{"http:\/\/dx.doi.org\/10.14288\/1.0087810":{"http:\/\/vivoweb.org\/ontology\/core#departmentOrSchool":[{"value":"Arts, Faculty of","type":"literal","lang":"en"},{"value":"Sociology, Department of","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/dataProvider":[{"value":"DSpace","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeCampus":[{"value":"UBCV","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/creator":[{"value":"Lehman, Allen J.","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/issued":[{"value":"2009-03-16T20:05:01Z","type":"literal","lang":"en"},{"value":"1996","type":"literal","lang":"en"}],"http:\/\/vivoweb.org\/ontology\/core#relatedDegree":[{"value":"Master of Arts - MA","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeGrantor":[{"value":"University of British Columbia","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/description":[{"value":"The death of a friend and the death of a spouse frequently\r\naccompany later life . The former has received scant empirical\r\nattention and the bereavement overload occasioned by the\r\ncombination of these losses only has been addressed in the\r\nconceptual literature. This study used data from the Canadian\r\nnational Survey on Ageing and Independence (1991), and examined\r\nthe short-term reactions of women and men aged 65 and older to\r\nthe death of a close friend, a spouse, or both, in comparison\r\nwith a non-bereaved group. Drawing on Weiss' (1993) conceptual\r\nperspective on loss, bereaved individuals' \"recovery\" or\r\nadjustment to loss was assessed on multiple dimensions of\r\neffective personal and social functioning: perceived health,\r\nnegative affect, positive affect, social involvement, emotional\r\ninvestment, family satisfaction, and friend satisfaction . Based\r\non Weiss* (1993) model that maintains the loss of a relationship\r\nof attachment (i.e., spouse) evokes more intense bereavement\r\nreactions than the loss of a relationship of community (i.e.,\r\nclose friend), it was predicted that the negative effects of\r\nbereavement would be the greatest for the spouse bereaved, which\r\nwould be significantly greater than the peer bereaved, which\r\nwould be significantly greater than the non-bereaved. A further\r\nexpectation was that reactions of the multiple bereaved\r\nindividuals would be equal to or greater than the spouse\r\nbereaved. Several expected as well as unexpected findings were\r\nrevealed. Four primary patterns of findings included: (a) only\r\nfor the measure of negative affect were the bereavement status\r\ncomparisons consistent with predictions; (b) spouse and multiple bereaved individuals' indicated comparable levels of functioning;\r\n(c) the peer bereaved unexpectedly reported greater effective\r\npersonal and social functioning than the non-bereaved; and (d)\r\ngender influenced levels of functioning, with women tending to\r\nindicate better functioning than men. This study provides a\r\ngreater understanding of older adults' reactions to the loss of\r\nclose interpersonal relationships and sheds light on the nature\r\nand meaning of close interpersonalties in later life .\r\nSuggestions are offered for future research.","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/aggregatedCHO":[{"value":"https:\/\/circle.library.ubc.ca\/rest\/handle\/2429\/6072?expand=metadata","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/extent":[{"value":"6396091 bytes","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/elements\/1.1\/format":[{"value":"application\/pdf","type":"literal","lang":"en"}],"http:\/\/www.w3.org\/2009\/08\/skos-reference\/skos.html#note":[{"value":"SHORT-TERM REACTIONS TO THE DEATH OF A SPOUSE AND\/OR CLOSE FRIEND IN LATER LIFE by ALLEN J. LEHMAN B.A., University of C a l i f o r n i a , Irvine, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES The School of Family and N u t r i t i o n a l Sciences Family Studies We accept t h i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October, 1996 \u00a9 A l l e n J. Lehman, 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholariy purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada DE-6 (2\/88) 11 Abstract The death of a f r i e n d and the death of a spouse frequently accompany l a t e r l i f e . The former has received scant empirical attention and the bereavement overload occasioned by the combination of these losses only has been addressed i n the conceptual l i t e r a t u r e . This study used data from the Canadian national Survey on Ageing and Independence (1991), and examined the short-term reactions of women and men aged 65 and older to the death of a close friend, a spouse, or both, i n comparison with a non-bereaved group. Drawing on Weiss' (1993) conceptual perspective on loss, bereaved individuals' \"recovery\" or adjustment to loss was assessed on multiple dimensions of e f f e c t i v e personal and s o c i a l functioning: perceived health, negative a f f e c t , p o s i t i v e affect, s o c i a l involvement, emotional investment, family s a t i s f a c t i o n , and f r i e n d s a t i s f a c t i o n . Based on Weiss* (1993) model that maintains the loss of a r e l a t i o n s h i p of attachment ( i . e . , spouse) evokes more intense bereavement reactions than the loss of a relationship of community ( i . e . , close f r i e n d ) , i t was predicted that the negative e f f e c t s of bereavement would be the greatest for the spouse bereaved, which would be s i g n i f i c a n t l y greater than the peer bereaved, which would be s i g n i f i c a n t l y greater than the non-bereaved. A further expectation was that reactions of the multiple bereaved in d i v i d u a l s would be equal to or greater than the spouse bereaved. Several expected as well as unexpected findings were revealed. Four primary patterns of findings included: (a) only for the measure of negative a f f e c t were the bereavement status comparisons consistent with predictions; (b) spouse and multiple I l l bereaved i n d i v i d u a l s ' indicated comparable l e v e l s of functioning; (c) the peer bereaved unexpectedly reported greater e f f e c t i v e personal and s o c i a l functioning than the non-bereaved; and (d) gender influenced l e v e l s of functioning, with women tending to indicate better functioning than men. This study provides a greater understanding of older adults' reactions to the loss of close interpersonal relationships and sheds l i g h t on the nature and meaning of close interpersonal t i e s i n l a t e r l i f e . Suggestions are offered for future research. i v Table of Contents Abstract. i i L i s t of Tables v i i L i s t of Figures ix Acknowledgment x Dedication x i Chapter I. Introduction 1 D e f i n i t i o n of Terms 3 Bereavement Theories 5 Review of the Literature 9 Death of a Spouse 9 Empirical: Psychological and physical health consequences 11 Conceptual: Spouse\/partner loss i n l a t e r l i f e . 14 Death of a Close Friend 16 Empirical: Death of a fri e n d i n l a t e r l i f e . . . . 16 Conceptual: Death of a f r i e n d i n l a t e r l i f e . . . 