 3.160 3.188 Financial Ability 1.622 1.599 1.593 Below the 12 27 39 Median of 2.750 2.593 2.641 Financial Ability 1.357 1.152 1.203 TOTAL - 35 52 87 3.057 2.865 2.943 1.533 1.401 1.450 aFrequency within the cell. bMean of the criterion variable within the cell. cStandard Deviation of the criterion variable within the cell. 104 and factor E, ADL, are in the same relationship to each other as reported for the total sample and the older sample. However, the marginal frequencies for factor A, age environment, and factor F, financial ability, are distributed differently for this younger sample than for the older or total samples. For the younger sample, more persons are found to be in a nonconcentrated, demanding age environment than in a concentrated, nondemanding age environment. The opposite results were obtained for the older sample and the total sample. This finding regarding age environment held true for all four tests of the hypotheses. For the younger sample, more persons were above the total sample median for financial ability than below the total sample median, and this is true at time #1 and at time #2. In both the older and the total samples, marginal frequencies indicate that the majority of cases above the total sample median for financial ability at time #1 changes to a majority of cases below the total sample median for financial ability at time #2, approximately six months later. Assessment of the data reported as means and standard devia- tions of the criterion variable, contentment, shows that contentment is invariably lower and more variable for this sample than for the total sample. The finding for the total sample and the older sample that contentment is higher at time #2 than at time #1 is not upheld for this younger sample; and the finding that the total sample and the older sample evidenced a higher mean contentment if the spouse did not survive is reversed for the younger sample. 105 To sum up, this younger sample has more unreported analyses due to small cell frequencies than the older sample. This sample is more apt to be in a nonconcentrated, demanding age environment than is the older sample. Members of this sample are more apt to be high in financial ability at time #1 and to remain at that finan- cial position at time #2, whereas in the other two samples financial ability is lower by time #2. The contentment level of this younger sample is lower and more variable than the older sample. Lower con- tentment is associated with nonsurvival of spouse for this younger sample, while higher contentment is associated with the same event for the older sample. Analyses of Covariance and Variance In addition to the data not reported in Tables 32 to 35, due to small cell frequencies, three analyses of covariance have gone unreported due to significant interaction between a covariate and a factor. In H02. (Table 33), prior (time #1) score on content- ment interacted significantly with factor F, financial ability. In H03. (Table 34), age interacted significantly with factor C, spouse survival, and with factor E, activities of daily living. All data not reported here may be accessed in Appendix C. There are no differential findings of significant f_values between the analysis of covariance and the analysis of variance for this younger sample. Consistency Hypotheses. In this younger sample, there were no significant interaction effects to support the consistency hypotheses. Table 32: 1(36 Analysis of Covariance and Variance for Younger Sample-~(H01C. A. D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Age A=Age Environment B=Residentia1 Location Data Not Reported Data Not Reported of Primagy Care Giver Cell Sizes 44 59 Cell Sizes = 44 59 A X B Interaction 4 9 4 9 Residual Tota1 Covariate=Age A=Age Environment CeSpouse Survival Data Not Reported. . Data Not Reported 11 x c Interaction Cell Sizes = 41 47 Cell Sizes = 41 47 Residual 7 21 7 21 Total Covariate=Age 1 14.861 7.728** A=Age Environment 1 .146 .076 1 2.333 1.144 D=Days of Bed Disability 1 7.504 3.902 1 6.086 2.986 A X 0 Interaction 1 .005 .003 l .137 .067 Residual 111 1.923 112 2.038 Total 115 2.051 115 2.051 Covariate=Age 1 14.861 7.964** A=Age Environment 1 .370 .198 1 3.060 1.553 E=Independence in Activ- ities of Daily Living 1 13.825 7.409** 1 13.744 6.974“ A X E Interaction 1 .012 .006 1 .035 .018 Residual 111 1.866 112 1.971 Total 115 2.051 115 2.051 Covariate=Age A=Age Environment F=Financia1 Abiligy Data Not Reported Data Not Reported A X F Interaction C811 51285 37 37 C811 Sizes = 37 37 Residual 8 29 8 29 Total ‘p < .05 **p_< .01 Table 33: 1(17 Analysis of Covariance and Variance for Younger Sample--(H02C, A, D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source Of Variation Freedom Square Value Freedom Square Value Covariate=Prior Score Covariate=Age A=Age Environment R=Residential Location Data Not Reported Data Not Reported of Primary Care Giver Cell Sizes = 29 49 Cell Sizes = 29 49 A X 8 Interaction 4 6 4 6 Residual Tota1 Covariate=Prior Score Covariate=Age A=Age Environment Data Not Reported Data Not Reported C=Spouse Survival Cell Sizes = 26 42 Cell Sizes = 26 42 A X C Interaction 7 13 7 13 Residual Tota1 Covariate=Prior Score Covariate=Age A=Age Environment Data Not Reported Data Not Reported D=Days of Bed Disability Cell Sizes = 25 40 Cell Sizes = 25 4O A7X 0 Interaction 8 15 8 15 Residual Tota1 Covariate=Prior Score Covariate=Age A=Age Environment Data Not Reported Data Not Reported E=Independence in Activ- Cell Sizes = 24 40 Cell Sizes = 24 40 ities of Daily Living» 9 15 9 15 A X E Interaction Residual Tota1 Covariate=Prior Score Covariate=Age A=Age Environmnt Data NOT. Reported 1 .44? -170 F=Financial Ability Interaction of Factor F 1 2.775 1.066 A X F Interaction with Prior Contentment 1 ~000 .000 Residual p_= .033 84 2.603 Total 87 2.547 Table 34: 1138 Analysis of Covariance and Variance for Younger Sample-~(H03C, A, D). Analysis of Covariance Analysis of Variance Covariate=Prior Score Covariate=Age A=Age Environment F=Financial Abiligg A X F Interaction Residual Tota1 Data Not Reported Cell Sizes = 25 29 8 25 Data Not Reported Cell Sizes = 25 29 8 25 Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Prior Score Covariate=Age A=Age Environment B=Residential Location Data Not Reported Data Not Reported of Primagy Care Giver Cell Sizes = 30 49 Cell Sizes = 30 49 A X 8 Interaction 3 6 3 6 Residual Tota1 Covariate=Prior Score Covariate=Age Data Not Reported Data Not Reported A=Age Environment Cell Sizes = 26 41 Cell Sizes = 26 41 C=Spouse Survival _ 7 14 7 14 A X C Interaction‘ Interaction of Factor C Residual with Age Total 9 = .048 * Covariate=Prior Score 1 37.930 17.733’ Covariate=Age 1 1.101 .515 A=Age Environment 1 .407 .190 1 .198 .075 D=Days of Bed Disability 1 .184 .086 l .124 .047 A X 0 Interaction 1 .163 .076 1 .255 .097 Residual 78 2.139 84 2.632 Total 83 2.573 87 2.547 Covariate=Prior Score Covariate=Age Data Not Reported A=Age Environment Interaction of Factor E 1 .257 .098 E=Independence in Activ- With 996 ities of Daily Living 9_ = .003 1 .361 .138 A X E Interaction 1 .985 .376 Residual 84 2.620 Total 87 2.547 *p< .05 “p< .01 ' Table 35: 1139 Analysis of Covariance and Variance for Younger Sample-~(H04C, A. D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Age A=Age Environment B=Residential Location Data Not Reported Data Not Reported of Primary Care Giver Cell Sizes = 29 46 Cell Sizes = 29 46 A X 8 Interaction 6 6 6 6 Residual Total Covariate=Age A=Age Environment Data Not Reported Data Not Reported §§§pouse Survival Cell Sizes = 29 39 Cell Sizes = 29 39 A'X C Interaction 6 13 6 13 Residual Total Covariate=Age A=Age Environnent Data Not Reported Data Not Reported D=Days of Bed Disability? Cell Sizes = 27 37 Cell Sizes = 27 37 A X 0 Interaction 8 15 B 15 Residual Total Covari ate=Age 1 13. 603 7. 38 3“ A=Age Environment 1 .071 .038- l .746 .386 E=Indepepdence in Activ- ities of Daily Living 1 14.804 8.035** 1 19.319 9.994** A X E Interaction 1 1.202 .652 1 .176 .091 Residual 82 1.842 83 1.933 Total 86 2.101 86 2.101 Covariate=Age 1 13.503 5.9974 A=Age Eniginonrnon.t 1 .141 .073 1 .196 .094 F=Financ141 Ability 1 7.615 3.917 1 5.853 2.791 A X F Interaction 1 .053 .027 1 .049 .023 Residual 82 1.944 83 2.097 Total 86 2,101 86 2.101 *p < .05 **p_ < .01 llO Disengagement Hypotheses. N0 main effects for age environment are reported in Table 36 or Table 37. There is no support for the disengagement hypotheses. Activity_Hypotheses. Two main effects are reported for the younger sample in Table 37. Both of them are in support of the activity hypotheses. These are significant main effects for factor E, independence of ADL, when HolA. and Ho4A. are under test. In this younger sample, two of the four activity hypotheses were supported in two of the eight sets of tests of those hypotheses. To summarize, the findings for this younger sample give no support to the consistency hypotheses or the disengagement hypotheses. Activity hypotheses yielded significant findings for two of five sets of main effects which were reported. Summa:y_of the Results The results reported in this chapter are summarized in Table 38 for the interaction effects, and in Table 39 for the main effects. Table 38 illustrates very concretely the extent of unre- ported data, shown as a dash (-) in the table. In regard to testing the four consistency hypotheses, a . glance at Table 38 indicates that, in general, none of the hypotheses was strongly supported, with no support at all f0r H02. and H03. There is support for Hol.in theinteraction of factor A, age environ- ment, with factor C, spouse survival. There is support for H04. in the interaction of factor A, age environment, with factor B, resi- dential location of PCG. A check of the cell means for the three lll mo a V dfi: _.o. v more _op.~ oo _o_.~ oo _eooe ~__.~ mo ooo.F oo _eoo_moo moo. oon. _ _Po. ooo. _ eooEooc_>oo moo n o coeoea smmm.m mom.m_ F mm< u muowco>ou Nooozm Noo.m No . meo.~ mo _eeoe oeo.~ oo ooo.~ oo Poooomoo ooo. oep. _ mop. ooo. _ oooEeocosoo oo< n o eooooe moo. PoP.F _ oo< . ooe_cosoo ..oo_.oo ooo.~o . ocoom coeco . ooeeeesoo dome: ooo ooozv Poo.N o_F Foo.~ o_P _eooe Noo.~ o__ ooo. op_ .eoo_moo ooo. omo._ _ ooo. ooo. P oeoEooc_>oo moo u < eoeoee .aooo.n poo.o_ P moo u ooeeeosoo ~o_ozm wo_o> mcoocm Eoummcm mo~o> mcoocm Eoummcm :owuoeco> mo mocoom m com: mo mmmcmmo u com: eo mmmcmmo mocoweo> eo memXFoc< mocowco>oo we m_mx_oc< .mposom emmcoo> com “coacocw>cm mmimocoweo>ou mo mmmzpooq oozimco "mm mpnmh 112 S. v ime... mo. v we. we move .oee>o_ opoao to Fe meek ea oe ooo ee om o_oee mmwpw>wuoo xmm PFo cm ucmvcmomucH «ax wax acmEucmucou mocmvcoamucH m .e meek .oeesep opeeo Co Fe weep ea ea ooo ea om o_oee mmmpw>wpoo xwm FFo cw ucmncmomocH tex «ax “cmsucmuooo mocmocmomocH m Home: eoeoo: m_ oooEoooeooo <>oz< <>ooz< "no ”coo eooooa eoocoooco oooteo ewe: ”no oomo coo: oooooeoooem mwmzpoc< cos: .mPQEom emm==o> cm ooooa meooeeo oeez ooaoeeoeoeo um mpnmh 113 .o. v.mr. mo. v N! nonconag no: name a . acauoeocmoocoz a oz oucooem> yo ooox_oc< a < oucoocu>ou mo oooxpoc< n u .ouoz oz oz oz oz oz oz oz oz oz oz oz oz - - - - oz oz oz oz - - - - oz .oz oz oz oz oz oz oz 1 1 mz i mz mz i i mz mz mz mz mz m2 1 i mz mz m2 m2 m2 mz mz mz mz i . mz mz mz mz i i i i .mz mz m2 m2 mz mz ¢< tau 1 i mz mz mz mz i 1 mz 1 m: i mz i a< 1 m2 au opanm pogo» i i opoeom cmmcno» i i oPQEom Coupe 1 i mPanm —muo» i i oyasom tomcao> i i opanm cmu_o oz oz mooseo .aooe i i m_a2om cmmcao> i i mpasom cmvpo mz mz mpascm pouch i i mFQEom cmmcaoz i i m—aacm cmvpo A—o>oa ucgucoucou I. 95... 9.50.6 .9... 233... No 95: .302 :oouoaououc< pouomo_ococzu co: A—m>04 uzmsucmucou we moo» acouooumca ocouoou _‘ os_~v pouoz gumbo“ mo; Pouomo_o:oczu . no: A—m>m4 ucmsucmocou we goo» acouuoumca ocouuou Na msopv pone: pocoouomm ooocu ooze—mo No: Apm>04 “cmzucmocou _‘ ego» m:_uuoumem ocoouoo .z moopv quos pocoooumm ooocu mucoum;o_ Po: < u < u < u < u < u QFQEom u couucu m couoem o coooom o couuou on on >2 on < Louuom < Lagoon < coooou < couucm m couuom on < couuou 67‘711' I :o_ouocooc_ 11l-.lli"l I I- I9-.. Ir‘l--- ooomzooaxx .ouomeem coouuocmu:_ acoucomoa oncoucou mo zgo253m. "mm upon» 114 8. v m: mo. v 9.. meow swamp m an aco>oz x 1*x 11x «z we?» z_omo mo omooo>opuo x «*x zzx _z ms_z on om 4oz co xoo Ppm co mucmozmawccw 1x 1x oz x 11x 1oz No meop z_omo we om_po>ooum No meop pm pm zoz co xoo Fpm co acmucmamncn 11x 11x oz x «1x zox .z wszo ozone oo omopo>opum x 11x 11x oz mew» om pm 4oz co xoo Fpm co ucmucmamucH 1x <~ “cuspcmucou mucmucmamucm m _* meoh omucmwg um >H_Fonm -mzxm mzm zpo_ozm _z meoz pm -ooo umo -ooo umm oo ozmo oz x 1x <_ “cmEozmocou oo ooze a No msoz uwucmog pm xuo__nm .-mzxm wzm zoopozm No weop pm -ooo umm 1o_o umm $0 oxmo oz x «1x 1«x o> _* meop um Pm>o>gzm 1g=o “o: omen moaoam x 1x 1 acmEpcopcou monogm u "cwzz Pouch gmmczoz LmuFO <>oz< <>ouz< "om uoH “Log Lopumu gmsmoz ow acmEpcmucou o: coogmpogu pumywu com: ”or umo: "oH umo: cmcz ems: acouo$wcmwm mpnzmm can: owoxpmc< can: .ouummmm com: acougwmmm omcwucou ucmuowwcmom mo aguessm "mm mpamp 115 appropriate samples, hypotheses and factors sets indicates that all of the reported interactions which are significant are in the hypothe- sized direction. This direction is such that contentment is higher when age environment is concentrated, nondemanding and activity resources are low, or when age environment is nonconcentrated, demand- ing, and activity resources are high. The reported main effects are summarized in Table 39. This table compiles results previously tabled in Table 17, 27, and 37. Activity theory has been well supported in three of the four hypothe- ses, with most of this support involving factor E, activities of daily living, and less involvement of factor D, days of bed disabil-: ity, in support of the hypotheses. It should be noticed that these findings for activity hypotheses are of such strength that a finding of significance is obtained in both the analysis of covariance and the analysis of variance for all but one effect. Disengagement hypotheses have been rejected because of no support. The findings reported in this chapter will be discussed further in Chapter IV. CHAPTER IV DISCUSSION Introduction This chapter consists of two general discussions. The first is pointed toward the results obtained in this study. It includes a summary of the characteristics of the several samples used in the study, a comparison of the results obtained for the hypotheses which relate to the three types of theories of successful aging: consis- tency, activity, and disengagement theories, and a discussion of other obtained results. The second general discussion makes suggestions for future research regarding successful aging. These suggestions are in regard to instrumentation, data collection strategy, appropriate analysis, size and characteristics of the sample, and questions amenable to further research. Discussion of the Present Study The sample for this study was purposely tailored to produce relatively equal numbers of persons at "good" and "poor" functional levels, while allowing for the possibility of measurement of change within a six month time period. A consequence of tailoring the sample in this manner was a relatively large number of deaths during the six month time period, experienced as part of the experimental mortality. ll6 117 Because the tailoring process was designed for over- representation of persons at "poor" functional levels, the sample is certainly not representative of the older population in the commu- nities studied. Therefore, no attempt has been made to generalize the results or conclusions of this study to a larger population. Such generalization is not warranted. Since the hypotheses of this study were tested on several subsamples, it is apprOpriate to summarize the characteristics of these subsamples before proceeding to discuss the results which were obtained from the subsample data. Analysis of results have been organized on the basis of time subsamples (time #l and time #2) and age subsamples (older and younger). Although the sample of this study was formed in such a way as to allow analysis by rural and urban subsamples, the data has not been reported in this format in Chapter III and no report has been made of the char- acter of those two subsamples. Such an analysis is conceivable as the subject of a future report based on the data of this study. It may also be important, in future research, to analyze this data after division into subsamples by sex. Sample Characteristics at Time #l and Time #2 . In discussing the differences in character of the total sample at time #l and time #2, we are pointing up matters which may influence differential findings for Hol (C, A, D), which was tested on the time #l sample; and for H02 (C, A, D), H03 (C, A, D) and H04 (C, A, D), which were tested on the time #2 sample. l18 The basic difference between these two samples is that the sample size diminishes between time #l and time #2. The total sample at time #I is 280, but due to missing data, the effective sample size fOr hypothesis testing is 25l-260 at time #I. The total sample at time #2 is 20]. Loss of cases due to deletion of missing data reduces the number of cases on which the hypotheses are tested to l9l-l96 at time #2. While most of the intake characteristics of the two samples which are described in Chapter II vary little between the two time samples, it may be noticed that at time #2 the sample consists of a slightly greater proportion of persons who are married, live with only one other person, own their own home, head their own household and have a PCG who is a spouse. Since the experimental mortality which was experienced between time #l and time #2 was actual mortal- ity to a great extent, we may conclude that persons with the above characteristics at intake into the study had an advantage in "stay- ing ability" (see Table 2 and 3). Certain other differences are discernible in several of the factors which were measured at both time #1 and time #2. It is not possible from this analysis to determine whether the differences observed in these factor levels at the two times may be attributable to actual changes for the persons in the time #2 sample, or to the loss of 79 persons between time #1 and time #2. By time #21 fewer persons are (l) bed disabled for l to 14 days, (2) are dependent in activities of daily living, and (3) are above the median of financial 119 ability. Hol (C, A, D) and H03 (C, A, D) are tested using the levels of the factors at time #1; while H02 (C, A, D) and H04 (C, A, D) are tested using the levels of the factors at time #2. In regard to the characteristics of the criterion, content- ment, it was found that by time #2 the mean level of contentment is both higher and more variable. Time #1 contentment is involved in the testing of Hol (C, A, D) and H04 (C, A, D), while time #2 con- tentment is involved in the testing of H02 (C, A, D) and H03 (C, A, D). Refer to Tables 8, 9, l0, and ll regarding the factors and criterion levels. Characteristics of Older Sample and Younger Sample For purposes of analysis of the hypotheses, the size of the sample is crucial. The older sample (persons 65 or older) is larger than the younger sample (persons 45 to less than 65). At time #1 the older sample contained 155 subjects, while 109 remained in that sample at time #2. Because of loss of subjects due to missing data in the data analyses, the older sample has an effective size of 140- 144 at time #1, and l04-108 at time #2. The younger sample con- tained 125 subjects at time #l and 92 at time #2. Due to missing data in the analyses, the younger sample shrank to a size of lll-llG at time #1 and 87-88 at time #2. Along with the age difference in the two samples, several other intake variables differentiate the two samples. The older sample has a greater proportion of females, whites, widowed persons, 120 persons at higher occupational levels, and persons at lower educa- tional levels. The older sample is less likely than the younger sample to have a spouse as PCG, and more likely to have a female PCG and/or more likely to have a child or other relative as a PCG. The older sample is more likely than the younger sample to have lived in their residence l0 years or more, to have only one cohabitant, to have an "other resident" as owner or head of the household, and less likely to haveaalandlord. At least one of these differences, the finding for educational level, may be recognized as a historical bias for the two cohorts. Most of the other differences are accounted for by the particular life cycle or stage of development of members of the two age groups. The younger sample compared to the older sample is, on the average, less content and exhibits a larger standard deviation for the criterion variable, contentment. However, the mean for content- ment remains stable from time #l to time #2, and the standard devia- tion increases from time #l to time #2 for the younger sample. The older sample is more content and has a smaller standard deviation for the criterion of contentment than the younger sample. The mean of contentment increases from time #I to time #2, while the standard deviation of contentment remains stable between the two time samples for the older sample. It was noted earlier that, for the total sample, between time #1 and time #2, both the mean and the standard deviation of contentment increase. 121 Results f0r the Consistency Hypotheses We have seen that two of the four consistency hypotheses received no support, while HolC. and Ho4C. received support from only one of the five sets of factors. The two factors involved in significant interaction with age environment are factor 8, location of PCG, and factor C, spouse survival. Both these factors are variables related to persons other than the self. The two hypothe- ses with significant interaction effects are the two which have in common a lower, less variable criterion measure, contentment at time #l. Two significant findings, out of 20 sets of main effects in which it would have been possible to obtain significance, is not much better than chance. We reject H02C. and Ho3C. The two sig- nificant findings will be discussed further in comparison to the significant findings for the activity and disengagement theories. Results for the Activity, Hypotheses The activity hypotheses received support for three of the four hypotheses, in five of the eight sets of tests of the hypothe- ses. For factor D, days of bed disability, two hypotheses (HolA. and H02A.) produced a significant main effect. For factor E, inde- pendence in activities of daily living, three of the four hypotheses (all except H03A.) produced a significant main effect. Notice that both the hypothesized main effects involve factors which are varia- bles involving the self. One of the five findings was borne out in 122 the older sample as well as in the total sample. Two others were upheld in the older, younger, and total samples. Four of the five findings are supported by both analysis of variance and analysis of covariance. Although one of the hypotheses, H03A., was rejected, the important matter is that three of the four activity hypotheses were supported with significant findings. These strongly supportive results will be compared with the results from the other two theor- ies. Results for the Disengagement Hypotheses For the total sample, none of 20 sets of tests of the dis- engagement hypotheses is supported. For the older sample and younger sample, no supportive results are obtained for the disengagement hypotheses. It seems obvious that there is no support for any of the disengagement hypotheses. Comparison of the Results for the Three Theories We have seen that there is no support for disengagement theory. Opposed to this we found support for HolC. and Ho4C. related to consistency theory, and for HolA, H02A, and Ho4A. related to activity theory. See Figure l9 for a summary of these results, taken from Table 38 and Table 39. Although the results support both consistency theory and activity theory, the results reported in support of the activity theory are overwhelmingly stronger. More significant findings were 1223 Hypothesis Hol H02 H03 H04 ' Innediate Delayed Theory of Cross- Cross- Chronological Chronological Successful Sectional Sectional Lag Anticipation Aging Paradigm Paradigm Paradigm Paradigm Consistency Factor A, Factor A. Theory Age environ- age environ- (4 X 5 = 20 ment X ment x Sets of Tests) Factor C. Factor B, spouse Location of survival. PCG Total Sample Total Sample ANCOVA ANCOVA ANOVA Older Sample ANOVA Activity Factor E, Factor E. Factor E. Theory AOL AOL AOL (4 x 2 = 8 ------------------------------------- Sets of Tests) Total Sample Total Sample Total Sample ANCOVA ANCOVA ANCOVA ANOVA ANOVA ANOVA Older Sample Older Sample Older Sample ANOVA ANCOVA ANCOVA Younger Sample ANOVA ANOVA . ANCOVA Younger Sample‘ ANOVA ANCOVA _,, ANOVA . Factor D, Factor D, '__T Days of Bed Days of Bed ' Disability Disability Total Sample Total Sample 3 ANOVA ANCOVA ANOVA Disengagenent Theory (4 X 5 = 20 Sets of Tests, 4 x l = 4 One-Hay Tests) ’Figure l9. Comparison of significant findings for three theories by four hypotheses. 124 obtained for activity theory than f0r consistency theory, in both absolute and relative terms. Unlike consistency theory, the results for activity theory were supported by both ANCOVA and ANOVA in four of five significant findings. The results for one of the activity hypotheses are supported by findings from the older subsample, as well as the total sample. Two more hypotheses are supported by findings from the older, younger, and total samples. Considering all these matters, activity theory comes out the strong leader among the three theories. Although consistency theory is not rejected by these results, it is not supported as strongly as activity theory. Disengagement theory is rejected by these findings. Discussion of the Results for the Three Theories The variables utilized in this study were those chosen from among the pool of available variables from the demonstration project. There were no specific research objectives or criteria which were satisfied by this variable set. Later in this chapter, many sugges- tions will be given for alternate variables which could be utilized in future research regarding theories of successful aging. This set of variables also is not a representative set from a larger popu- lation of variables. The type of bias which may have been intro- duced to this study by the variable set will be discussed below. It is noteworthy (see Figure l9) that the two variables which have lent support to the consistency hypotheses are residential location of the primary care giver, and spouse survival, both in interaction with age environment. All three of these interacting 125 factors represent variables having to do with persons other than the subject (self): (l) the primary care giver, (2) the spouse, and (3) the cohabitants of the subject's residence. Factors such as: (1) days of bed disability, (2) activities of daily living, and (3) financial ability, all have a strong "self component" and none of them contributed to support of the consistency hypotheses. The activity hypotheses, on the other hand, have been sup- ported by two of the "self component" variables: (I) activities of daily living, and (2) days of bed disability. The deficiency of the financial ability variable to measure significance when it does, in fact, exist will be discussed later in this chapter. Based on these observations from this limited set of "self component" and "other component" variables, we put forward the following two generalizations for future testing: l. Variables which depend mainly on characteristics of the self and which are little influenced by the actions of others will be related to successful aging through activity theory. 