18 Conceptual: Friendship i n l a t e r l i f e 19 Multiple Bereaved 21 Empirical 21 Conceptual 22 Summary 23 Theoretical Framework 24 Hypotheses 26 I I . Method 3 5 Sample 3 5 Measures 39 Dependent Variables 39 V Independent Variables 44 Control Variables 44 Analyses 46 I I I . Results 48 Univariate Distributions 48 Hypothesis Testing 48 Perceived Health: \"giving energy to everyday l i f e \" 50 Negative Affect: \"freedom from disturbing thoughts and f e e l i n g s \" 51 Positive Affect: \"to experience g r a t i f i c a t i o n \" 54 Social Involvement: \"to function with reasonable adequacy i n s o c i a l r o l e s \" 60 Emotional Investment: \"to function with reasonable adequacy i n s o c i a l r o l e s \" 63 Family S a t i s f a c t i o n : \"to function with reasonable adequacy i n s o c i a l r o l e s \" 66 Friend S a t i s f a c t i o n : \"to function with reasonable adequacy i n s o c i a l r o l e s \" 69 Regression Analyses 72 Perceived Health 73 Negative Affect 73 Positive Affect 77 Social Involvement 77 Emotional Investment 78 Family S a t i s f a c t i o n 78 Friend S a t i s f a c t i o n 79 v i Exploratory Analyses of the Peer Bereaved 79 Perceived Health 81 Negative Affect 81 Positive A f f e c t 83 Social Involvement 85 Emotional Investment 88 IV. Discussion 90 Bereaved and Non-bereaved Adults\u2022 E f f e c t i v e Functioning 94 Spouse and non-bereaved comparison 94 Peer and non-bereaved comparison 97 Multiple and non-bereaved comparison 100 Comparison between multiple, spouse, and peer bereaved 104 Context of Loss ....108 Weiss' Framework on Loss I l l Limitations 114 Conclusions and Implications 117 References 122 Appendices 13 0 v i i L i s t of Tables Table 1. Characteristics of bereaved and non-bereaved adults 38 Table 2. Analysis of covariance on perceived health 52 Table 3. Adjusted means and standard deviations of perceived health for bereaved and non-bereaved adults 53 Table 4. Analysis of covariance on negative a f f e c t 55 Table 5. Adjusted means and standard deviations of negative a f f e c t for bereaved and non-bereaved adults 56 Table 6. Analysis of covariance on p o s i t i v e a f f e c t 58 Table 7. Adjusted means and standard deviations of p o s i t i v e a f f e c t for bereaved and non-bereaved adults 59 Table 8. Analysis of covariance on s o c i a l involvement 61 Table 9. Adjusted means and standard deviations of s o c i a l involvement for bereaved and non-bereaved adults.... 62 Table 10. Analysis of covariance on emotional investment 64 Table 11. Adjusted means and standard deviations of emotional investment for bereaved and non-bereaved adults 65 Table 12. Analysis of covariance on s a t i s f a c t i o n with family relationships 67 Table 13. Adjusted means and standard deviations of s a t i s f a c t i o n with family relationships for bereaved and non-bereaved adults 68 Table 14. Analysis of covariance on s a t i s f a c t i o n with friends. 70 Table 15. Adjusted means and standard deviations of s a t i s f a c t i o n with friends for bereaved and non-bereaved adults 71 v i i i Table 16. Standardized beta weights for predictors of dependent measures by bereavement status 74 Table 17. Analysis of covariance on perceived health for peer bereaved adults (N = 3105) 82 Table 18. Analysis of covariance on negative a f f e c t for peer bereaved adults (N = 3 067) 84 Table 19. Analysis of covariance on p o s i t i v e a f f e c t for peer bereaved adults (N = 3 053) 86 Table 20. Analysis of covariance on s o c i a l involvement for peer bereaved adults (N = 3095) 87 Table 21. Analysis of covariance on emotional investment for peer bereaved adults (N = 3 029) 89 IX L i s t of Figures Figure 1. Hypothesis t e s t i n g of bereavement status differences on measures of e f f e c t i v e functioning.... 92 Figure 2. Hypothesis testing of gender differences on measures of e f f e c t i v e functioning 93 X Acknowledgment I would l i k e to sincerely thank Dr. Brian de Vries for the encouragement and enthusiasm he provided throughout my journey. Brian's support and insights were invaluable. As well, I would l i k e to thank the other members of my committee, Dr. Daniel Perlman and Dr. Darrin Lehman, for t h e i r i n s i g h t s . P a r t i c u l a r l y , I would l i k e to express my appreciation to Dan for h i s time and assistance. I wish to express my deepest appreciation to both my family i n C a l i f o r n i a and my family i n B r i t i s h Columbia for t h e i r continuous support, encouragement, and patience. Thank you. x i Dedication I dedicate t h i s thesis to the memory of my father Jim Lehman who, during h i s ba t t l e with Amyotrophic Lateral S c l e r o s i s , taught me about l i v i n g and dying as well as about love, compassion, and i n t e g r i t y . His l i f e continues to make me ask questions and seek answers to \"why?\" 1 Chapter I Introduction The death of a spouse and the death of a f r i e n d are normative interpersonal losses that accompany l a t e r l i f e . Over 65,000 Canadians aged 65 and older became widowed i n 1993 ( S t a t i s t i c s Canada, 1994). One-third of a sample of 10,000 Canadians aged 65 and older experienced the death of a f r i e n d i n 1991 ( S t a t i s t i c s Canada, 1993), leaving an estimated one-million older adults peer bereaved. While research into l a t e r l i f e bereavement primarily has centered on the short-term reactions of i n d i v i d u a l s to widowhood (e.g., Gallagher-Thompson, Futterman, Farberow, Thompson, & Peterson, 1993; Lund, Caserta, & Dimond, 1993), l a t e r l i f e peer bereavement has received scant empirical attention (Roberto & Stanis, 1994) and only a l i m i t e d conceptual focus (Deck & Folta, 1989; Sklar, 1991-1992). In addition, while there i s l i m i t e d conceptual discussion of multiple interpersonal losses i n general (Kastenbaum, 1969; Moss & Moss, 1989), no research has addressed the bereavement reactions to the loss of both a spouse and a close friend i n l a t e r l i f e . The majority of research that examines bereavement reactions of older adults centers on widows and widowers. A review of the e x i s t i n g l a t e r l i f e widowhood research reveals a focus on two main aspects of bereavement reactions: psychological and physical health consequences. Research on psychological health consequences includes examinations of depression and d i s t r e s s (e.g., Bruce, Kim, Leaf, & Jacobs, 1990; Gallagher, Breckenridge, Thompson, & Peterson, 1983), stress and coping (e.g., Caserta & Lund, 1992), g r i e f (e.g., Vezina, Borque, & Belanger) , l i f e 2 s a t i s f a c t i o n (e.g., Lund, Caserta, & Dimond, 1986), and morale (e.g., Scott & Kivett, 1985). Investigations of physical health consequences include examinations of medical i l l n e s s and v i s i t s to physicians (e.g., Thompson, Breckenridge, Gallagher, & Peterson, 1984), and perceived health (e.g., Caserta, Lund, & Dimond, 1990). Notwithstanding t h i s corpus of research, bereavement investigations into these reactions to loss often have lacked the guidance of an integrative t h e o r e t i c a l framework (Stroebe, Stroebe, & Hansson, 1993) that must c e r t a i n l y be multidimensional i n scope. Multidimensional models of bereavement reactions provide an opportunity for the examination of the range of psychological ( i . e . , a f f e c t i v e , cognitive), physical (including behavioural), and s o c i a l reactions of bereaved i n d i v i d u a l s . Weiss' (1993) conceptual writings on loss provide such a multidimensional perspective. Weiss' (1993) perspective delineates s p e c i f i c types of interpersonal losses based on degree of attachment and predicts that the loss of a spouse ( i . e . , r e l a t i o n s h i p of attachment) w i l l evoke intense g r i e f and severe d i s t r e s s , while the loss of a f r i e n d ( i . e . , r e l a t i o n s h i p of community) w i l l evoke bereavement reactions of a lesser magnitude. An exception to the above, Weiss (1993) suggests that i n d i v i d u a l s who lose entire relationships of community (e.g., a l l of one's friends) w i l l experience intense bereavement reactions and face d i f f i c u l t y i n personal and s o c i a l functioning s i m i l a r to i n d i v i d u a l s who lose a relationship of attachment (e.g., spouse). S t i l l , s i m i l a r to so many others i n t h i s respect, Weiss (1993) o f f e r s l i t t l e by which to predict reactions to and characterize 3 multiple interpersonal losses that tend to accompany l a t e r l i f e . M u l tiple dimensions on which individuals are affected are i d e n t i f i e d