2. Variables which depend mainly on characteristics of others, especially significant others, and which are little influenced by the actions of the self will be related to successful aging through consistency theory. Suggestions for conducting research about these two proposi- tions will be presented later in this chapter. Our attention should also be directed to another character- istic of the variables which provided support for activity theory and did not provide support for consistency theory. In addition to the Specification of days of bed disability and independence in l26 activities of daily living as self component variables, we must also notice that they are of such a basic nature, and such a usually simple level as to be expected activities, regardless of the type of age environment impinging upon the individual. It appears quite possible that the two activity variables utilized in this study will not differentiate between the two age environment levels. If this is the case, we would not be able to obtain interaction effects. This is exactly the picture presented by the data. Limitation of activity resources in this study to activity variables which do not differentiate between age environment levels, may have lessened the probability of support for the con- sistency hypotheses, while increasing the probability of support for the activity hypotheses. Several alternate activity resource variables could be sug- gested which are of a less basic nature, or which lie in a middle position on a scale of activities from simple to complex. Among others these include: walking, climbing stairs, getting about out- side of a building, traveling, using a telephone, writing corre- spondence, driving a car, earning small amounts of money. While such variables are complex, they define or facilitate a social activity. The simpler activities of daily living, and staying out of bed for a good part of the day do not. Rather, these simpler activities are more readily described as pure physical activities. We should recall that these two variables of physical activity were selected hiorder to operationalize an underlying concept of 127 physical health. In saying that the interaction effects posited by consistency theory are more likely found in variables which represent social activity, rather than pure physical activity, we introduce the possibility that an activity resource which we shall label as "social health" is a more appropriate medium in which to detect sig- nificant consistency (interaction) effects. At the same time, we may say that physical health is a concept which may be expected to produce main effects of the activity theory type. Along the same lines, the rather meager support for disen- gagement theory may be due to the dearth of psychological variables in the present data. Recognizing that the present discussion of the relationship of the three theories to types of variables is speculative, incom- plete and lacking structure, we have attempted to produce a more complete structure of the theories and variables. Figure 20 leads us to the folloWing statements: l. Variables which represent physical characteristics of an individual and which may be thought of as con- tributing to physical health will be related to successful aging through activity theory. 2. Variables which represent social characteristics of an individual and which may be thought of as con- tributing to social health will be related to successful aging through consistency theory. 3. Variables which represent psychological character- istics of an individual and which may be thought of as contributing to psychological health will be related to successful aging through disengage- nent theory. Many of the variables in this study have called for major assumptions to be made about the relation of the individual 128 Type of Health (Activity Resource) Type of Type of Type of Type of Theory Activity Strength Support Activity Theory Physical Physical Physical Activities Speed Mechanical aid of Daily Endurance Human aid Living Lifting Days of Bed Gripping Disability Pulling Pushing Reaction time Consistency Theory Social .Social Social Ambulation Financial Friends and Traveling solvency neighbors Telephoning Personal Friendly Writing attractive- relatives Driving ness Escorts Earning Social Co-habitants Buying awareness PCG's Housekeeping Education Employers or Repairing buyers Creating Merchants Service workers Disengagement Theory Psychological Psychological Psychological Remembering Intelligence Spouse Memorizing Affect Significant Reasoning Verbal others Computing response Listeners Expressing speed Advisors Appropriate Size of Confidants emoting vocabulary Clinicians Learning Psychiatrists Orientating Teachers Recognizing Figure 20. Types of appropriate variables for type of theory by type of health. 129 variables to the particular theory being tested. In some cases these assumptions may not be wholly defensible. The age environment variable used to operationalize the consistency and disengagement theories has called for the most far-reaching assumptions. The first such assumption is that all age environments clas- sified as concentrated and nondemanding are classified validly. This may be questioned on the basis that some persons, whose age environments are classified in this way, may be members of a two- generation family, by living either with very elderly parents or with adult children over age 45. Given such an anomaly, the nondemanding classification would be just as inaccurate as the concentrated classification. The next assumption about the age environment variable is that it is age environment within an individual's residence which is of greatest importance in either the consistency or disengagement theory. For the person old enough to be retired, who does not get I out of the house much, this may be an acceptable assumption. For the person who is retired and still gets about outside the residence for considerable amounts of time, the age environment of friends and neighbors may be of more importance. For the person of 45 to 65 years of age, who is not yet retired, the more important age environ- ment may be at the place of work. Finally, it is questionable that living in a nondemanding concentrated age environment is equivalent to being in a state of disengagement, and that living in a demanding, nonconcentrated age environment is equivalent to being in a state of engagement. Age 130 environment is not a psychological activity or strength. It is con- ceivable that persons who live in a concentrated, nondemanding age environment receive more psychological support than do persons who live in a nonconcentrated, demanding age environment. This psycho— logical support, however, will not be perfectly coincident with disengagement or engagement of the individual, by which terms we intend to describe the polar opposites of a psychological state. This set of assumptions regarding age environment should be kept in mind when evaluating the validity of the consistency and disengagement results obtained in this study. They are especially important, since age environment is a variable involved in every analysis performed. The use of age environment to operationalize social context in this study has lessened the probability of support for the disengagement theory and, perhaps, for the consistency theory. We have already discussed the particular choice of variables which may have worked to the benefit of the activity hypotheses. To summarize: we suggest the operation of each of the theories of aging is correlated with one of three types of "health." Operation of activity theory is associated with physical health, which consists of physical activity, physical strength and physical support. Consistency theory operates when matters of social health are considered such as social activity, social strength, and social support. Disengagement theory is expected to be in operation when it is important to consider psychological health, through such vari- ables as psychological activities, psychological strengths, and psychological supports. 131 Choice of variable type (see Figure 20) may be crucial to the success of testing each theory. The poor showing of disengagement theory may be due entirely to choice of a social variable to Opera- tionalize an essentially psychological theory. The Additional Finding of a Main Effect for Spouse Nonsurvival Spouse survival has provided us with one main effect in the direction of high contentment associated with nonsurvival of spouse. This effect was not anticipated by the hypotheses of the study. The finding is significant in the chronological anticipation paradigm for the total sample (see Table 39). Although this main effect for nonsurvival of spouse was not anticipated, here are three possible explanations of the effect: 1. Nonsurvival of spouse is a condition which facilitates future increased disengagement of the widowed person, while spouse survival facilitates future continued engagement which is not self-motivated. 2. Nonsurvival of spouse is a condition which facilitates future increased activity of the widowed person, while spouse survival facilitates future continued inactivity which is not self-motivated. 3. Nonsurvival of spouse bears no relation to either future activity or disengagement. It is related to a recovery to normal levels (with a possible recovery spike) of contentment in anticipation of release from the worries and depressed contentment levels experienced during the episode of the spouse's dying. Suggestions for research designed to compare these three alternate hypotheses will be presented later in this chapter. It is possible that survival of spouse may operate differ- ently for the two sexes. If females become more content in 132 conjunction with anticipated nonsurvival of spouse, while males show no change in contentment or become less content in conjunction with anticipated spouse nonsurvival; the preponderance of females in the sample for this study could then account for this unexpected main effect in the results. Since both responsibility and dependency shifts from the locus of the work place to the locus of the home following retirement age: males may become increasingly dependent and females increasingly responsible, especially if the division of labor and role structure of the marriage has been of a traditional type. Thus prospective loss of a spouse for a female would mean loss of a responsibility which may lead to higher contentment, while the prospect of losing a spouse for a male would mean loss of a support which may lead to lower contentment. While the £_values are not significant, the main effect of spouse survival is in opposite directions for the older and younger sample. For the older sample, spouse survival is associated with lower mean values of contentment, while in the younger sample, spouse survival is associated with higher mean values of contentment. Discussion of Suggestions for Future Research Some of the general approaches to furture research which may be useful are to: (l) broaden the list of variables which are used to support the theories, (2) narrow the definition of the presently useful variables, or (3) test further researchable ques- tions regarding the circumstances under which one or the other of 133 the theories will be operant. More will be said about researchable questions following some specific suggestions for future research methodology. Suggestions for Future Instrumentation In the present study, the variables used as factors have been reduced to dichotomous data from instruments with longer scales. This was done so that the problem of small frequency in individual cells would be reduced to a minimum. In future studies, it would be desirable to allow all scales to assume their full-range. For’factors which appear to be true dichotomies, such as spouse survival, and residential location of PCG, it is possible to extend the range by adding the variable of time or distance to the scale. Such a change in the instruments would overcome ceiling and floor effects of the present measurement scales and allow the use of information from cases at the extremities of the scales. Such a change may work to specify the results of future research in greater detail. Of all the instruments used in this study, the results from one, financial ability, were least useful and most disappointing. None of the main effects or interaction effects related to this instrument were significant. This seems unusual because the types of information from which the financial ability scale is constructed have been strongly related to contentment by past research. Table l reported results of past research as follows: (l) employment related to contentment in nine studies, (2) income level related to content- ment in four studies, and, (3) home ownership related to contentment 134 in one study. Financial ability is composed of family income level and the index of economic dependence. Two of the component parts of the index of economic dependence are employment and home ownership. We suspect the substitution of large numbers of median values for missing values of family income and index of economic dependence has lessened the variability of the financial ability instrument to the point where it is impossible to show a relationship of this instrument to contentment, if such a relationship exists. Not only is this particular measure of financial ability not useful, but the practice of substituting a measure of central tendency for sizeable amounts of data in any measurement instrument should be avoided in future research. The effects we are attempting to capture in this area of research are not strong, in any event. Therefore, we should take special care to avoid all practices which weaken the ability of measurement instruments to detect relationships. The