including negative and p o s i t i v e a f f e c t , cognitive, and behavioral domains. This study was guided by Weiss' (1993) perspective i n examining the short-term reactions to the loss of a spouse, a close f r i e n d , or both, i n l a t e r l i f e . This study provides the f i r s t empirical comparison of both spouse bereavement and peer bereavement i n older women and men. In addition, t h i s study investigates multiple bereavements ( i . e . , loss of both spouse and close friend) and individuals' reactions to multiple interpersonal losses i n comparison to other bereaved and non-bereaved people of similar age. Examination of the bereaved and non-bereaved comparison groups are based on measures of perceived health, negative a f f e c t , p o s i t i v e a f f e c t , s o c i a l involvement, emotional investment, s a t i s f a c t i o n with family r e l a t i o n s h i p s , and s a t i s f a c t i o n with friends. An investigation of these multidimensional facets of bereavement reactions provides a better understanding of older adults' reactions to the loss of close interpersonal relationships, as well as sheds l i g h t on the nature and meaning of interpersonal t i e s in l a t e r l i f e . D e f i n i t i o n of Terms The following sections provide an accounting of the d e f i n i t i o n of terms to be discussed i n t h i s study, as well as an overview of bereavement theories. The l i t e r a t u r e on short-term bereavement reactions to the loss of a spouse, a f r i e n d , and multiple interpersonal losses makes up the substance of the l i t e r a t u r e review that follows, encompassing both an empirical 4 and conceptual discussion of reactions to loss and meaning of the l o s t r e l a t i o n s h i p . The most commonly used d e f i n i t i o n s of terms associated with bereavement research are provided below i n order to c l a r i f y what i s examined. Bereavement refers to \"the objective s i t u a t i o n of having l o s t someone s i g n i f i c a n t \" (Stroebe et a l . , 1993, p. 5). Loss i s used interchangeably with bereavement. Bereaved i s a d e s c r i p t i v e term for an individual who experienced the death of another person. Grief \" i s the emotional response to one's lo s s , \" and mourning \"denotes the actions and manner of expressing g r i e f \" (Stroebe et a l . , 1993, p. 5). The term bereavement reactions r e f e r s to in d i v i d u a l s ' psychological, physical, and s o c i a l functioning as a r e s u l t of the loss. Recovery from or adjustment to bereavement r e s u l t s when an individual i s able to function i n d a i l y l i f e at lev e l s similar to those preceding one's bereavement status (Weiss, 1993). This study examines the short-term reactions to the death of a spouse, a close friend, and both, i n a n a t i o n a l l y representative sample of Canadians aged 65 and older. Short-term i s defined as twelve months or less; thus, reactions to the death of a close interpersonal relationship(s) in the preceding year are investigated. Spouse i s the term used to i d e n t i f y a marriage partner. For the purposes of t h i s study, common-law-partners also are grouped under the category of spouse. Close friend i s subjecti v e l y defined by each ind i v i d u a l , but respondents are primed to think of someone with whom \"an i n d i v i d u a l f e e l s at ease, can t a l k to about private matters, or can c a l l on for help.\" 5 Bereavement Theories The study of g r i e f and bereavement has i t s roots i n the c l a s s i c a l psychoanalytic perspective of Freud (1917\/1957), l a t e r elaborated i n the attachment models of Bowlby (1969, 1980), Parkes (1972, 1993), Parkes and Weiss (1983), and Weiss (1993). The evolution of bereavement research has witnessed a movement from intrapersonal perspectives to interpersonal perspectives (Stroebe, Stroebe, & Hansson, 1988). For example, Freud (1917\/1957) proposed a d i s t i n c t i o n between bereavement reactions: mourning\u2014the normal or more natural emotional reaction to the death of a loved one ( i . e . , normal g r i e f ) \u2014 a n d melancholia\u2014a more pathological reaction with persistent depression following the death of a loved one ( i . e . , chronic or c l i n i c a l depression). This intrapersonal perspective emphasized that the bereaved i n d i v i d u a l must \"work through the loss\" and detach or r e l i n q u i s h his\/her psychological attachment to the deceased by withdrawing energy from the image of the deceased. Freud (1917\/1957) believed that af t e r the bereaved in d i v i d u a l had \"worked through the l o s s , \" he\/she then would have the emotional energy to e s t a b l i s h new relationships. Bowlby's (1980) attachment perspective on g r i e f moved beyond Freud's (1917\/1957) focus on the i n d i v i d u a l and into a broader, more interpersonal perspective (Stroebe et a l . , 1988). Central to Bowlby's theory was the relationship between c h i l d and parent i n which the c h i l d exerts attachment behaviours i n order to \"maintain c e r t a i n degrees of proximity to, or communication with, the discriminated attachment f i g u r e ( s ) \" (1969, p. 40). With a c h i l d ' s loss of an attachment relationship ( i . e . , parent), the 6 c h i l d no longer f e e l s protected and secure. Bereavement i s an unwanted separation from an attachment figure that causes separation anxiety (Bowlby, 1969) and feelings of anger and depression. Unlike Freud, Bowlby believed that these attachment behaviours were normal for bereaved individuals i n response to t h e i r desire for a reunion with the deceased. Bowlby (1980) i d e n t i f i e d four phases that bereaved individuals experience i n reaction to the death of an attachment figure: numbness, yearning and protest, depressive withdrawal, reorganization and recovery. Recovery occurs when the bereaved in d i v i d u a l invests i n new r e l a t i o n s h i p s and has a return of interests that preceded the lo s s . Parkes (1972, 1993; Parkes & Weiss, 1983) extended Bowlby*s (1969) attachment theory into adulthood and imbued i t with a more cognitive orientation. He i d e n t i f i e d losses that included not only the loss of a spouse or another loved one, but also the loss of a limb and losses that arise with a terminal i l l n e s s . The experience of these losses occasion psychosocial t r a n s i t i o n s (Parkes, 1993), i n which individuals are forced to readjust t h e i r assumptions about the world i n coping with the loss. For example, \"[t]he death of a spouse invalidates assumptions that penetrate many aspects of l i f e . Habits of action...and thought...must be revised i f the survivor i s to l i v e as a widow\" (Parkes, 1993, p. 94). This coping and readjustment may negatively a f f e c t mental and physical health, p a r t i c u l a r l y when there i s resistance to change following the loss (Parkes, 1993). Individuals' reactions and recovery may vary depending on the type of loss, for each loss brings with i t an unique meaning for 7 each i n d i v i d u a l (Parkes, 1993). Like Bowlby, Parkes (1993) noted that i n d i v i d u a l s go through phases of grieving\u2014numbness, pining, disorganization and despair, and recovery. Recovery r e s u l t s when the bereaved i n d i v i d u a l has both accepted the loss and adjusted or revised her\/his assumptions about the world. Weiss' (1993) perspective on loss draws on Bowlby's (1969, 1980) development of attachment theory and Parkes 1 (1972) extension of i t into adult functioning. Rooted i n Bowlby's attachment theory, Weiss (1993) delineates two categories of r e l a t i o n s h i p s \u2014 r e l a t i o n s h i p s of attachment and re l a t i o n s h i p s of community. Weiss' (1993) relationships of attachment include r e l a t i o n s h i p s between married couples or partners, parent-child r e l a t i o n s h i p s , transference relationships (bonds between patient and t h e r a p i s t ) , and some relationships between parents and t h e i r adult children. His second category of r e l a t i o n s h i p s - -r e l a t i o n s h i p s of community\u2014includes friendships, co-workers and colleagues, adult s i b l i n g relationships and other non-household f a m i l i a l r e l a t i o n s h i p s . Weiss' (1993) perspective on loss holds that only the loss of an attachment relationship (e.g., spouse) w i l l evoke a g r i e f response and intense d i s t r e s s , whereas the loss of a r e l a t i o n s h i p of community (e.g., close friend) w i l l evoke much les s d i s t r e s s and less negative bereavement reactions. Whereas attachment re l a t i o n s h i p s provide feelings of security and are not replaceable, Weiss (1993) contends that the loss of a f r i e n d or re l a t i o n s h i p of community may be replaceable and evokes less intense bereavement reactions: \"membership i n a meaningful community w i l l l i m i t the d i s t r e s s \" (Weiss, 1993, p. 272). I t i s 8 only with the \"loss of a l l of one type of r e l a t i o n s h i p of community-all friendships...,\" notes Weiss (1993, p. 271), that intense bereavement reactions are experienced s i m i l a r to those of ind i v i d u a l s who lose a spouse or other r e l a t i o n s h i p of attachment. Following the death of a loved one, in d i v i d u a l s experience phases of grieving and, in the \"movement to recovery\" or adaptation, go through a process of cognitive acceptance, emotional acceptance, and identity change (Weiss, 1993). Weiss i d e n t i f i e s cognitive, a f f e c t i v e , and behavioural dimensions that are impacted following the death of an attachment r e l a t i o n s h i p . Bereaved i n d i v i d u a l s ' recovery following loss r e s u l t s when they are able to e f f e c t i v e l y function as adults; t h i s includes the a b i l i t y to (1) give energy to everyday l i f e , (2) maintain psychological comfort, (3) experience g r a t i f i c a t i o n , (4) make preparations for the future, and (5) function i n s o c i a l roles (Weiss, 1993). This perspective on loss o f f e r s an a r t i c u l a t i o n of multiple types of loss and the degree to which they evoke intense bereavement reactions as well as provides multiple dimensions on which to assess bereaved i n d i v i d u a l s ' recovery and le v e l s of e f f e c t i v e functioning. While others have studied bereavement reactions from a va r i e t y of d i s c i p l i n e s including stress theory (e.g., Stroebe & Stroebe, 1987) and symbolic interactionism (Rosenblatt, 1993), Weiss' (1993) attachment.perspective i s the most germane to t h i s study. Weiss' (1993) framework provides an interpersonal perspective that i d e n t i f i e s p a r t i c u l a r types of losses and delineates multiple dimensions of bereavement reactions. 9 Review of the Literature Death of a Spouse The following two sections provide a review of the l i t e r a t u r e on older adults' reactions to the death of a spouse. The review of the empirical l i t e r a t u r e i s r e s t r i c t e d to the short-term reactions of older adults, and i s organized by the two core areas of bereavement research i n v e s t i g a t i o n \u2014 p s y c h o l o g i c a l and physical health consequences. This empirical discussion i s followed by a conceptual discussion of the meaning of a spouse i n l a t e r l i f e which offers a context for understanding the bereavement reactions to the loss of a spouse. The early examination of bereavement was e s s e n t i a l l y the study of widowhood and tended to be problem-generated (Stroebe et a l . , 1988); that i s , i t focused on single issues. For example, early researchers investigated the r e l a t i o n s h i p between widowhood and increased mortality rates i n spouse bereaved i n d i v i d u a l s (Farr, 1858: c i t e d in Stroebe et a l . , 1988; S h u r t l e f f , 1955). Lindemann (1944) examined negative a f f e c t following bereavement. Research was often limited to young and middle-aged widows (e.g., Parkes, 1964). The studies that have focused on l a t e r l i f e spousal bereavement have been composed of small and unrepresentative samples (e.g., Heyman & Gianturco, 1973), and did not compare individuals' bereavement reactions to non-bereaved individuals (e.g., Caserta, Lund, & Dimond, 1989). Consequently, sample selection often limited the g e n e r a l i z a b i l i t y of findings, while the lack of control groups minimized the understanding of bereavement reactions. Furthermore, these and other studies tended to focus on women under age 60, l i m i t i n g the 10 understanding of bereavement reactions (and interpersonal t i e s ) i n l a t e r l i f e women and men where the death of a spouse i s s t a t i s t i c a l l y more common. More recent research has moved away from the single issue-focused research on mortality or depression (with some minor exception, e.g., Bruce et a l . , 1990) and toward l i m i t e d t h e o r e t i c a l l y - d r i v e n research (e.g., Caserta et a l . , 1989; Reich, Zautra, & Guarnaccia, 1989). Researchers now examine widowhood i n l a t e r l i f e (e.g. Lund's, 1989, edited book, Older bereaved spouses: Research with p r a c t i c a l implications), and recently have begun to establish and refin e t h e o r e t i c a l frameworks (see Stroebe et a l . , 1993) that recognize the \"multidimensionality of bereavement reactions\" (Stroebe et a l . , 1988) . Recently, more intensive attention has been paid to l a t e r l i f e widowhood (e.g., Lund, 1989; Thompson, Gallagher-Thompson, Futterman, Gilewski, & Peterson, 1991) with larger and more representative samples including non-bereaved comparison groups (e.g., Breckenridge, Gallagher, Thompson, & Peterson, 1986; Caserta & Lund, 1992). These studies address some of the methodological l i m i t a t i o n s of e a r l i e r research and, a d d i t i o n a l l y , provide a growing body of l i t e r a t u r e from which to examine the short-term bereavement reactions to the loss of a spouse i n l a t e r l i f e i n d i v i d u a l s . The following l i t e r a t u r e review of widowhood examines the psychological ( i . e . , a f f e c t i v e , cognitive) and physical functioning consequences of l a t e r l i f e bereavement. 11 Empirical; Psychological and physical health consequences. Many studies have examined the psychological health consequences following widowhood focusing on the dependent variable of depression i n the bereaved between 1 month and 1 year following the loss (Breckenridge et a l . , 1986; Bruce et a l . , 1990; Carey, 1977; Caserta et a l . , 1989; Gallagher et a l . , 1983; Reich et a l . , 1989) as well as throughout the f i r s t 2 years following the loss (Lund et a l . , 1986; Vezina et a l . , 1988; Zisook, Schuchter, Sledge, Paulus, & Judd, 1994). Consistently, studies comparing spouse bereaved with control or matched samples of married and non-bereaved older adults indicate that bereaved samples evidence deleterious e f f e c t s following the loss. For example, Breckenridge et a l . (1986) examined the bereavement reactions of 196 i n d i v i d u a l s at 2 months following the loss of t h e i r spouse. As compared to a non-bereaved comparison group of 145 older adults, the bereaved group had s i g n i f i c a n t l y higher l e v e l s of d i s t r e s s . In an e a r l i e r study by Gallagher et a l . (1983) that examined bereavement reactions 2 months following the loss, 211 bereaved older individuals reported s i g n i f i c a n t l y more depression than a non-bereaved comparison group. Reich et a l . (1989) investigated the e f f e c t s of l a t e r l i f e spouse bereavement on psychological d i s t r e s s