criterion variable, contentment, has been measured by a five item scale which was assumed to be an adequate measure of "general satisfaction." The five items which comprise this scale include only two items which seem to reflect satisfaction of a "general" nature. (See Figure BlO, Appendix B.) These two items are: "Would you say that you have been happy or unhappy during these past two weeks?" and "In general, how satisfied are you with your way of life today? Are you satisfied, partly satisfied, or dissatisfied?" The remaining three items include references to specific attributes of an individual's life. These items are obviously not 135 all inclusive of matters which contribute to an individual's general satisfaction. They include health, age, and arrangements for house- hold maintenance as their subject matter. These three items lessen the face validity of the five item scale as a measure of general satisfaction. However, it should be pointed out that the developers of the contentment scale had culled these five items from an originally much larger set. They employed item analysis to find those items which correlated best with the total score on a longer "contentment" measure. Such a procedure should have worked to extract the most generalizable items from the longer set of items. It is important that these three scale items may be viewed as attitude items, specific to several of the factors used as inde- pendent variables in this study. Without such artifactual dependence of the criterion measure, it may be impossible to show any effects on the criterion. The item on worry about health may be directly dependent on the level of activities of daily living and days of bed disability. The item on satisfaction with household maintenance chores may be linked to factors which affect such chores, such as the location of PCG, survival of spouse, and age environment. The item on "things just keep getting worse for me as I get older" may be related to the age environment, and spouse survival variables, both of which can be expected to be well correlated with age. It is worth raising the question of whether a different measure of general satisfaction, one which did not contain itens related to the independent factors, could show any significant 136 effects for the theoretically grounded hypotheses. This question ought to be addressed in future research. A beginning could be made by determining the effect of the factors of this study on each of the five items considered individually. For the present, it is question- able whether the criterion measure of contentment may be considered to have face validity for the construct of "general satisfaction." One final.suggestion regarding instrumentation remains. Stronger effects may result if subjective rather than objective data is gathered., The age environment variable would become subjective instrumentation if each person were asked the extent to which all the persons who lived with them in their residence were like/unlike them- selves. This similarity could be based on age or on a more global perception by each subject. The age covariate could become data based on a self-perceived age, scaled from young to old. In that case we would expect there to be little differentiation in self-perceived age for persons chrono- logically older than 65. We base this expectation on the fact that the objective age covariate produced very small nonsignificant E values for persons aged 65 and older. That would indicate that once a person reaches a certain threshold of chronological age, addi- tional years of age become meaningless. One final candidate for subjective instrumentation in the data set of this study would be financial ability. A request for subjective judgment about level of financial ability would, perhaps, have avoided some of the problems with missing data that plagued the objective measurement of this variable. 137 The confidence that subjective instruments will allow stronger effects to be obtained has a basis in the reality of the situation. It is the subjective perception of a variable's value upon which the subject depends during the internal process of adjustment of content- ment level. The objective measure of the same variable cannot show such strong effects as the subjective measure because it is less directly related to the outcome level of contentment. See Table l for findings from earlier research of the relationship of self- perceived variables and satisfaction. Suggestions for Data Collection Strategies In order to obtain funding for any future research which calls for additional data collection on very sizeable samples of the elderly, it will probably be necessary to conduct such research as ancillary to funded data collection f0r applied purposes. The fact that the elderly are not available to serve as subjects in large captive institutionalized audiences, such as schools, simply adds to the expense of gathering data in this research area. Additional communication difficulties, and a slower pace of data gathering with the elderly adds still more to interviewing costs. Data gathering of such an expensive nature must be well funded, and the realities of the present funding picture are that adequate funding will most likely be available through research which tests the usefulness of some manner of social reform. One strategy which should be adopted for doing theoretical research within such a research setting is to remove from consideration all data gathered 138 from those subjects who receive the experimental treatment. Such an approach has been adopted in the present study. One further step, which was not implemented optimally in the present study, is the formulation of research design, data gathering, and instrumentation procedures which assigns the needs of the theor- etical research nearly equal priority with the needs of the practical research. This should be done through straightforward discussion and agreement with the principal investigator and with the agents of the funding agency. The strategy is one of nondetraction from the stated research goals, for which funding was obtained; while accommo- dating to the additional goals of the theoretical research. Accom- plishment of such theoretical goals in conjunction with demonstration of social reform A may allow more insightful planning of additional social reforms, B, C, etc., at a later time, while also adding to the store of knowledge about human behavior. Suggestions for Analysis of Future Research For purposes of future research, it is important to consider the appropriateness of the two analysis methods which have been employed in this study, analysis of covariance and analysis of vari- ance, as well as the appr0priateness of the specific covariates drawn upon for the analysis of covariance. The f_values for the covariates in this study were significant at p_ < .05 for all covariates except the age covariate for the subsample of persons who are 65 years of age or older. Thus, the covariate of age appears to be an inappropriate choice for the analysis of data from this older sample. 139 Age is the only covariate used in the analysis of covariance for the older sample in the paradigms for Hol (C, A, D) and H04 (C, A, D). Therefore, it appears that analysis of covariance is an inappropriate procedure for those two hypotheses within the older sample. It may be that some other variable, such as socio- economic status, would be a more appropriate covariate than age for persons 65 and older. A test of that possibility remains to be carried out in future research. Suggestions for Sampling in Future Research In general, while the total sample in the present study was of an adequate size, the younger sample was too small to produce adequate cell frequencies for Uwaanalysis of variance, and nearly too small to allow detection of significant effects if they did, in fact, exist. The slightly larger older sample came closer to an adequate size. Any increase in the scale of measurement will demand a larger sample in order to afford a fairly continuous distribution along the scale. Since the standard deviation of contentment measured at time #1 is also lower than when the measure is made at time #2, hypotheses using the later criterion (H02. and H03.) require larger samples. The method of sampling is important not only for production of a certain sample size, but also for production of a sample of specified characteristics. In the present study, the sample was 140 chosen in the manner it was to produce fairly equivalent frequencies in the subclassifications of the factors. The marginal distributions in Tables 8, 9, l0, and ll indicate that this goal was accomplished with a fair degree of success for all factors except residential location of PCG. This sample was also designed so that it would be possible to discern change in the factor levels within a six month time span. We have been fairly successful in accomplishing this goal, for the factors of days of bed disability, independence in ADL and financial ability. The criterion of contentment also evidenced change within a six month period. Because this sample was designed to accomplish the two goals of (l) equivalent frequency of "good" and "poor" classes of activity resources and (2) observability of change within a relatively short time span, an attendant difficulty has been designed into the sample. Due to heightened frequency in "poor" categories of functioning and a heightened rate of change in the sample, as compared with a repre- sentative sample of persons 45 and older, this design also produced an experimental mortality of 28.2% within six months. This experi— mental mortality was largely real mortality which may be related. positively to rate of change and negatively to physical functioning. In future research, if similar samples are tailored, mortal; ity must be expected to remain a problem to be dealt with. One suggestion is to ascertain the entry level characteristics of persons lost from the sample, so that the effects of mortality may be separated from the actual change occurring over time. 141 Second, mortality effects different factors and different parameter of the criterion, contentment, in the older sample than it does in the younger sample. Because of this, it appears that these two samples are from two different populations. Future research should proceed to use larger samples of each age group, so that valid analyses may be carried out on each age group separately. Of course, the mortality itself should be planned for so that samples to be used in anaysis of data from a time subsequent to mortality of the sample will maintain large enough size to be useful. It will be important to tailor samples as above if short time-span longitudinal research is part of the research design. In longitudinal designs with a time-span of two years or longer, a repre- sentative sample drawn from the normal aged population would be acceptable. Such a sanple would also allow for generalization of the results to a larger population, and for a decrease in loss of sample over time due to mortality. Future Research Questions Two immediate extensions of the present study, which have already been mentined, are the analysis of the data for an urban and rural subsample, and for male and female subsamples. The latter may be of special importance when the spouse survival/nonsurvival varia- ble is involved, since the effect of this variable may differentiate on the basis of sex. However, using the present data, we can expect that the frequency in several cells of the analysis will be so small as to make the data unreportable, as is true of the age subsamples. A larger subsample is really necessary for optimal analysis. 142 A second immediate extension of the present study would be a test of the older sample under analysis of covariance, to assess the behavior of Hollinghead's index of socio-economic position as a covariate to replace age, in the older sample. Main Effects for Spouse Nonsurvival. Three alternate hypotheses were suggested earlier as possible explanations of the significant main effect for spouse nonsurvival. In order to better understand how this main effect relates to activity and disengage- ment theories, the analyses of variance and covariance should be run on subsamples of: (l) married persons, and (2) widowed persons. In this approach, all data from persons who are divorced, separated, or were never married would not be used. It would also be efficient to discontinue use of the analyses of factor A, age environment, crossed with factor C, spouse survival, in this approach to the data. Since the subsample of widowed persons is proportionally quite small, it would certainly be important to add more cases to the total sample before attempting this work. In the study which we have suggested of a married subsample and a widowed subsample, better support for disengagement hypotheses among the widowed than among the married would lend support to the first hypothesis (nonsurvival facilitates disengagement). Better support for activity hypotheses among the widowed than the married would lead to support of the second hypothesis (nonsurvival facili- tates activity). Little difference between the support given to activity hypotheses or disengagement hypotheses between the married 143 subsample and the widowed subsample would give indirect support to the last hypothesis (nonsurvival unrelated to activity or disengage- went). Type of Theogy_Related to Type of Variable. Earlier in this chapter it was proposed that "self component" variables will provide evidence to support activity theory, while "other component" varia- bles will provide evidence to support consistency theory. If these propositions are to be tested, one would wish to gather additional data, adding to the data set specific self component variables and specific other component variables which are