and psychological well-being between 1 and 10 months following the l o s s . In a sample of 58 bereaved and 59 controls matched for age, gender, and socio-economic status, the bereaved maintained s i g n i f i c a n t l y higher l e v e l s of depression and feelings of helplessness\/ hopelessness between 5 and 10 months following the l o s s . S i m i l a r l y , the bereaved reported lower lev e l s of psychological 12 well-being that were r e f l e c t e d i n lower p o s i t i v e a f f e c t . However, there were no differences between the bereaved and control groups on level s of anxiety between the same period of 5 and 10 months following the loss. Research also indicates greater lev e l s of stress i n bereaved i n d i v i d u a l s 12 to 24 months following the death (Caserta & Lund, 1992). Reich et a l . (1989) reported that while bereaved i n d i v i d u a l s evidenced higher levels of psychological d i s t r e s s than non-bereaved controls, the dist r e s s l e v e l s of the bereaved around one year following the loss s i g n i f i c a n t l y lessened and approached the level s of the non-bereaved i n d i v i d u a l s . Others have found coping d i f f i c u l t i e s i n older adults 2 years a f t e r bereavement. For example, Caserta and Lund (1992) found moderate to high stress l e v e l s i n bereaved older adults to continue 1 to 2 years beyond the loss of a spouse. In a portion of the same sample, Lund et a l . (1985) also found that nearly o n e - f i f t h of the 13 8 bereaved adults experienced s i g n i f i c a n t l e v e l s of depression and perceived i n a b i l i t y to cope two years a f t e r t h e i r spouses' deaths. The negative impact of loss i s f e l t both i n the immediate months following loss as well as throughout and beyond the f i r s t year of bereavement. The e f f e c t s of gender on levels of depression and d i s t r e s s i n bereavement are mixed. Studies report that women i n general have higher l e v e l s of depression than men (e.g., Gallagher et a l . , 1983). Sim i l a r l y , i n a sample of spouse bereaved older adults, Jacobs, Hansen, Berkman, Kasl, and Ostfeld (1989) found that spouse bereaved women had higher levels of depression than the bereaved men. However, other bereavement studies 13 incorporating control samples and longitudinal measurements indicate no s i g n i f i c a n t differences between bereaved women and men i n terms of levels of depression over the f i r s t 2 years following bereavement. For example, Gallagher et a l . (1983) found no s i g n i f i c a n t differences between bereaved men and women on l e v e l s of depression 2 months following the loss. Others have found no gender differences among the bereaved through 2 years a f t e r the death of a spouse (Lund et a l . , 1986; Zisook et a l . , 1994) . Less frequent attention has been focused on short-term physical health consequences following the loss of a spouse. Perceived health i s the prototypic operationalization of physical health. Caserta et al . ' s (1990) research on l a t e r l i f e bereavement indicates that perceived health i s strongly associated with physical health and the reporting of symptoms and medical conditions. A study of 212 older i n d i v i d u a l s bereft of t h e i r spouses for two months, as compared to a non-bereaved comparison group of 162 individuals, reported s i g n i f i c a n t l y more i l l n e s s e s , increases i n medication usage, and lower perceived health among the bereaved. Thompson et a l . (1984) also found an increase i n new or worsened i l l n e s s e s i n spouse bereaved women and men two months after the loss. Women, i n general, reported worsened health (Thompson et a l . , 1984). They also found that while the bereaved individuals, as compared to non-bereaved, indicated poorer perceived physical health and poorer physical health r e l a t i v e to others, there were no gender differences nor gender by bereavement interactions. 14 Conceptual: Spouse\/partner loss i n l a t e r l i f e . Spouses i n l a t e r l i f e \"provide each other with more and more of the love, companionship, and stimulation\" that i n d i v i d u a l s need, p a r t i c u l a r l y as networks of friends and family diminish (Nock, 1987, p. 255). The subjective meanings that i n d i v i d u a l s attach to a spouse vary extensively. The loss of a spouse, then, i s the loss of one's partner i n l i f e whose experiences and meaning are unique to each i n d i v i d u a l . Researchers only recently have begun to discuss the importance of understanding the context of a r e l a t i o n s h i p i n order to understand the meaning of the l o s t r e l a t i o n s h i p (Wortman, S i l v e r , & Kessler, 1993). As Wortman et a l . (1993, p. 350) contend, \"the impact of a major loss i s . . . l i k e l y to depend on the meaning of the loss to the i n d i v i d u a l \" . While current research provides l i t t l e understanding of the meaning p a r t i c u l a r relationships hold for i n d i v i d u a l s , inferences about the importance of being married are drawn based on studies i n d i c a t i n g that married individuals have greater l e v e l s of psychological and physical health than non-married adults. For example, married individuals tend to have higher l i f e s a t i s f a c t i o n and better health than non-married i n d i v i d u a l s (Harvey & Bahr, 1974), including those widowed (Hyman, 1983; Larson, 1978), divorced and separated (Larson, 1978). Married men and women also tend to have lower levels of depression as compared to previously and never married in d i v i d u a l s (Pearlin & Johnson, 1981). The causes of such differences favouring married in d i v i d u a l s and the dynamics that bring about such changes are currently open 15 to speculation. The s e l e c t i v i t y hypothesis suggests that i n d i v i d u a l s who are both mentally and ph y s i c a l l y h e a l t h i e r are the most l i k e l y to marry (Stroebe & Stroebe, 1983) ; others suggest that married individuals experience fewer l i f e s t r a i n s . S t i l l others have posited that \"married people are less emotionally responsive than nonmarried people...[and thus] l i f e s t r a i n s have less emotionally damaging ef f e c t s on married than nonmarried people\" (Kessler & Essex, 1982, p. 485). While the nuances of the supportive nature of l a t e r l i f e m arital relationships remain to be further examined, research findings indicate gender differences i n the receipt of s o c i a l support. For example, both younger (age 65-74) and older (age 75+) married women report receiving emotional support less from t h e i r husbands (Depner & Ingersoll-Dayton, 1985) and more from t h e i r children, other family members, or both (Lowenthal & Haven, 1968), while husbands tend to report receiving more emotional support from t h e i r spouses. This, combined with men more l i k e l y reporting wives as confidants than women i d e n t i f y i n g t h e i r spouses as such (Tower & Kasl, 1996), suggest that the loss of a spouse for men may be p a r t i c u l a r l y d i f f i c u l t given the additional loss of a primary provider of emotional support. Given that close f r i e n d network size for older married women and men has been found to be limited to a couple or none (Rubinstein, 1987), reactions to the death of a spouse may be further d i f f e r e n t i a l l y impacted by the presence (or absence) of other close network r e l a t i o n s h i p s . 