different from the present set along some additional dimension. Since both the other component variables used in this study appear to be related to other persons who may be classified as part of the subject's social support system, it is advisable to capture additional data on variables related to other persons who may be classified as part of the subject's social responsibility system. A few examples of such variables are: (l) single children living at home vs. no single children living at home, (2) alive, never married children vs. no alive, never married children, (3) supervises other persons in paid or volunteer work vs. does not supervise others in paid or volunteer work, (4) clients, customers, or students relate to the subject vs. no clients, customers, or students relate to the subject. Regarding the variable of age environment, additional data should be gathered on the age environment of the neighborhood, and 144 the age environment in the work place or volunteer place of the sub- ject, as well as the data which has been analyzed on the age environ- ment of the subject's cohabitants. Because both of the self component variables used in this study relate to minimum levels of activity (being out of bed during the day and carrying on usual self-maintenance activity), it is advisable to gather additional activity data which increases the range of the activity scale, and/or differentiates it into several types of activity, such as social interaction in general, economic activity, educational activity, household maintenance activity, recreational activity. The purpose of such additional data gather- ing and analysis is to determine whether the two general proposi- tions (self vs. other variables), which have been stated previously, will hold up under these diverse conditions, or whether the type of theory supported is related to the classification of physical, social, psychological variable, as suggested in Figure 20. Relationship of Consistency Theory, Activity Theory, and Disengagement Theory. Finally, if one assumes that both consistency theory with its attendant interaction effects and activity theory and disengagement theory with attendant main effects are operant within the same individual, it will be necessary to determine whether they Operate concurrently or in sequence. If it can be shown that they operate in sequence to each other, it will be neces- sary to ask questions such as: (l) Does the sequence of operation of the three theories depend on the chronological age of the subject? 145 (2) Does the sequence of operation of the three theories depend on the life span stage (married, widowed) of the subject? (3) Does the sequence of operation of the two theories depend on the cultural and ecological setting (rural, urban, male, female, etc.) of the subject? If it can be shown that the theories operate concurrently within the same individual, it will be necessary to ask another set of questions such as: (l) Is the effect of the three theories on the level of contentment multiplicative? (2) Is the effect of the three theories on the level of contentment additive? (3) Is the effect of the three theories on the level of contentment determined by whichever effect is greatest? (4) Is the effect of the three theories on the level of contentment determined by whichever theory creates the highest (or lowest) level of contentment? Suggestions for Future Research Strategy It is always tempting to suggest true experimental designs, in order that the obtained results of research will be relatively more definitive. However, the types of data dealt with in this study do not lend themselves to manipulation without some major breech in current research ethics. A suggestion of true experimenta- tion brings to the mind scenes such as an individual randomly assigned to have a primary care giver housed in the same building or a different building, or an individual being randomly assigned to stay in bed 24 hours a day or out of bed for at least l2 hours each day. Such possibilities are not within the realm of propriety or simple common sense, for a noninstitutionalized sample. 146 Eventual use of true experiments may be possible, if the range of variables related to the theories is broadened to include variables which are more trivial, and therefore may be manipulated reasonably and ethically. Until such changes occur in the content of the researchable variables, field studies will be the most elegant tool available in this area. As long as the field study approach is followed, there is no good method to isolate the effects of measured- variables from the effects of contaminating variables which occur in combination with the studied variable in a natural setting. Summary This study has shown no support for disengagement theory hypotheses; has shown some support for consistency theory hypotheses, which are based upon suggestions made by Gubrium; and greatest sup- port for activity theory hypotheses. We have discussed various sug- gestions for future research which are intended to specify the basis of successful aging more thoroughly. These include: suggestions for instrumentation, data collection strategy, analysis, sampling, research strategy, and research questions. The discussion has pointed up the possibility that the results obtained in this study are highly related to the types of variables chosen for measurement in the study. Thus, these results may be misleading and should be used with caution. A major suggestion has been made that there is a relationship between the type of "health" considered to be of highest priority by an individual and the aging style adopted. It was suggested that 147 physical health is tied to activity theory, social health tied to consistency theory, and psychological health tied to disengagement theory. It remains for future research to lend credence to this new approach to theory. Two additional findings should be mentioned. The covariate of age appears to be neutralized for the older sample (65 and older). This suggests a curvilinear relationship between age and content- ment. It was also found that for the total sample there is a main effect which may be represented as a relation between anticipated nonsurvival of spouse and high contentment. Since nonsurvival of spouse was predefined as a low activity resource, this was unexpected. While not significant the effect is opposite for the older and younger samples-~survival of spouse is tied to high contentment in the younger sample, and survival of spouse is tied to lower content- ment in the older sample. APPENDICES 148 APPENDIX A COMPARISON OF SEVERAL MEASURES OF SATISFACTION/MORALE 149 APPENDIX A COMPARISON OF SEVERAL MEASURES OF SATISFACTION/MORALE Carp (1969) has stated the generalized problems created by poor attention to comparable measurement in research on elderly samples: A wasteful and dangerous tendency has been noted for ger- ontological research to be diffuse and incoordinate and for its findings to be repetitive or contradictory. Some confusion in the gerontological literature may result from insufficient attention to the nature of criterion varia- bles. In gerontological studies in the behavioral sciences, important criterion concepts tend to be imprecisely defined. This tendency allows for the possibility that different studies deal with different traits under the same name and with the same trait under different names. In addition, important gerontological criterion concepts tend to be global, which provides opportunity for elements to be variously included and weighed in criterion measures for different studies. This variable inclusion of elements may be particularly obscurantist when components do not all operate according to the same rules (p. 341). The special characteristics of satisfaction indices are pointed up by Bloom (1975) as a combination of curse and blessing on the head of the researcher: The various quality of life terms must rank among the soft- est, most personal or existential concepts which research- ers have ever tried to capture in operational form. Yet, there are few concepts more important, especially for the practitioner who asks questions of this general form: "Intervention (treatment) to obtain what objectives?" Any answer to this type of question presumes some benefit to the client's quality of life. Interest in accountability of the helping professions has also put great pressure on the development of concepts and measures of quality of life (p. 2). 150 151 Several criterion measures of general satisfaction/morale are listed in the order of their length in Table Al. Of the meas- ures listed, there is a particularly strong relationship between the 3rd, 4th, and 9th. The LSI-A, LSI-B, and LSR were developed as equivalent measures of life.satisfaction, and all were designed to contain five components of the variable: zest, resolution and forti- tude, congruence between desired and achieved goals, positive self- concept, and mood tone. The LSI-Z is a shortened form of the LSI-A. Of the four, only the LSR is based upon raters' observations, which were made following a depth interview of one to four hours duration. Interjudge Reliability Interjudge reliability of less than I point of disagreement was obtained in 95% of lOO LSR judgments (Wood, Wylie, & Sheafor, I969) and in 94% of 885 LSR judgments interjudge reliability coeffi- cient = .78) (Havighurst, l963). Wood, Wylie, and Sheafor also report the interjudge reliability percentage for each component of the LSR as: (l) zest = 83%, (2) resolution and fortitude = lOO%, and (3) the three remaining components = 97% each. Concurrent Validity The following concurrent test validity correlations are reported: Wood, Wylie, and Sheafor report §_= .56 between LSI-A and LSR; and y_= .57 between LSI-Z and LSR. Havighurst (1961, 1963) reports §_= .58 between LSI-A and LSR; y_= .7l between LSI-B and LSR;‘§.= .73 between LSI-A and LSI-B; and y.= .62 between combined LSI-A/LSI-B and LSR. Adams (l969) reports g_= .93 between the LSI-A and the LSI-Z. 152 TABLE Al: List of Various Measures of Satisfaction/Morale. Name of Measure N of Items Source 1. Satisfaction 56 Cavan, Burgess, Havighurst, and Goldhamer (1949) 2. Morale 24 Lawton (l972) 3. Life Satisfaction 20 Neugarten, Havighurst, and Index--A (LSI-A) Tobin (l96l) 4. Life Satisfaction l3 Adams (l969) Index--Z (LSI-Z) 5. Life Satisfaction l2 Neugarten, Havighurst, and Index--B (LSI-B) Tobin (l96l) 6. Morale (Adjustment) 7 Kutner (l956) 7. Age-Associated Morale 7 Thompson (1973) 8. Contentment 5 Blenkner, Bloom, and Weber (1964) 9. Life Satisfaction 5 Neugarten, Navighurst. and Ratings (LSR) Tobin (l961) lO. Morale 4 Cumming and Henry (l96l) ll. Happiness 3 Youmans (1974) l2. Life Satisfaction l Spreitzer and Snyder (refers to happiness) (1974) 153 At the same time, Bloom (l975) has reported a set of very low intercorrelations between four measures of quality of life: §_= .20 between the LSR and morale, g_= —.23 between the LSR and contentment, §_= .29 between morale and contentment, [ = .40 between the LSR and despair (Sherwood & Nadelson, l972), §_= .62 between morale and despair, §_= .ll between contentment and despair. Such low Spearman rank order correlations resulted from a pilot study utilizing an indeterminately small number of subjects. In all of the concurrent validity data reported above, there has been no attempt to assure independence of the criterion measure, in the sense that the same interviewer probably administered each test. This type of bias contamination is especially crucial in a test like the LSR, where the judgment of the interviewer is the criterion measurement. The following are reports of a test of con- current validity based upon independent judgment: Blenkner, Bloom, and Weber (l964) report a validity of .69 (Pearsonian 3) between the contentment scale and a five-place scale of judged contentment made by noninterviewers; Lawton (l972) reports a correlation of .47 between the morale scale and a judged criterion, and a correlation of .57 between morale and the LSR rating scale used by non- interviewers. Table A2 summarizes the concurrent test correlation data reported above. Item Analysis Item analysis of the LSI-A was carried out by both Adams and Wood, Wylie and Shaefor. Although Wood, Wylie, and Sheafor 154 am. no. cooemuocu ammuzn .pcmucmamucH FF. mm. ow. ewmaome mm. mm.1 “cospcmpcou om. mm. meLoz No. No. _N. Wm” zoo m\<1Hm4 mm. N1Hm4 me. m1Hm4 <1Hm4 Lomaomo ucmEacmucoo womeoz mm; 1wmx 1HM4 1Hmz 1H“; .mpmeoz\cowuummomumm eo omgzmmmz Fmem>mm Low mcowumpmgeoogmch om< m4moo .N A.uuo .ouoscoo zuouopoe .oopuo~__ou_oooz oo zoom .nooooug sozuo so» oououoooo on mos oooooov ooooco: .— an aunt a: on oooscoe cosoz Aoozoooz so oooco>oo momma oopo bozo nu oozoooz oocoooooeo ooze» ooocaoz "ooocv oooozzzz zozoz no» ozzz zo ~1 ooooooooo to ooocosoo ozoooo .zzppooooocoozov oozzz oo . o1 ooooooo. oozoov zoopo “oooouoooa moopoa zoo «o: oo o>Loooo 1 oozo_>gouo~v am mu Mouaoaoo— Loguov mu_z3 1 Afiovomoomoop »»_oooov om» .N on Azooooomv zoo—m apoo zoo—mom ozoooo 1 oz .— mu Azooooomv ou_zz oou>_»z< “ooze o_ .oo _1 zz_=o zoo—memo zoa_o ”oscoo zoozcouoo o. 1 o. co oooow oogooo ouoo "zoozoouoz op ooo o_ co» ooou pooooooo .m menu—nu couosoe< o>_uoo oo: .og=o_:o mcoxoooo zoo—mcm 1 ou_z3 .m xppoouoooz coxooo omooacop cmooco» omzuo 1 moon: .e wzppoou_ooz coxooo ooomoucoo so goocoom.wwmmmmomv ooze—:0 o_:ooo_: 1 ou_z3 .m oc:u_oo ooo_goe< osouoo 1 ooozz .N on ouoo oz zoa_o ._ N1 zo to .o .e .o .Nv oo_zz Aoooom ooozoo o>coooov p1 A_v zoo—m "oeLoo zoo>cooo_ o_ 1 o. oo ooooo ooeooo noon "3o_>couo_ up use on go; ooou zoo—m .N on ouoo oz oo_zz ._ N1 z_v ooozz Aoooem ooozoo o>coooov uueooo~ o_ co» ooou ooozuu spu _ «one zo_.o_.op.zpv N 1 z o eooz .~ omoo Aopv oooooooom uoooooouooo ooooo ootoa - ooozo .oeuzoo mono P 1 o oo 2o oo~ co o moo; c «Eoz Lasso: oucoom.ouoo mu a cu _ opooocoz o—oopgoz 16C) 161 .opoo»oo» ooouooouoo uoozoo: .~m noon—m .a—ouoeoxogooo .oooozv 5§5§§E2§§2£§8§§gzé samph_uv auh<004 h~ m<3 umuzx .mo ozozzz »zoo o» oozoFQEo ooooooo cozy zozoo» zoozo o»oommm so .coeogo» o Lo»o no: oooooa »_ .Fou_oo_o so vo___zoooooo .vo——»xo1»5oo .oop—ozo so» zuozuv Aooomo_o5o _ouomzooo>om Lozuo coo .ooossoo zoouo»_e .oozopo5o1upoo .ooooocoe goo ozooeoooo »__oooooooo ozuz»zza zo .o<_o_oao .ooooo so» .zuu~uuo no oz_x >z< =o> may: «mop—h < m><= no» c_o .mo no.3oP—o» ozu oo ooao» o» o_ooo gozoco opooooo oz .v o»:» .ooo»ooooooo zopooopo ou oozopezo1mzom .n ooo: on: oFooo voogooo___o: oz» Aocoouogooeou u_»oco1ooo oco o_oozoo ouozooov oooo_oom .N .oogoo_oooo oo—o o» oooo»ooooo . ooosocozoo .— oooz zoo oooooooocoo ozo oozo 3.3.1:... 1.1.... an.” .M...z._._.._.zs “new. ossooo uoo oov Aoouozeoo zoooo—oe Lo .oo_u»ooo>_o: — z z .z z __ v . . . oozgox uooo_o_ugoo oz» Pozop gun: on ououo ooo o_oozoo o»_ooo on »_oz oo .ooo_ ooo xuooou .ooo ucooouoaooo ozo zoo: uo cozy .ooooo ._oeoooo oooo_oo_ oooooaosooz ooo»ooazo-aooo oo» oozx »o ooooto o» ocooo zo_ooooooo mam: «o .mmuz_o=m uh<>~zo .hzwzzzmzou :»_2 use mzzp moz zoe ooeoz o .ouoouo:» Lou .oowsoo oE»u o1 pouoozuo» .oopom .pouoeo—u noomoo— on» so» «zoo zoo: oz» »—_Loooooo= uoo o» o»zwv mu pooooogoo o>ouoooo»:»so< oooogoo ammo—gem »_»uo ooo oog»uoc once so» ago: yo - o—ooo_ooooooo gooooo o u voo— any on apoooooo p—o: wonky nzuu¢ououoxu eono»: coouooouuc ~¢~ ousoow ouoo cozuauv»ouuc~ 05oz Lanes: o»oo»oo> o—oooooz 162 .opn-rsa> acmacwaauo pacovunuavu .na moan—u Loxooo opoooou oz .z z: ooooozo ooouoooa assume —o=o.ooooooo go ouoooooo — uooo» no coco—osou 1 on» cognac o.sopozuoo o oouopoeoo 1 oz oooooo o.co—ozuoo o =o»o no: 1 oz . ozoozooo ozzooooooozo . O l—NMQQ zo o» opoo7»o> 1653 .ucooooopooo op oozoooouopoe woo mo opooooo> .mm oeom_u A» omoo co oo>»m ooocooo cum zvrgo>v ooozompoh xom om< "oooeoo< z.ooo< oozoozzoz m>.:o>.zomzmo maze wzzzuz>zmzzz mm>oo>zH zuzooo on» m» xpco zoopopog oozuo .N mu Fooooooowoco ooooom .» oco ooo>eom o o o ~30» oz zozz ooo on . Homooo 1umzo: oz< mz ooo mu moo—opm hmoz NIH 30> mm>Hw 013 zommma mIH mH 013 m— w u NPN *> Nu Upwcu “cmawuwugcm ossow zoo>gopow ucmwuoouoo co ocoow meow“ a 1 ppm z> mm _u ooooom oo oozocoopopom11ouo PFN oucoom ouoo oopooogo> oouo_oz1emoc_ coououoooemuoz oEoz Logan: o_ooooo> o—oooooz 16541 .ucoo»0»ugoo ou coouo_oc c» zoo woo oo o_ooogo> .vm atom»; A» oooo co co>»u ooocuoo woo zu—co>v agency—op zoo ooz "mongooz A.oma< ouzonpzu: u>.:o» zomzuu m_:» wz_:m_>zuhz_ muzoo>z~ zuz»— uoo oooo cooooo toss: o>ooo mg» »» zpoo zoouo—os sozuo .~ ooooom .— wzo» o» zo_»_u oz: zomzuo mzh m_ oz: .m— meLow zo»>gou:» «cuouoouso co venom souz A3383 135° .o ooxo_o5u .m Loozoooz .o ooo.;u .m Ammouooov o>»uo_oo Lozoo .~ ooooom .— ozo» o» zo_p<4uz m.zomzuo mzz» m~ pzzz “oxuzo oz< uz~z¢u ox: zomzuo uz» oz: oz: mam: mzcouc» acoouooc» oo ooooz souz z 1 u zoo 1 cu; mpw _ oooo zo» .o_ .o» .o_o v eo- .~ mono Ao_v oooooooum N 1 z ocoo_o_oooo oooom ooooz11uoozo .oguoou oooo _ 1 o oo zoo mow oocoom oooo coououooocooc_ ozoz Loazoz o_oo»oo> o_oooeo> xo_ooo_o oo oooo ooo,o_ooooeo 1 o ocooo_e 1 a go o1» 1 MNN Lo> yo mocoz coo oc. o»~.» Co oo_o» ozo oo __oo oozo»: o=_1»1 oo ocozo oco ozosoww mm 1 o o» 8o co a mom o. 1 o o o o N 10o o.oa . go o _ xWo 12mm o o» woo ooo o_oe o» oo 1 o o m » 1N _ mow z>w zoo ou_ooooo o» ouo oco opoe o» no 1 n m m N _ zoo «Eco o. woo uoo opoeoz oo oo 1 ~ :o_>eou=_ .oouooz om o.ouo 1 m_~ z> xoo oooooooo no acooouougoo ou m—~.mo~ nu~z> oo oo_o> on“ woos: oo oooo moose—poo ozo mo: o» woo woo apnea» oo oo 1 p ozouo_oz gum o.uuo «KN ougoom ouoo no oo—uoEcom oo oopoz ooououoeooouc_ «so: o_oooco> zones: woo: oo—ooogoz 3oz oo>oooo u—ooosoz 165 .mpoo_go> mucmowomg an mswu we cameo; "om ogomom 1-1- Aocooo»ev o 1 go o 1 N 1 mocox ogoe go ogooz o. 1 o ogooz o— cozu ooo— woo .ogoox m oooo_ u< 1 m ogoo» m co>_o gozoco opooooo oz .o can» ooo» uoo .gooz _ oooo_ o< 1 e um xo ooooozo coouoooo .m goo» o oozo oooo 1 m mgoa go ogooz o» .o coca: o coca oooz 1 N ogooz o» cozy moo» .ogooz m poem» u< .m ogooz m cog» moo» .gooz » Homo» o< .o .Aoogozoooo ou goo» o coo» oooz .m ooooo ogo zoo» zoos: oo oucooooog zucoe o oozu oooz .N gov oooooooog ooooogo goooo uo ouxoz hzmmuzo zoo» h< om>~z so» m>_p on; acoo_o_ogoo ozu oeoo mo xu__oooum "msgog :o»>gouco accouoouao co ooze» souw zumoo» go o—ouo oogoogo oo o» <~o~ z> ouoooooom <~o~ ougsom ouoo ooououogogoooz oeoz gooe:z opoomgo> o—ooogo> 1665 .opoowgo> ouoouooozou oo gonzo: .sm ugaapm oo>oo gozooo o—oooou oz ”a xo oooomzo ooououoo upmm CONCOO‘ oe_» .zog oo»o_o2o 1oz a z oooooooogoo xom on oozoooouopoz A.3opoo xmo oco o_zooo_uo_mg mpooev woo» zoo: om>o— oz: .op woo» so»: m>_~ w—oooo gozuo xooa 3o: .o orpzoo zoozgouo» oomouooooo co oooom sop” A.3o_oo xoo ooo oozocoouo_og opoopv N—ouwoooz on» m>oo_ no» con: :0» zo_3 o>o— __o3 oz: .o— g_oo_oooz ozo «zoo» so» coo: so» no»: o>o_ .__3 o_oooo gozoo zoos go: .o 51p 1 ooooz a a «poo oz o.1 ouoouooozou oz ocopo o>op .Foz go oooo ouoouooozou "oEgoo 3o_>gooom oooooooco oo ooooo swam cozwo oo 2 mo 2 o—~ mogoom ouoo :omuoumgogooco 02oz ganeoz m_ooogo> opooogoz .o_zo_zo> ouoooooog mo goz:o "mo ogaoou 167 .oozooo go__ogo ozo zo_z: co ooo. oz» co ooxoo zoo ozo ooo. ozu czo oooooo o_z go oco_ooo ozo ooo_oo o_zo 1go::o ozoz oo ooc_zoo ooz o_ o_zogoz:o gopoog» .goo:o ogoz o oo ooooopo oo oooooooog oo ooo.; ooz oo goo:o1o0 oo oz: acooooo < .oc___o:o gozoo oaoo go goz:o ooooomoooo z__omo. ozo on mos oz zooozu go::o oaoz ozo zoos: oo ooooo_u . oo c:o uoc oooo oz zooz: oooooooog osoz oz» :30 ooooongo whwm o co oo>op oz: ocooooo o ooooo_oog oz: ogozuo go ooaooo go opoo An ogozou Sop” pocoouucom zoo: or ogoz ogooooo ozu oocom ooucog m— oucoomoog11ogoz oozo— oz: ozo oz .p oozoo go:ooo o_ooooo oz . oeoz oz» oc:o ogoz oo>o_ oz: oo_o oooeooo . NMO" o 1 ouoo o: go go:ooo o_ooooo oz Noam: mac: zoo» mzzo oz: .m— m 1 ogozu oo>_» oz: oo_o ooooEom N 1 oooooo go opom oozoo_ooz oEgom :oo>gouco ucoouoouoo co ocoom Eou~ P 1 ogozo oo>o_ oz: oco oz mo oozogoz:o oNN ougaom ouoo zoouooogogoooz oooz gooEoz ogooogo» o_ooozo» 168 .opooogo> opozoosoz mo oooz .oo ogooog zo>wo go:ozo opzooou oz o .z »o ooooozo oooooooo o ,. u_om N o m o m r111nII111 1 .1.1 1 111101.111 l 0. 111.11 11 .Nl p osgo _»og oo»o»oso ooz .orrz oooo_o_ogoa xom on oozocoouo—oz go:ooo o 1 opooooo o: go oooo oz m 1 gozoo Azopoo gooEo: o—ugouv N 1 ooooom ozozoooo: N>o~2 o—zoogoz 1159 .uomzoooocoo uo—zomgo> coogouogu oo_m ogomoz I'lollI '- .ll 01 go:oco o—zoooo oz . ooogogoooo_o . oopoo_uoo x—ugoo . zoooogogooo oo oo oooz .1 ooo_oooo ooooo_ooo . 11. iihwmmmmmmmmm ozo oo ooooozo ozo ooog oo ogoo ooo ooogoo_o mozu yo co_oouogogooo_ gozogoz oz< >zozogouz_ uooFuoozoo co ocoog sou_ P 1 Amy oog_o_oooo_o zo_uuouo_oom. NNo .Ago:ozo opooooo ozv m 1 oooo ooocoozgoo oozo . ooou .ucoo_u_ugoo ozu.:¢ ogozozo go gozoco go:oco o oo o. zggo: oeoo ooooo»ooo zupz: gwmmwo zcw .w ozu oouooa—ooo oo_o ooooeoo go coo ozu g— oo»go—oooz= 1 oz ._ m 1 oo>_m go:o=o o_ooooo oz Nmzuuz ozh N 1 oo_g__o=oco 1 1oz: zopooz hm o_zoogo> ooooo- 1 go o 1 o 1 ooo g» go ooooz mmmmm1 1 ooo z> zozo .ooooo- 1 o~o1»No o.g» o ozo go »oo go .oo_oo_z oo; oooz > co ooog: oco ooocoooog o>_g oz» go oco oooop oopzoogo> m oooz» co oozozoo N _—o uoaoo oo goo oooo ozo mmmmm- 1 ooo z> oo mmmmm1 1 m .m go ogooo o_zooooo o go ooo ogoum _o_ooo_ qu 1 .No mNo1—No z> ooouoz ooocoooog oooocoozoo: go z ozo 1 ovc 11ozo2ozouoou oov ougoom ouoo oo . oopuoEgoz mo oopoz zoououogogoooz oooz opzoogo> gooeoz oooo oopooogo> 3oz oo».goo opoo_go» '17O zooom go:ozo o_zoooo oz .m oogmoo_o .N oogm< ._ szmozo hue a m< u: «on mmzoz wzmhhmu oumz pmoo mwz_xh: > zoo: oooog oz zoo o—oo_go> owzu mo comuouogagouo_ gozugoz m N gozooo o_oooou oz 3.533 No ozz oooooo oooz» ogoz oozozooo zo oozoz oo» ooooz _ 1 oogoz zoo Eooeoooo oNz goxooo opooooo oz .m oooooooooo_o .o ooogooooo »_ogoo .N ooogo_ooo ._ ._No o» oo.2 oooog oo »12 “ooooouougoo opzomgo> oozo mo coouooogogoo:_ gozogoz o» oouwozu ozu ooog o» ogoo ompllNomzzm_h opoozgoz »:oEoog»>:o omo u< go»ooz ._pm ogomzz 171 zoo» go:oco o_ooooo oz m E zo oooozo 8:38 o too N goo g; 8». o» zoo: oooo» ogo zoo: ozao» u:o»»oozoo o oopooggo» oo»o»oggo»c_ o» co_»ooogogo»z~ zoooo om< mNm m Non zz on oz zoo: ozoo» ogo zoo: ozoog »oo»»oozoo o oo_oo»go> oo»o—oggo»z_ o» oo»»o»ogogo»o_ z»m go oo< «NM m Non oz on oz z»»: oooog ogo z»»: ocoog »oo»»oozou . N oo»ooogo> oo»o—oggo»o_ o» co_uo»ogogouo_ z»v oo omz nNm Non zz on oz _ zoo: ooooo ogo z»»: ooooz »=o»»oozoo -1 oo_oo_go» oo»o_oggo»c_ o_ zo_»o»ogogo»c_ ogm go oo< NNm 8.: :og ooo.?“ ooz g z ooootzgoo goo .» 8" oz xoo o» oozoooooozoz z»»: ocoog ogo z»»: ooooo »oo»»oozoo oo—zoogoz oo»o_oggo»o~ oo oo»»o»ogogooz_ ozN No om< —Nm Non zz mm 1 oo»» co oEoz on go oooo oz .328 ooo 3%: 5.; 958 ogo oo 1 .96 go oo 28328 . Huh “ .__ oo_oo»go> oo»o_oggo»o_ oooom oo «goo» o» om< »o_ oo om< oNn .ooz : oz o oooo o zoos :o . m 1 mom 1 man oz 11 1 z: : » .55 z o o 1 goo - Noo o» o 1 gozooo "oegoz :oozgo»z» »zo»uoo»oo oo oooo» so». m 1 .mn 1 on zz oz go goxooo opooooo oz N_o»»oooz ozo ozoo_ ooz ooz: 3oz mm 1 mNm 1 on oz o 1 ozo_o oo»» zzox go oooo »»: oz : o oooo go o zoos :o . m 1 m—m 1 «Pm oz N 1 + oooogoo oozom z »— .z» p z» z m mm 1 nzm 1 mom oz N 1 ozoogoo o3» ”msgoo :oozgouo» »co»»ooz_ co ooooo moo». o 1 Non oz go _ 1 ooogoo ooo o»oo»_zozoo »o z Non 396 mm 1 o»oo oz ozzzooguo »mg»z »oozm .og»ooo oooo mm 1 gozo go om om< »o »zoo»u»»goz Aogooz. c: cozzm no om< mo oo< mo ougoom o»oo oo_zo_goz oo»o»oggo»c_ oo»»ooogogo»c~ oEoz goossz o_oo_goz o_oo_goz .»oonooo 2»:E»g3§5z ogooz ogoe go o. .o oz< mo ooz» gomoooz .o o» eoz: N 1 omoEzogozou om< ,go ooo »ooo_ »o .ogoz»o .2 A zoo 1 131: - oo» : 5:. oo»: ooooooo 2: 1 N z 1 »zoezog_>zu om< »oonooo .o— v 3». v 9:2; 1 ooz : o5 :oz» gomSoz Aoooo zooo zooo goov ocooz oops» zuoo o. .v ogooz o. zoz» ooo— oNo .oNo _ 1 oogooggsoo ooz 25 .oo 55 gomSo» .o zo .MNm .NNm .oooog oo—oz zuoo oo o_ .m goo—o ozo me .o .me .on mo v oo_o> gog mNmz1oan zugoom .N ogo Eoz: No p_o .ogoz»o .Non .mo »coEcog_>om om< Ego» »oo oo .o 1 Nomz oz .F z»»: oo>»_ »uooz:o oz» 1 — »coezog_zcu omz ougoom o»oo zooooEgoz go oo_oz oo»»o»opogo»o~. oaoz o—nozgo) no oooo oo>zgoo oo_oozgoz 3oz 1172 .oog go oo.oooo» .o.oooo.oog "o.goooog "N_o osaooz Till ougoom o»oo 11 Azzzoooov o»_»o_og gozoo . .» omoa co oo»»o ooogooo woo zgzgo». 111.1 ocozoozoz 11o 1oz Homogooz ..ooo< oozo_»zoz o».oo» renews m_zz nz_:u_>zu_z_ muz»0>z_ zu¢_»o—og goz»»o~ ooogoz oz» no ooczzoo o» Aooov go». ogou ogoz.go < 33:35 .33...» »o as: as nun .0 mmmmm1 1 go n1. 1 .oN oz go» ooooz oo»»oooo »ooo.ooz ougooo o»oo o. oo»»oEgoo go oo_oz eon: oozoozgaz 3oz coz»o»ogogo»o_ ozouoo_znhg~> >_g _m~ 500832 Opacwgd’ '173 .—g oo»»11pozgzgoo ooooou Nu go»uog .n—o ogong oogggoe gozoz .o 111.1. oo:oo_3 .m Ax—_ozog»zo»z»v oo»ogoooo go ooogozzo .N A.u»o .ougzgoo zgo»g»_e .zog»oN»—o»goooz oo zooo m 1 o»oo oz oooooog goz»o gog oo»ogoooo oz hoe oooooov oogggoz .» o 1 oooggoa gozoz Aoozoog: go oougozzo mcgoo oopo gozo m 1 oozoo_3 oucooouogo oozo» oogggoz ”o»ozv Nomzzz 111G u m—N §> mm > 3m QPU a on n N 1 ooN »» 1 o.o.N 1 SN g» s» g ooo . .5; ._o 3: o _ zosgoos 1 3:11.12sz . » o ooooo- 1 go :32: 2: oo ozN ooN zz o»og o—N zz ooouoz N 1» 1 ooN zz go omzoz .ozozgoo ooooom ooN ougoom o»oo oo oo»»oEgog go oo—oz cog»o»ogogo»z» oEoz opzoggoz gooeoz com: mo—zoggoz 3oz oozggoo o—zozgoz 'F74 .Ng mago11gc>g>g=m omaoam "u gouoog .cpm ogam.g o 1 moon on . go Am mo omogo>_o moou . uo: oov gozmco m_ooooo oz ocwsgouoo ou «pace: “a m 1 con mcucoe xgm casu egos omzoowz .e w H oc_agoumo o» m_ncmmagz o>P_oowwuumeoMMwmm .m m mcucoe x_m coco mgMM oonwog: . . . 1 o ego m “ouoo Agwuoomv mgucoe xgm umop cw oomomuoo mmaoom — go .oougo>_o .oogggoa gm>oz N>_Pu ._gum omaoom m<3 aux: anxmo mgacos xwm umo— mgu op vopv mo: umaoom uuv .m p 1 mcucoe xgm umo— cg omgo omooom umooom no space «no ougaom ouoo cowuoumgogouca onz gonzo: o_oo,go> opporga> mmmmm- mo N1p 1 moom> go mmcam . 1 moou on mpnupwo>o mono uco_u_oo:m=~ 1 mmmmm- mmmmmm 1 www“» HIMVM.W H mmww» wmooom mcw>w>gom o: no; ucoomomugum 1 N p 1 momm> N 1 mmoN> mm>p>g=m acuowowugmo go mmaoom 1 p zo_>gwu:v mmmmm1 1 go zucoe1m use an nmo mmoN> one. mmo§> ooouum N.F 1 moog> go mmcom ~o>m>gam omaoom woo ougaom coco no covaoEgog go mm_=¢ coguouogogouc_ use: opnovga> gone:z vow: mo—aovgo> 3oz oo>.gmo opoavgu> 175 43:23.6 ooo go as a .33.: as 8%: =.mzoo own: no ooocaoo hmzoo go goneozw z_—oowuo2ou=o mgo mzmo .augomoz .co_gou_go gozuoca m. zoo zgo>o mammogo ucoooowugoo oz“ N>z¢z no: mxumz m. 1 sumo oz o3» hm.m no «zoo: N umm— ozu cw Ammo .58 21:52: 8 1 3.. 