16 Death of a Close Friend The following three sections provide an accounting of the l i m i t e d empirical attention given to the study of older adults' reactions to the deaths of close friends, and a conceptual discussion of the meaning of a close f r i e n d \u2014 a n d thus the meaning of the loss of a close r e l a t i o n s h i p \u2014 i n l a t e r l i f e . Empirical; Death of a friend i n l a t e r l i f e . A review of the empirical l i t e r a t u r e revealed that while some research has examined the impact of the death of a f r i e n d on young children (O'Brien & Goodenow, 1991) and adolescents (Pohlman, 1984) , only three studies have examined the impact of the death of a close f r i e n d on adults. This i s surprising, given the increased frequency with which older adults experience the loss of close interpersonal relationships (Johnson & T r o l l , 1994). Sklar and Hartley (1990) explored the bereavement reactions of 48 people who had a close f r i e n d die within the preceding f i v e years. T h i r t y - f i v e students aged 18 through 4 5 years completed either in-depth interviews or essays, while the remaining t h i r t e e n students, faculty, and college s t a f f p a r t i c i p a t e d i n a three-session mutual-support group designed s p e c i f i c a l l y to focus on i n d i v i d u a l s who had a close f r i e n d die. Sklar and Hartley (1990) found that many of the reactions to the death of a close f r i e n d were s i m i l a r to the bereavement patterns that follow the death of a spouse or c h i l d ; that i s , changes in psychological health included feelings of loss, anger, g u i l t , and decreased coping a b i l i t i e s . In a longitudinal study, Murphy (1986) examined the stress, coping, and mental health outcomes of 69 i n d i v i d u a l s (mean age 17 40.5 years) following the death of a family member or f r i e n d i n an unexpected natural disaster ( i . e . , Mount St. Helens' volcanic eruption). At both one and three years following the death, stress decreased for both the friend and family groups. However, at one year following the death, family members had s i g n i f i c a n t l y higher l e v e l s of depression and mental d i s t r e s s than did friends of the deceased. Additionally, friends were more l i k e l y to note some p o s i t i v e or \"growth-producing\" e f f e c t s one-year following the loss i n sharp contrast to family members. Interestingly, some of the po s i t i v e e f f e c t s noted by peer bereaved i n d i v i d u a l s included an increased focus on the present, preparations for one's own death, and enjoyment of the present. Only one study has focused on peer bereavement i n l a t e r l i f e . Roberto and Stanis (1994) studied 38 women aged 67 to 92 who had experienced the death of one or more close friends within the preceding 12 years (average time since death was approximately 5 years). Even after t h i s extended period of time, they found that the majority of bereft women ( i . e . , 69%) reported fe e l i n g s of deep loss. One-third indicated an increased f e e l i n g of being alone. Single women ( i . e . , widowed, divorced, never married) aged 75 and older were s i g n i f i c a n t l y more l i k e l y to report a sense of loss than married women and those under age 75. The older women were s i g n i f i c a n t l y more l i k e l y to report a f e e l i n g of deep loss when the close f r i e n d had died within the l a s t f i v e years or when two close friends had died i n the preceding twelve years. However, when the close f r i e n d died more than f i v e years ago there were no age differences i n fe e l i n g s of loss. 18 Roberto and Stanis (1994) also examined the older women's perceptions of changes i n other relationships and personal b e l i e f s and feelings following the death of a close f r i e n d . Approximately half of the women reported f e e l i n g e i t h e r closer to ex i s t i n g friends and\/or a greater appreciation of l i f e , and about a quarter of the individuals made a \"new close f r i e n d . \" Nearly a t h i r d of the women turned to family for friendship and\/or reported an increased appreciation for a family member. Married women (regardless of age) and individuals under age 75 were s i g n i f i c a n t l y more l i k e l y to indicate an increased r e l i a n c e on family members for friendship following the death of t h e i r close f r i e n d . Several effects of a more sel f - e v a l u a t i v e nature were noted, including an increased awareness of aging and awareness of one's mortality for approximately 45% of the peer bereaved women. Similar to Murphy's (1986) findings of p o s i t i v e e f f e c t s deriving from the experience of deaths of a close f r i e n d ( s ) , Roberto and Stanis (1994) found that nearly 45% of the peer bereaved women reported an increased appreciation for l i f e . Conceptual: Death of a friend in l a t e r l i f e . The discussion about peer bereaved individuals i s illuminated by two conceptual a r t i c l e s (Deck & Folta, 1989; Sklar, 1991-1992). Deck and Folta (1989) a r t i c u l a t e some of the macro-social issues involved with peer bereavement, highlighting the role society plays i n struc t u r i n g i n d i v i d u a l and group reactions to bereavement and the prescribed g r i e f patterns. \"Friend-grievers,\" Deck and Folta (1989) note, tend to be ignored or to have t h e i r experiences negated by family, doctors, and the legal system when addressing issues of the deceased person's wishes, funeral arrangements, 19 disposal of the body, property rights, and access to s o c i a l support. Close friends are not recognized as grievers, not i d e n t i f i e d by a term, norms, expectations, or r i g h t s ; a l l of t h i s may contribute to increased levels of d i s t r e s s and g r i e f . Similar to Kastenbaum and Aisenberg's (1972) discussion of older adults' increased recognition of t h e i r own mortality, Deck and Fo l t a (1989) suggest that the death of a close f r i e n d , someone of s i m i l a r age, gender, and socio-economic standing, confronts an i n d i v i d u a l with one's own death and questions of meaning about the future. There may be tension between fee l i n g s regarding the close friend's death i n that there may be both fear that \" i t could have been me\" and r e l i e f that \" i t wasn't me\" (Deck & Folta, 1989) . Si m i l a r l y , Sklar (1991-92; Sklar & Hartley, 1990) underscores the need to recognize that friends also are bereaved when families experience death, and bereft friends may experience profound loss and have intense g r i e f reactions s i m i l a r to family members. \"Grief,\" Sklar comments, \" i s an emotional r o l e whose ri g h t s , p r i v i l e g e s , r e s t r i c t i o n , obligation, and entry requirements tend to be confined to family members\" (1991-92, p. 110). Close friends are not recognized as having l o s t something and t h e i r a f f e c t i v e bond to the deceased tends to preclude any le g a l r i g h t to tangible property. Sklar (1991-92) suggests that the l e g a l i n s t i t u t i o n makes peer bereaved i n d i v i d u a l s a hidden population, further f a i l i n g to recognize close friends' r i g h t s to grieve. Conceptual: Friendship i n l a t e r l i f e . The examination of the meaning or importance of friendships to in d i v i d u a l s i s a 20 neglected area by researchers (Matthews, 1986). While \"close f r i e n d \" denotes a p a r t i c u l a r type of r e l a t i o n s h i p , the meaning derived from the relationship varies across i n d i v i d u a l s , just as the meaning of a spouse varies. For some, a close f r i e n d i s a confidante and sharer of emotions, for others a close f r i e n d i s a companion and sharer of a c t i v i t i e s . It appears that the meaning and the r o l e of a f r i e n d are often inferred from studies of numbers of friends i n an individual's network or frequency of contact with friends (de Vries, 199 6). However, just as a high frequency of receipt of s o c i a l support does not necessarily imply p o s i t i v e or \"meaningful\" support (e.g., Wortman & Lehman, 1985), much contact with and\/or many friends may not warrant the assumption that they are highly supportive (Crohan & Antonucci, 1989) . Although investigations of the meaning of friendship are lim i t e d , research does reveal that close friendships, p a r t i c u l a r l y i n l a t e r l i f e , have a p o s i t i v e e f f e c t on well-being, l i f e s a t i s f a c t i o n and morale (Crohan & Antonucci, 1989; Mancini, 1980). With increased losses and diminishing networks i n l a t e r l i f e , friends take on a greater role and importance than family (Arling, 1976; Wood & Robertson, 1978), often providing emotional support and assistance i n adjusting to new roles (Bankoff, 1983; Lopata, 1977). Furthermore, close friends in l a t e r l i f e tend to have shared decades of experiences together, providing a sense of continuity between the past, present, and future (Matthews, 1986), as well as opportunities for self-evaluation and s e l f -assessment (Roberto & Stanis, 1994). 