3.. So 2, ._ 1 o2 ; t 15 go :8. go :3 oz F. 28... goo oo 2 «covuaouao co oooo» souu co 1 no: zoo Noe m> .N 1 may m> o~ Ae—1oo sag» momcozv havooago: Noe ougoom auoo mopaoogo> oouopug1gouc~ coououogogoucm ogoz goosaz opoooga> upaoogo> 176 .EtzgzagoszzzzgosfiBass agssgufiotaz .mcgoo go ogoooou no ozmgos «no: no: .uzoor oxp «moo ozu cg ocoo appoauuo no: «concoouog oz» nag: uao vcgg on coo: o3 .naooz one «man one maggov mcgcoguucsg go poso— «coocoooo unos ozu opuggu goxmca o—oooou oz .m pomogoz .N g—om ozone uoc moot .o 2oz» o» uzmoogo goon: goon no; .u soon on» ou go .ooo ozu go «so one cg moguuom cogog>goo=m go ugoooam .ouooomgomo moom .o go .zooo ozu go ogoo ooo cazu ogoe mcgcuun cg :ogmg>gooam go ugoooom .oucoumgmmo moom .a ocoooomoo .N go .Azugeogoxo oopoomgo go goon mo zoomv ogoo opmc—m a mcgzuoo cg oogmg>gooom go ogoooom .ouooomgmmo moom .o .zqu omcoom go .nou .gozozo cg .z—ooo—oeoo g—om mozooo .o momoooooooo .. Agoogoz zgoo ooog oog ozooo ozogz< oooo gzgoo ooog og oog zoo: oo ozogzo mooo ogozg zoxg ogoz oo goo. ooozoo o gozo ogoz ooox goo oo» oo - oog goo Agoggoo oog oooz ozogzo oooo gzgoo oozoom go goog ogozozm < zg gozgoo oog oo 3oz .oopgoo o 1 goxmoo ogoooou oz zoooo: oxo omoo ooo ogo goozozm oo oogzgz< moon .moz opooggo> mmmmm- 1 o 1 o 1 mmo gm go omooz mmmmm11 mmo g> cog» .mmmmm1 1 oNe1m—o m.g> o oz» go zoo go .mmo g> .mogmogs mm; o. ooggz ooo mogooggo> o mogog>gooo xgm ozu go ooo «moo_ ogoum _ogog:~ ~uog omozo og mozoooo . ._o ooooo on gog noon oz» .mmmmm1 1 mac g> g_ 1 m=g>go —goo mmmmm- 1 m .o go ogoum go mogug>guu< mmmmm- 1 N opogomoo a go poo mogug>guuo cg oucoocoooucg oNo 1 m—c oNo1mpo g> oooooz soooooooooogo go z oz» 1 mac go xooo—v oa< mac ougoom coco oogooEgog go «ogoz cogoooogogoucg oooz opooggo> goasaz n< non: oopooggo> 3oz oooggoo ogaaggo> 1177 A.oouoouv uo—m ogamgg .m_o o> oo» .oogooo oooo oz» ooooo oogooogogog gozogog gog goxooo opooooo oz .m pooogoz .g cogoogoxo go ooomgo ocgooo—o go .oozao—o oogoooggo o. oogug>goooo go ouooomgmuo moogooog .o . um .oooo oopgou ggo ooo co mcgouoo o. oogmgogooom go ooooomgnoo moggouog .u um .ooog oopgoo o» mogooom cg oogmgggoooo go ooomoogmoo ooogooog .o go .o. mogzooso ooozogx .ozmgo «o gozogo mom: go .oagozoo ocggoo oooesou go cooooo oomo .o ocoooomoo .N .uoo .oggozupooz: .mgox_ox .mozooogu .oooo mo zoom mogooooo pougoozoos on: oo o gog o_oooooooo mg a. "ooo: .o o! moguoewlmmm1umoo usage on oooseou go oooooo :3o omocos zoe .o um .cogoogoxo go oooogo oooogo “mozoogu oooooggo .u mmMIHwo_goo ggo woo co moom .o oco .eoog oopgoo oo moom .o ocooowmooco .p gzooozzoo. zooooo zoo ooog ggoxo oo» oo o 1 goxooo ogoooou oz .Amg¢gzov goumzzog oz~zz< moon Nzoozzg ooo .oogooo oooo ozo ooooo oogooegogog gozogog gog gozmoo ogooooo oz .m pooogog .g Amazonoo go>o ooogzooo cg .ggo noozo ..m.o commogooo apogoo mogoeog ..o.g. oggouo oocozo ooc moon .u .mm oommogo mcgooom cg cogmgogooom go ooooyMWmoo oo>gooog .o go ogoxogo ooo noooo—u go uao mogzoopu mogooom :— oogog>gooom go ouooomgomo moowmoog .o oooooomoo .N Aooou socoooooooog. co gog ggoooouoc poo mg goon: mgooogo xuoo ngoog~ go .moozm mcgogo ooouxov mgoooooog mogzoopo ggo moooooe ooo .ogggog—m go ooozm oco omcggoooo .oocosgom goooo .mozoopu co mono .u mmMIwhzgommoooo ggo goo zgooo mooogo co mono .o coo mgoxogo ooo moomogo sogg mozoo_o ooom .o oooocomooog — o 1 gonooo o_ooooo oz Aggoo googo oomoogo goo oog oo goomoooo g 1 .ooogoo hum no» so”: mzo>z< mmoo gmzmzzzo oz< mgumoou go goo ozgzgooo zoo» goo oo» N 1 ooooooooo od=§2§o3o138& csgzgouao: ssagogsogfizagszg .1 éssego osaeo o: oogaom some coguooogogooog ogoz gooaoz _ o—ooggo> opooggoz 'l78 3.2.8. as 25o: .m—o o> oom .oogoom oooo ozo ooozo zogoosgogzg gozugog gog gozmoo o—zoooo oz .m _om:goz .N ngooog oaozo>ogozg co ozo__og11—_o oo ooo ooc mooo .o go ngooog go ouo cg gozoo zz zo>gm ouzoumgmmo .o ozoozomoo .N .ooogz go nggooooz ozo oooe go moguuauogm no zoom .ooog go zogoogooogo cg ouzoomgmmo mo>gouog ozoogugogoo ozo gg .ozoozooooc_ ._ mo oooo oo o_ooooouoo ”oooz .u m an zoo: oz mocoEopoeg _oom: oozo zgommouoc ooz "oooz .o zoooe oozg .oco_o>gooo mug gov ooo_o Eogg oomwxmeo .o m H gozmco o—nMWMWgwfl ozoozomoozg .— N 1 ozoozoooo Amxooz oxo oooo zgv Nozggz< N>zwuum so» on — 1 ozoozoooozg ngoou m—v .m_o g> oom .oogooo oooo ozo ooozo zogooEgogzg gozogog gog gozmzo ogzoooo oz .m _om:gom .g ggozo go ooo sogg o>oe ooz mooo .o um ggozo memmm ooo go one one no mcg>oa cg oozoomgmmo .o gamuooomom .N .uoo .mggozo_oozz .mgoz_o3 .mozooago .mocoo no zoom mogoooom .ougzozooe ngmo oz ooc zoe go hoe “oooc .o x—ozoozoooozg mggozo ozo ooz go ago new zwlmmvmz .o ozoocoooozg p m 1 gozmzo opooooo oz zgmooozo go goo goo g 1 gooogoz oo» oo 3o: gooo go goo goo oo» oo 3o: ”ooogoo zmgooz ozo omoo zoo N 1 ooooooooo Namag gm cum go goo oz< z. goo :og z o_zoggo> 179 «wo.ucouv "m—m ogomgg .mpo g> oom .ougoom ouoo oz» ooozo cogooagogcg gozogag gog I gozmco opzooou oz .m _omogoz .N onomo_oo go .mcooooz go\ozo mpozgg: go om: go .mgouozooo .moaoco zz Pogocoo go zogmgogooom go oucoomgmmo _oooo go .ogogoo .z .Hm zooz go zogoooogoo go :ogoocggz cg oucocgucoucg —oooo go .ogogoo .o .wmouoummm .N .uoo ._ocgg: .zooooz .zsoomo_oo .moggoogmoooom .moEoco mo zoom ozoEooocoe pocgooxo go pogocoo pocgoozg zz gozogo .oop—ogocou1g—om z_og_o:o cogoouogoo oco zogoozggo .o wmwumwmmmmm .— wmemoguu< m 1 gozmco ogzoooo oz u> oo ”muoo go oomgzou Nmoo 30> oo g 1 .omagoz Noozooogo :ggz mgzoogooo moo: oog oo zmgooz ozo oooo ggo N 1 ooooooooo Nzogg u>gu o» zmum mooog z~ opzoggo> 180 .ocozonzoo goo... 323g poozzo113g5zo pogozozgo u... goooo... "NS Eomgo ooNo oozo ogoz ooN._o oozo ogoz ooo.o_o oozo ogoz .o ooNo oo ooNo ooN._o oo oooo ooo.o_o oo ooo.o_o .o oo_o oo oo.» moo» oo oomo ooo.oo oo ooo.go .o on.“ oo oo.» mom» oo ggo» ooo.oo oo ooo.oo .n oo» oo ooo o_oo oo omNo ooo.oo oo ooo.mo .N oo» oozo oooo ooNo oozo omoo ooo.mo oooo oooo .— zooz goo zoooz goo goog goo mam z.oozo_gzoz gmoo o>.o ozoozo oo mooooom oooommoo ozg ooz goooo ozozgo .goo zoxog moo: mzoogooooo oozgo oo moxog gzo oooooo ngzoz N_ gmoo oxg ooo ozoozo oogog ozg oo ozozg “moooo z.ozooz_ ooog oo gzoozo o_o_ooom ozg oooz g.zoo o3 .ozooz_ o 1 oooo oz oo.moozoo oo oooog < oog zoomo oooo og oxoo o._ no: m 1 zow ooo.oo oo ooo.mo gooeoo ggoooomo ozoogooo< N 1 N ooo.oo o» ooo.mo oo oo oZoggo> 1E3] 33 1 ooo.-«gooey. No 2:. on ago—3.. ..oN 1 oooaogoooog .1 11_go1_.oosgogoao 33 332g; :3. 11...; .ll-rt ' I oogoooo .ooozooaoo ooooozoooo 28.33 go 18511323. 7.2.2.: "g 33: 12o 23: .mmN 2g gog 2:: 05 3 Zoo 55:; ooo.->1 19:3 Foo mmNz ooo...— .o 1 oNN go 3N; g. 13:. .15., .3... 13:. n .32. .3... .g. 63: 13.2 o .3... 6me o v v :1 m m m n n n m m ..11|.11 111,9; onN; acouuaaoo «no! u v . . . oooooo ooo oooz .3... a — $2: n N £1: m «3: v on AWN. “an”... ogooo “coo: 1 p 1.11.1L.'|l1 “a 33:2; 33 ocogogggoooo 1 $3? 1.1111,; .11 71111.. 111111.111- 1111 . opouméoo «co—“3:123 1.11.50 1 m E E E E E E E E E .E E E 21:1??? o is? :13 .3; «1: oz or: no» oz .3; 2g 3:. no» oz. m > go ooooogo go 222: 359. .3 ll .93... 56 ac: oooo 133—Boo “oz 1 o A: oozogoau E ago—95 ooz A: ooze—95 AN“ Boo—95 «or. 3N; .ugooooo Azooo go 32:3 go 9.32: 359. .3 .82 «£6 133—Bo ooz 1 n .N. gooxo ogoz z_o goozo ogoz ooz .o1ooooo 11111 1 Aoooigo go 322: zuzomo co ooc «oNN o> oozogozxo 86: % €23.98 .3: 111.52 6.6 ooo 88 1 n 733...... go ”:32: :82: 3:8... co 1 j .goc 19.5: 516 .go: .113: €11}.an 1 N .36 SN SN; go 2:? . N. 2.: g . . Aooolgo go 9233 ogooooo 3:2... Aogouw :32: SN ocN 23 out) 3 83 05.828 o5 3: co “co .oogogneo ooc 1oeoz tic 1 N 3:00:88 = 8;... 23. 12d — 1 <0NN; 1 n go _ 1 oNN; 43.13.... go 232: 1:93.; 35:3 8.3.55 mnN .oNN «oNNoo ooog oNNgo oooooo oo ooo ooo ogoz ooxo .oogogoEo 1 _ go 1oooo. ooo ooN 3.53 33 no zogoaogog go 033. 5333335 053. 933g; gnaw: 237:; 182 4323. .2. 11.5: P ozooos zoo ooo— oz» ooggao zoo gog ozozgos ooooo»o go oogg»og11oz moo gog zon o opoz go>oz ogooo ago) 0» coooo »oz .xoo gog az.zgo3 vozzc»o go tog_»cg11oz ooo ozocoa z.o ooz» ogoe oogg»og11oz Aozoo— zogo zo go .z»coe\gaoz — zozov oozo—oso ——g»u11uo> Aoozo—zeg ——.»o ogo zoz» .zoz ooze» zoom oo»»»oo m 1 gozozo o—zooou oz 1. 3.23:: o: 35%... z: 5: oz» 9:55 3:335 go 38 oz ozooooo zoo zoog zoo. ozzgoz oog o_o .1oo1 ._o .zooooggooz _ N 1 .o1o1n.No oz oozo. ozooz zoo» zozg g_oozo zo gozzt zoo oogozozo oo» oz< .o p 1 ._o oo» »ooa>o.oou no. goxmzo o_zoooo oz .m ~¢ no ooozgzm zoo»ogoo 1a a 1 goxnzo o_zoooo oz «m: .N a 1 o»oo oz go > — ao—v oozoo »oz zoouoooa az»co£\gooz . zoom. Nozgzomoz oz. zuzugzo so» wooooz gmzo N 1 oz zoo: ozog zooo »_oooo zo gozoz zoo oogooozo oo» woo: .oo _ 1 11g oouazo.osu .oN goo—ozocz ozzoom og<>_oo oo .mzzozoo m 1 gosooo o_zoooo oz “ogoz 1mo>~gz< N>_zuNz N 1 oz go mNnv o»g.u Azmzoom «:3» «a. so» ogz mmgzc11xwm.mmw~zgaw1cmgzzo .N Aoov — 1 oo» «save. go oogaom onN A_mz wagon—oz» 15:—z zzo go a»:9zzo: ogogzo3 .oo_ooogo ozo go. ooz 1o=._z ozo go. ooz .oooz ozo o. ooo. ooozogz_mo< o__zzo zoo. ozoozo gzz o>_oooo zomooom zoo» zoo oo» o.o ngzo: z_m gmzo oz» oz_ooo .n zooo cocooooooz ..z_ooo mo ooooo ozos no o_og.uv omz_zoz xzm gm.— oz: »zo.»oo o oozoogoog oEoz oz» czo o o g g. m .N nozo»»oz:g z».3 o. ogoz zgoozoo oz» ooz_m ogoz ozgp oz: ogoz»o go ooaooo go g—oo zz ugoz»o sogg oo»zog «g oozoogoog11ogoz ooo.. oz: ozo oz 1» m 1 o»oo oz go goxoco ozzoooo oz . 1 ogo » no» 3 oo o oooooo SS: Eoz zoo» 2.5 oz: 2 M u z : 2113? .3 row 811311. ongog roozgo»zg »zo.»oo»oo zo ozoog no»_ _ 1 ogoz» oo>__ oz: oco oz go o.zogooso ONN 111. . oogoom o»oc zog»ooogogo»z_ ogoz goaaoz o.ao.go> opooogoz APPENDIX C DATA RESULTS NOT REPORTED IN THIS STUDY 183 TABLE Cl: Data Results Not Reported for Total Sample--(Ho3C.A.D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Prior Score l 57.437 33.06l** Covariate=Age l 20.l05 ll.573** A=Age Environment 1 .154 .089 I 3.033 1.375 B=Residential Location ——bf Primary Care Giver l .699 .402 l 1.59] .721 A X 8 Interaction l .002 .00l l .008 .004 Residual 183 1.737 l92 2.205 Tbtal 188 2.198 l95 2.194 Covariate=Prior Score Covariate=Age A=Age Environment C=Spouse Survival.' A X C Interaction Residual Total Covariate=Prior Score Covariate=Age A=Age Environment D=Days of Bed Disability R’X 0 Interaction Residual Total Covariate=Prior Score Covariate=Age A=Age Environment E=Independence in Activ- ities of Dailv_Living A X E Interaction Residual Total Covariate=Prior Score Covariate=Age A=Age Environment F=Financial AbiTitx A x F Interaction Residual Total ”a < .01 *2 < .05 184 185 TABLE C2: Data Results Not Reported for Older Sample--(HolC,A,D). Analysis of Covariance Analysis of Variance Degrees of Hean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Value Covariate=Age l .408 .262 ' A=Age Environment 1 5.288 .399 1 4.789 3.084 BfResidential Location of Primary Care Giver 1 .080 .051 l .042 .027 A X 8 Interaction l .737 .474 l .568 .366 Residual 139 1.556 140 1.553 Total 143 1.559 143 1.559 Covariate=Age l .408 .279 A=Age Environment 1 .999 .683 C=Spouse Survival l 4.558 .116 A x C Interaction 1 9.233 .313* Residual 139 1.463 Total 143 1.559 Covariate=Age l .408 .265 A=Age Environment 1 5.532 .600 D=Days of Bed Disability» 1 1.523 .991 A X 0 Interaction 1 1.978 .287 Residual 139 1.537 Total 143 1.559 Covariate=Age 1 .408 .269 A=Age Environment 1 3.959 .613 E=Independence in Activ- ities of Daily Living l 6.384 .213* A X E Interaction 1 .093 .062 Residual 139 1.515 Total 143 1.559 Covariate=Age l .234 .153 A=Age Environment 1 3.097 .023 F=Financial Ability, 1 .039 .025 A x F Interaction 1 2.425 .584 Residual 135 1.531 Total 139 1.531 **p_< .01 f2_< .05 186 TABLE C3: Data Results Not Reported for 01der Sample--(H02C,A.D). Analysis of Covariance Analysis of Variance Degrees of Hean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Va1ue Covariate=Prior Score 1 20.358 l4.018** Covariate=Age 1 .016 .011 A=Age Environment 1 .626 .431 1 .000 .000 B=Residential Location of Primary Care Giver l .673 .463 1 2.222 1.347 A X 8 Interaction 1 .762 .525 1 1.438 .872 Residual 99 1.452 104 1.650 Total 104 1.601 107 1.638 Covariate=Prior Score 1 20.358 13.984** Covariate=Age l .016 .011 A=Age Environment 1 1.267 .870 1 .304 . .183 C=SpousekSurvival 1 .710 .487 1 1.291 .778 A’X C Interaction 1 .369 .253 1 1.318 .794 Residual 99 1.456 104 1.660 Total 104 1.601 107 1.638 Covariate=Prior Score 1 20.358 l4.604** Covariate=Age l .016 .011 A=Age Environment 1 .400 .287 1 .022 .014 D=Days of Bed Disability 1 7.163 5.139* 1 9.545 5.995* A X 0 Interaction 1 .038 .028 1 .090 .056 Residual 99 1.394 104 1.592 Total 104 1.601 107 1.638 Covariate=Prior Score Covariage=Age A=Age Environment‘ E=Independence in Activ- ities of Daily Living A X E Interaction Residual Total Covariate=Prior Score Covariate=Age A=Age Environment F=Financial Ability A X F Interaction Residua1 Total "p_ < .01 *p < .05 187 TABLE C4: Data Results Not Reported for Older Sample--(H03C,A,D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Va1ue Freedom Square Value ' Covariate=Prior Score 1 20.358 14.195** Covariate=Age l .016 .011 A=Age Environment 1 .684 .477 1 .009 .005 B=Residential Location of Primary Care Giver 1 .178 .124 1 1.086 .659 A X 8 Interaction 1 3.043 2.122 1 2.860 1.737 Residual 99 1.434 104 1.647 Total 104 1.601 107 1.638 Covariate=Prior Score 1 20.358 14.135** Covariate=Age l .016 .011 A=Age Environment 1 1.904 1.322 1 1.285 .802 C=Spouse Survival 1 1.758 1.220 1 5.053 3.153 A X C Interaction 1 .861 .598 1 3.502 2.185 Residual 99 1.440 104 1.602 Total 104 1.601 107 1.638 Covariate=Prior Score Covariate=Age A=Age Environnent D=Days of Bed Disability A X 0 Interaction Residual Tota1 Covariate=Prior Score Covariate=Age A=Age Environment E=Independence in Activ- ities of Dai1y Living A X E Interaction Residual Tota1 Covariate=Prior Score Covariate=Age A=Age Environment F=Financial Ability A X F Interaction Residual Tota1 *fp_< .01 fp_< .05 188 TABLE C5: Data Results Not Reported for Older Sample--(Mo4C,A,D). Analysis of Covariance AnaIysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Va1ue Covariate=Age 1 1.615 .977 A-Age Environment 1 3.198 1.936 1 2.561 1.545 B=Residential Location of PrimaryACare Giver 1 .404 .245 1 .294 .177 A X 8 Interaction 1 2.976 1.801 1 3.163 1.909 Residual 100 1.652 101 1.657 Total 104 1.668 104 1.668 Covariate=Age 1 1.615 1.016 A=Age Environment 1 .239 .151 1 .074 .047 C=Spouse Survival 1 3.326 2.092 1 4.937 3.131 A X C Interaction 1 6.286 3.953* 1 6.660 4.224* Residual 100 1.590 101 1.577 Total 104 1.668 104 1.668 Covariate=Age 1 1.615 .959 A=Age Environment 1 3.272 1.942 1 2.615 1.547 D=Days of Bed Disability 1 .168 .100 1 .089 .052 A X 0 Interaction 1 .016 .009 1 .034 .020 Residual 100 1.684 101 1.690 Tota1 104 1.668 104 1.668 Covariate=Age A=Age Environment E=Independence in Activ- ities of Dailnyiving A X E Interaction Residual Tota1 Covariate=Age A=Age Environment F=Financia1 Ability A X F Interaction Residual Tota1 “p < .01 *p.< .05 189 TABLE 66: Data Results Not Reported for Younger Sample--(H01C,A,D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Va1ue Covariate=Age 1 14.861 7.478** A=Age Environment 1 .006 .003 1 1.463 .701 B=Residentia1 Location of Primary Care Giver 1 .367 .185 1 .666 .319 A X B Interaction 1 .027 .014 1 .054 .026 Residual 111 1.987 112 2.087 Total 115 2.051 115 2.051 Covariate=Age 1 14.861 7.678** A=Age Environment 1 .068 .035 1 .658 .326 §3§pouse Surviva1 1 2.625 1.356 1 2.889 1.429 A X C Interaction 1 3.502 1.809 1 5.178 2.561 Residual 111 1.935 112 2.022 Total 115 2.051 115 2.051 Covariate=Age A=Age Environment D=Days of Bed Disability A X 0 Interaction Residual Total Covariate=Age A=Age Environment E=Independence in Activ- ities of Daily Living A X E Interaction Residual Tota1 Covariate=Age 1 14.111 7.243** A=Age Environment 1 .561 .288 1 .088 .043 F=Financial Ability, 1 8.192 4.205* 1 7.982 3.890 A X F Interaction l .002 .001 l .034 .017 Residual 106 1.948 107 2.052 Total 110 2.080 110 2.080 “p < .01 *p_< .05 190 TABLE C7: Data Results Not Reported for Younger Sample--(H02C,A,D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Va1ue Freedom Square Va1ue Covariate=Prior Score 1 37.930 18.098** Covariate=Age 1 1.101 .526 A=Age Environment 1 .311 .148 1 .129 .051 B=Residential Location of Primary Care Giver 1 1.222 .583 1 2.084 .829 A X 8 Interaction 1 2.490 1.188 1 8.277 3.294 Residual 78 2.096 84 2.513 Total 83 2.573 87 2.547 Covariate=Prior Score 1 '37.930 18.100** Covariate=Age 1 1.101 .526 A=Age Environment 1 .427 .204 l .175 .068 C=Spouse Survival 1 .838 .400 1 1.418 .550 A X C Interaction 1 2.888 1.378 1 3.494 1.355 Residual 78 2.096 84 2.578 Total 83 2.573 87 2.547 Covariate=Prior Score 1 37.930 19.184** Covariate=Age 1 1.101 .557 A=Age Environment 1 .486 .246 1 .322 .141 D=Days of Bed Disability 1 12.960 6.555* 1 29.792 13.076** A X 0 Interaction l .000 .000 1 .257 .113 Residual 78 1.977 84 2.278 Total 83 2.573 87 2.547 Covariate=Prior Score 1 37.930 18.704** Covariate=Age 1 1.101 .543 A=Age Environment 1 .222 .110 1 .148 .064 E=Independence in Activ- ities of Daily Living_ 1 8.810' 4.344** 1 26.148 11.264** A X E Interaction 1 .191 .094 l .297 .128 Residual 78 2.028 84 2.321 Total 83 2.573 87 2.547 Covariate=Prior Score 1 37.930 17.709** Covariate=Age 1 1.101 .514 A=Age Environment 1 .304 .142 F=Financia1 Ability 1 .000 .000 A X F Interaction 1 .114 .053 Residual 78 2.142 Total 83 2.573 *fp' < .01 *2. < .05 191 TABLE C8: Data Results Not Reported for Younger Sample--(Ho3C,A,D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Value Freedom Square Va1ue Covariate=Prior Score 1 37.930 17.970** Covariate=Age 1 1.101 .522 A=Age Environment 1 .417 .198 1 .168 .064 B=Residential Location of Primary Care Giver 1 .594 .281 1 .724 .277 A X 8 Interaction 1 1.950 .924 1 .772 .295 Residua1 78 2.111 84 2.618 Total 83 2.573 87 2.547 Covariate=Prior Score 1 37.930 18.057** Covariate=Age 1 1.101 .524 A=Age Environment 1 .401 .191 1 .174 .067 C=Spouse Survival 1 .671 .319 l .438 .168 AgX C Interaction 1 2.662 1.267 1 2.114 .811 Residual 78 2.101 84 2.606 Total 83 2.573 87 2.547 Covariate=Prior Score Covariate=Age A=Age Environment D=Days of Bed Disability A X 0 Interaction Residual Tota1 Covariate=Prior Score 1 37.930 17.999** Covariate=Age 1 1.101 .523 A=Age Environment 1 .697 .331 E=Independence in Activ- ities of Daily Living 1 1.644 .780 A X E Interaction 1 1.167 .554 Residual 78 2.107 Total 83 2.573 Covariate=Prior Score 1 38.168 17.745** Covariate=Age 1 1.267 .589 A=Age Environment 1 .344 .160 1 .458 .173 F=Financial Ability, 1 .017 .008 1 1.912 .724 A X F Interaction 1, 1.067 .496 1 .181 .068 Residual 77 2.151 83 2.643 Tota1 82 2.604 86 2.577 **p_ < .01 f2' < .05 192 TABLE C9: Data Results Not Reported for Younger Sample--(Ho4C,A,D). Analysis of Covariance Analysis of Variance Degrees of Mean F Degrees of Mean F Source of Variation Freedom Square Va1ue Freedom Square Value Covariate=Age 13.603 6.940** A=Age Environment .055 .028 1 .926 .441 B=Residential Location of Primary Care Giver 1.682 .858 1 1.216 .579 A X 8 Interaction 4.684 2.390 1 4.416 2.103 Residual 1.960 83 2.100 Total 2.101 86 2.101 Covariate=Age 13.603 6.762* A=Age Environment .009 .005 1 .780 .364 C=Spouse Surviva1 .165 .082 1 .011 .005 A X C Interaction 1.954 .971 1 1.928 .899 Residual 2.012 83 2.145 Total 2.101 86 2.101 Covariate=Age 13.603 7.060** A-Age Environment .008 .004 1 .475 .234 D=Days of Bed Disability 4.522 2.347 1 7.767 3.823 A X 0 Interaction 4.569 2.371 1 3.565 1.755 Residua1 1.927 83 2.031 Total 2.101 86 2.101 Covariate=Age A=Age Environment E=Independence in Activ- ities of Dain Living A X E Interaction Residua1 Total Covariate=Age A=Age Environment F=Financia1 Ability A X F Interaction Residual Tota1 **p_< .01 *p.< .05 1|!- 193 mo. v m..." S. v we... Page» _m:uwmmm acme:ocw>cm mm< u < Logos; mm< u mymwcm>oo Noeozm Fmpok _m=uwmmm “coacogw>:u mm< u < Louumm mm< u wumwgm>ou maoum Lowe; u mpmwcm>oo «one: new omozq mmm.~ map Pouch oam.~ Pap Fmsqumm mam.m _om.m _ Sewagoaa>=m mm< u < aoauaa mom. woe. _ mm< u aaaaaa>ou NQFOIM w3~m> wcmscm sovmmcm ms—m> mcmacm Eoummcu cowpmwam> mo mugzom m com: mo mmmcmmo u com: mo mmmcmmo mucmwcm> mo mwm>Pmc< mucmwco>ou mo mwm»_mc< .mpaEmm Lmu_oinmmma—mc< amzimcoiiumucoamm uoz mapzmwm mama "o—u mFQmH LIST OF REFERENCES 194 LIST OF REFERENCES Adams, 0. L. Analysis of a life satisfaction index. Journal of Gerontology, 1969, 24, 470-74. Adams, 0. L. Correlates of satisfaction among the elderly. Ifle_ Gerontologist, 1971, 14, 64-8. Birren, J. E., Noodruff, D. S., and Bergman, S. Research, demonstra- tion, and training: Issues and methodology in social gerontology. The Geronto1ogist, 1972, 1g, 49-83. Blenkner, M. "A disengagement theory of agigg?--Cumming, Dean and Newell. Paper presented at the meeting of the GerontoIogi- cal Society, Philadelphia, November, 1958. Blenkner, M., Bloom, M., and Weber, R. Protective services for older pegple: Final report on the planningtphase of a research and demonstration project. Cleveland: Benjamin Rose Institute, Research Department, 1964. Bloom, M. Discontent with Contentment Scales. Paper presented at the 1975 meeting of the Gerontological Society, Portland, Oregon. Bortner, R. w. Adult development or idiosyncratic change? A plea for the developmental approach. The Gerontologist, 1966, 6, 159-64. Bortner, R. R., and Hultsch, D. F. A multivariate analysis of correlates of life satisfaction in adulthood. Journal of Gerontology, 1970, 25, 41-7. Brown, A. S. Satisfying relationships for the elderly and their patterns of disengagement. The Gerontologist, 1974, 14, 258-62. Carp, F. M. Some components of disengagement. Journal of Geron- tology, 1968, 23, 382-6. Carp, F. M. Compound criteria in gerontological research. Journal of Gerontology, 1969, 24, 341-7. Cavan, R., Burgess, E., Havighurst, R., and Goldhamer, H. Personal adjustment in old age. Chicago: Science Research Associ- ates, 1949. 195 196 Cumming, E. Further thoughts on the theory of disengagement. lpter- national Social Science Bulletin, 1963, 15, 377-93. Cumming, E., Dean, L., and NeWell, D. A disengagement theory of agi 9. Kansas City, Mo., October, 1958 (mimeographed). Cumming, E., Dean, L., Newell, D., and McCaffrey, I. Disengagement-- A tentative theory of aging. Sociomentry, 1960, 22, 23-35. Cumming, E., and Henry, W. Growing_old: Theyprocess of diseggage- ment. New York: Basic Books, 1961. Edwards, J. N., and Klemmack, D. L. Correlates of life satisfaction: A re-examination. Journal of Gerontology, 1973, 28, 497-502. Goodman, M., Dye, D., Harel, B., and Bley, N. Perception of economic status as a determinant of morale in the aged. (Program Notes 1971 Annual Meeting of the Gerontological Society, Houston). The Gerontologist, 1971, ll, 34. Gubrium, J. F. Environmental effects on morale in old age and the resources of health and solvency. The Gerontologist, 1970, 10, 294-7. Gubrium, J. F. Toward a socio-environmental theory of aging. Ihg_ Gerontologist, 1972, 12, 281-4. Havighurst, R. J. Successful aging. The Gerontologist, 1961, 1, 8-13. . Havighurst, R. J. Successful aging. In R. H. Williams, c. Tibbits, and W. Donahue (Eds.), Processes of aging (Vol. 1). New York: Atherton Press, 1963. Havighurst, R. J. Personality and patterns of aging. The Gerontolo- gis , 1968, a, 20-23. Havighurst, R. J., Neugarten, B. L., and Tobin, S. Disengagement and patterns of aging. In B. L. Neugarten (Ed.), Midd1e_ age and aging: A reader in socia1_psychology. Chicago: The Universityof Chicago Press, 1968. Henry, W. A theory of intrinsic disengagement. Paper read at the International Social Science Seminar, Sweden, August, 1963 (mimeographed). Katz, S., Ford, A. 8., Downs, T. 0., Adams, N., and Rusby, D. I. Effects of continued care: A studyyof chronic illness in the home (DHEW Publication No.71HSM) 73-3010). Washington, D.C.: Department of Health, Education, and Welfare, Health Services and Mental Health Administration, National Center for Health Services Research and Development, 1972. 197 Katz, 5., Ford, A. B., Moskowitz, R. W., Jackson, 8. S., and Jaffe, M. W. Studies of illness in the aged: The index of ADL, a standardized measure of biological and psychosocial function. Journal of the American Medical Association, 1963, 185, 914-19. Kutner, B., Fanshel, D., Togo, A. M., and Langner, T. S. Five hundred over sixty. New York: Russell Sage Foundation, 1956. Lawton, M. P. The dimensions of morale. In D. P. Kent, R. Kasten- baum, and S. Sherwood (Eds.), Research planning and action for the elderly. New York: Behavioral Publications, Inc., 1972. Lemon, 8. W., Bengtson, V.L”, and Peterson, J. A. An exploration of the activity theory of aging: Activity types and life satisfaction among in-movers to a retirement community. Journal of Gerontology, 1972, 21, 511-23. Lipman, A., and Smith, K. J. Functionality of disengagement in old age. Journal of Gerontology, 1968, 22, 517-21. Loeb, M., Pincus, A., and Mueller, J. Growing Old in Rural Wisconsin. University of Wisconsin School of Social Work, Madison, Wisconsin, 1963. Loeb, M. B., Pincus, A., and Mueller, 8. J. A framework for viewing adjustment in aging. The Gerontologist, 1966, 6, 185-87. Lowenthal, M. F. Geriatric illness: Some shortcomings of isolation theory. In E. W. Busse (Chair.) and F. C. Jeffers (Ed.), Duke University, Council on Gerontology: Proceedings of seminars, 1961-65. Durham, N. C.: Duke University, Regional Center for the Study of Aging, 1965. Lowenthal, M. F., and Boler, 0. Voluntary vs. involuntary social withdrawal. Journal of Gerontology, 1965, 29, 363-71. Maddox, G. L. Activity and morale: A longitudinal study of selected elderly subjects. Social Forces, 1963, 42, 195-204. Maddox, G. L. Disengagement theory: A critical evaluation. ‘Iflg Gerontologist, 1964, 4, 80-83. Maddox, G. L. Fact and artifact: Evidence bearing on disengagement theory from the Duke geriatrics project. Human Develgpment, 1965, §, 117-30. Martin, W. C. Activity and disengagement: Life satisfaction of in-movers into a retirement community. The Gerontologist, 1973, 12, 224-27. 198 Neugarten, B. L. Personality and the aging process. The Gerontolo- gis , 1972, 42, 9-15. Neugarten, B. L., Havighurst, R. J., and Tobin, S. S. Personality and patterns of aging. In B. L. Neugarten (Ed.), Middle age and aging: A reader in social psychology. Chicago: The University of Chicago Press, 1968. Palmore, E., and Luikart, C. Health and social factors related to life satisfaction. Journal of Health and Social Behavior, 1972, 12, 68-80. Prasad, S. B. The retirement postulate of the disengagement theory. The Gerontologist, 1964, 4, 20-23. Rose, A. M. A current theoretical issue in social gerontology. Ihe_ Gerontologist, 1964, 4, 46-50. Rosow, I. Social integration of the aged. New York: The Free Press, 1967. Schooler, K. Effect of environment on morale. The Gerontologist, 1970, 19, 194-197. Sherwood, S., and Nadelson, T. Alternate predictions concerning despair in old age. In D. P. Kent, R. Kastenbaum, and S. Sherwood (Eds.), Research planning and action for the elderly. New York: Behavioral Publications, Inc., 1972. Spreitzer, E., and Synder, E. E. Correlates of life satisfaction among the aged. Journal of Gerontology, 1974, 29, 454-58. Tallmer, M., and Kutner,B. [Msengagement and morale. The Gerontolo- 91 t: 19709 Lg: 317-20. Thompson, G. 8. Work versus leisure roles: An investigation of morale among employed and retired men. Journal of Geron- tology, 1973, 28, 339-44. Tibbitts, C., and Schmelzer, J. L. New directions in aging and their research implications. Welfare Revue, 1965, 2, 8-14. Tissue, T. L. A Guttman scale of disengagement potential. Journal of Gerontology, 1968, 22, 513-16. Tissue, T. L. Disengagement potential: Replication and use as an explanatory variable. Journal of Gerontology, 1971, 26, 76-80. Tissue, T. L. Another look at self-rated health among the elderly. Journal of Gerontology, 1972, 2], 91-94. 199 Tobin, S. S., and Neugarten, B. L. Life satisfaction and social interaction in the aging. Journal of Gerontology, 1961, 1g, 344-46. - U. S. Department of Health, Education, and Welfare. Age Patterns in Medical Care, Illness, and Disability: United States, 1968-1969, Vital and Health Statistics Data from the National Health Survey, Series 10 - Number 70, DHEW Publication No. (HSM) 72-1026. Public Health Service, Health Services and Mental Health Administration, Rockville, Maryland, 1972, p. 30. U. S. Department of Health, Education, and Welfare. Mortality Trends: Age, Color, and Sex, United States--1950-69, Vital and Health Statistics Data from the National Vital Statistics System, Series 20 - Number 15, DHEW Publication No. (HRA) 74-1852. Public Health Service, Health Resources Administration, Rockville, Maryland, 1973, pp. 20 and 30. U. S. Department of Health, Education, and Welfare, Utilization of Short-Stay Hospitals: Summary of Nonmedical Statistics: United States--l972, Vital and Health Statistics Data from the National Hea1th Survey, Series 13--Number 19, DHEW Publication No. (HRA) 75-1770. Public Health Service, Health Resources Administration, Rockville, Maryland, 1975, p. 3. Winer, B. J. Statistical principles in eyperimental design. New York: McGraw-Hill Book Company, Inc., 1962. Wood, V., Wylie, M. L., and Sheafor, B. An analysis of a short self-report measure of life satisfaction: Correlation with rater judgements. Journal of Gerontolggy, 1969, 24, 465-69. Youmans, E. G. Some perspectives on disengagement theory. Ihe_ Gerontologist, 1969, 9, 254-58. Youmans, E. G. Age group, health and attitudes. The Gerontologist, 1974, up 249-54. "mama“