21 Unlike the ascribed status of f a m i l i a l r e l a t i o n s h i p s , friendships are more voluntary i n nature (e.g., A l l a n & Adams, 1989; Rook, 1989). While t h i s presents a richness and d i v e r s i t y i n the composition and function of friendships across the l i f e course, research findings reveal a homosociality (Lipman-Bluman, 1977) i n l a t e r l i f e friendships; that i s , friends usually choose others of si m i l a r socio-demographic dimensions\u2014age, gender, race, marital status, s o c i a l class or socio-economic status, and r e l i g i o n (e.g., Adams, 1989). P a r t i c u l a r l y i n l a t e r l i f e , cross-gender relationships are rare (Chown, 1981). Both the demographics of la t e r l i f e individuals (more women than men with increasing age) and marital status (more women widowed\/single than men) impact the opportunity for cross-gender r e l a t i o n s h i p s (Allan & Adams, 1989). Widowed individuals, note Bankoff (1983) and Lopata (1977), have much contact with other widowed friends. S i m i l a r l y , Richardson (1984) reported that older women indicate that same-gender friendships are more important than cross-gender r e l a t i o n s h i p s . Multiple Bereaved The l i m i t e d focus of research on reactions to multiple interpersonal losses i n l a t e r l i f e i s r e f l e c t e d i n the following section; t h i s includes a review of the empirical work i n the area as well as a conceptual accounting of the meaning of multiple interpersonal losses i n l a t e r l i f e and t h e i r impact on older adults. Empirical. No studies to date have e x p l i c i t l y examined i n d i v i d u a l s ' reactions to multiple interpersonal losses i n l a t e r l i f e . Past studies have focused on multiple losses i n the 22 context of catastrophes and v i o l e n t deaths. For example, there have been studies or discussions of i n d i v i d u a l s ' reactions to the death of multiple numbers of family and friends i n the context of the Mt. St. Helens' volcanic eruption (Murphy, 1986), the Hiroshima atomic bomb blast (e.g., L i f t o n , 1967) and the Holocaust (e.g., Dimsdale, 1980). More recent studies have examined multiple losses among gay men i n young- and middle-adulthood. For example, B i l l e r and Rice (1990) note that many gay men can witness the \"destruction of entire friendship networks\" due to death from AIDS (p. 283). In a q u a l i t a t i v e study of gay individuals under age 6 0 who experienced the deaths of multiple close friends from AIDS, Carmack (1992) found that i n d i v i d u a l s ' bereavement reactions and grieving a f t e r successive losses grew more intense. D i f f i c u l t y i n coping with loss for some gay men i s compounded by dealing with the multiple loss of i n d i v i d u a l s with AIDS and by not having time to recover from each death. This may represent not only the loss of a r e l a t i o n s h i p of community for some gay men, but rather the loss of a total community as entire networks of friends die. Reactions to multiple interpersonal losses that tend to accompany l a t e r l i f e ( i . e . , death of spouse and friends) have yet to be investigated. Conceptual. Increased numbers of interpersonal losses accompany l a t e r l i f e (Allan, 1989; Johnson & T r o l l , 1994) . Older adults' communities of friends and family may greatly diminish i n s i z e with the deaths of friends and spouses, p a r t i c u l a r l y for men (Allan & Adams, 1989). Kastenbaum (1969) suggests that i t i s l i k e l y that an older individual who experiences multiple bereavements w i l l show some cumulative e f f e c t and be 23 \" p a r t i c u l a r l y vulnerable to the psychological e f f e c t s of lo s s \" (p. 47). This bereavement overload (Kastenbaum, 1969) a r i s i n g from a succession of losses i n lat e r l i f e i s l i k e l y associated with decrements i n the psychological, physical, and s o c i a l domains of functioning. Moss and Moss (1989) suggest that the experience of family deaths and other close interpersonal losses (M.A. Moss, personal communication, September, 1995) over a l i f e t i m e may create a \"personal pool of g r i e f \" that p e r s i s t s over .time and i n t e n s i f i e s with added f a m i l i a l losses. While t h i s i s d i f f e r e n t than bereavement overload occurring with multiple losses i n a short period of time, the pool of g r i e f concept may be extended to include the losses of networks or large portions of one 1 s rel a t i o n s h i p s of community. The intense d i s t r e s s and severe g r i e f that w i l l l i k e l y accompany the experience of being multiple bereaved i n a short period of time and losing one's e n t i r e community of relationships i s similar to that associated with lo s i n g one's spouse or attachment rela t i o n s h i p (Weiss, 1993) i n that the rec i p i e n t s of one's confidences and providers of support and security fostering feelings are gone. Summary Recently widowed women and men d i f f e r e d s i g n i f i c a n t l y from the non-bereaved on a host of dimensions including perceived health, depression, and positive a f f e c t . While s i m i l a r i t i e s e x i s t between women and men i n terms of the i n t e n s i t y of reactions to loss, some studies indicate s i g n i f i c a n t gender differences on level s of perceived health and depression. The reactions of peer bereaved individuals are not well a r t i c u l a t e d 24 i n the empirical l i t e r a t u r e due to the limited number of studies. Research suggests that women and men whose close friends die experience mixed reactions of depression and sadness. Interestingly, some indicators of p o s i t i v e a f f e c t are evidenced. The reactions of multiple bereaved individuals i n l a t e r l i f e have not been examined. The bereavement overload occasioned by the loss of multiple numbers of family and friends suggests that s i g n i f i c a n t decrements to psychological health w i l l r e s u l t . A l l three losses\u2014spouse, close friend, both spouse and close f r i e n d \u2014 s u g g e s t s i g n i f i c a n t impacts on i n d i v i d u a l s ' e f f e c t i v e functioning i n the psychological, physical, and s o c i a l domains. Reactions to s p e c i f i c losses may be d i f f e r e n t i a l l y impacted by gender, close r e l a t i o n s h i p network size ( i . e . , number of close friends, number of close family members, or both), and age. Many of these dimensions, however, have not been examined nor c o n t r o l l e d for i n l a t e r l i f e bereavement research. The degree to which bereaved women's and men's reactions to various types of interpersonal loss are d i f f e r e n t or s i m i l a r i s yet to be examined. Theoretical Framework Weiss' (1993) perspective on loss provides a framework to examine both the death of a spouse and the death of a close f r i e n d . Weiss' (1993) perspective a r t i c u l a t e s d i f f e r e n t types of loss and delineates multiple dimensions on which i n d i v i d u a l s are affected. Drawing on attachment theory, Weiss (1993) delineates two categories of r e l a t i o n s h i p s \u2014 r e l a t i o n s h i p s of attachment and r e l a t i o n s h i p s of community. The death of a spouse or partner i s considered a loss of an attachment re l a t i o n s h i p . The death of a 25 close f r i e n d i s considered the loss of a r e l a t i o n s h i p of community. Weiss' (1993) perspective on loss holds that the death of an in d i v i d u a l embedded in a relationship of attachment (e.g., spouse) w i l l evoke a g r i e f response and intense d i s t r e s s . By contrast, the death of an individual embedded i n a r e l a t i o n s h i p of community (e.g., close friend) w i l l provoke less intense d i s t r e s s and w i l l tend not to evoke intense g r i e f reactions. Weiss' reasoning follows from his attachment perspective in which the \"[o]ne common element of relationships of attachment i s t h e i r linkage to feelings of security\" (1993, p. 272). Weiss (1993) believes that relationships of attachment are not replaceable, and the loss of an attachment relationship w i l l present strong f e e l i n g s of separation d i s t r e s s , pining, and despair. Whereas close friendships and other relationships of community may be highly valued and t h e i r loss evokes some d i s t r e s s , \" g r i e f , \" Weiss notes, \"does not o r d i n a r i l y follow the loss of a f r i e n d \" (1993, p. 273) ; friendships and other losses of community are replaceable. I t i s only when entire relationships of community are l o s t , notes Weiss (1993), that the losses are comparable to the loss of a relationship of attachment; thus, intense bereavement reactions may be evoked. Weiss (1993) provides a multidimensional model to assess i n d i v i d u a l ' s reactions to bereavement. He believes that i n d i v i d u a l s ' recovery and aspects of t h e i r personal and s o c i a l functioning can be assessed on several cognitive, a f f e c t i v e , and behavioral dimensions (Weiss, 1993). A return of the following f i v e a b i l i t i e s , or p a r t i a l return, i s necessary for an i n d i v i d u a l 26 to \"function as an adult and as a member of society\" (Weiss, * 1993, pp. 277-278): 1. Ability to give energy to everyday l i f e . E f f e c t i v e functioning requires investment i n the present, with adequate energy to meet current challenges... 2 . Psychological comfort, as demonstrated by freedom from pain and distress. E f f e c t i v e functioning requires freedom from disturbing thoughts and feelings... 3. Ability to experience gratification - to feel pleasure when desirable, hoped-for, or enriching events occur. E f f e c t i v e functioning requires the a b i l i t y to experience pleasure as well as to anticipate pleasure should hoped-for events occur... 4. Hopefulness regarding the future; being able to plan and care about plans. E f f e c t i v e functioning requires being able to give meaning to a c t i v i t y , and... a sense of a future that may bring with i t something desirable... 5. Ability to function with reasonable adequacy in social roles as spouse, parent, and member of the community. E f f e c t i v e functioning requires meeting s o c i a l expectations well enough to maintain emotionally s i g n i f i c a n t r e lationships... Hypotheses This study compared the short-term reactions of peer bereaved, spouse bereaved, and multiple bereaved women and men with a s i m i l a r aged non-bereaved comparison group. A d d i t i o n a l l y , the i n t e r a c t i o n between bereavement status and gender was investigated. Such a focus i s unique i n that most studies of l a t e r l i f e bereavement have examined only one type of loss. The two primary questions addressed were: (a) What are the differences i n e f f e c t i v e personal and s o c i a l functioning i n peer bereaved, spouse bereaved, multiple bereaved, and non-bereaved older adults? (b) Are there gender differences i n short-term reactions to d i f f e r e n t types of interpersonal losses? 27 The primary hypotheses regarding i n d i v i d u a l s ' reactions to the loss of d i f f e r e n t types of close interpersonal r e l a t i o n s h i p s were guided by Weiss' (1993) perspective on l o s s \u2014 t h a t i s , the i d e n t i f i c a t i o n of s p e c i f i c types of loss and a multidimensional model to assess individuals' e f f e c t i v e personal and s o c i a l functioning. In addition to bereavement group comparisons on multidimensional measures of psychosocial functioning, hypotheses about gender differences were presented based on empirical and conceptual l i t e r a t u r e . As well, research expectations were presented based on limited conceptual l i t e r a t u r e . Research questions that addressed the interplay between gender and multiple types of loss also were offered where there was no substantial empirical nor conceptual l i t e r a t u r e to p o s i t a r e l a t i o n s h i p between variables a p r i o r i . Weiss' (1993) conceptual perspective on loss holds that bereaved i n d i v i d u a l s ' recovery or return to ordinary functioning following loss may be assessed on several dimensions of e f f e c t i v e functioning. Individuals' experiencing the loss of a r e l a t i o n s h i p of attachment ( i . e . , spouse) are expected to experience lower levels of functioning due to intense d i s t r e s s and severe g r i e f reactions as compared to i n d i v i d u a l s who lose a r e l a t i o n s h i p of community ( i . e . , close friend) and experience less impact on personal and s o c i a l functioning. Gerontology and bereavement research has consistently revealed that widowed ind i v i d u a l s are negatively affected in terms of health, depression and psychological d i s t r e s s , and s o c i a l functioning (e.g., Breckenridge et a l . , 1986; Lund et a l . , 1993; Reich et a l . , 1989). The loss of a spouse presents an i n d i v i d u a l with the 28 loss of a key component of one's id e n t i t y (Lopata, 1973). Short-term adjustment or adaptation to widowhood e n t a i l s numerous l i f e changes (physical, psychological, and s o c i a l ) , many that negatively impact the bereaved i n d i v i d u a l . The paucity of research on older a d u l t s 1 reactions to the death of a close f r i e n d l i m i t s the body of empirical l i t e r a t u r e from which to draw. The loss of a close f r i e n d represents a void i n an i n d i v i d u a l ' s l i f e \u2014 a sharer of l i f e experiences, a c t i v i t i e s , and confidences i s gone. Friends can act as yardsticks by which to measure how one i s doing\u2014achieving goals, maintaining good health and well-being\u2014or act as mirrors i n r e f l e c t i n g the s i m i l a r i t i e s that.one sees i n one's s e l f \u2014 s h a r e d values, interests, and experiences. Research indicates that peer bereaved individuals, l i k e spouse bereaved i n d i v i d u a l s , report poorer psychological and physical health (Roberto & Stanis, 1994; Sklar & Hartley, 1990). A comparison of a l i m i t e d number of young and m i d - l i f e spouse and peer bereaved in d i v i d u a l s revealed that widowed women and men reported poorer mental and physical health than peer bereaved individuals (Murphy, 1986). Weiss (1993) also posits that individuals who lose an e n t i r e community of relationships ( i . e . , loss of a l l friends) w i l l tend to experience severe distress and reduced l e v e l s of e f f e c t i v e functioning comparable to those of i n d i v i d u a l s who lose a spouse. The loss of both a spouse and a close f r i e n d may approximate t h i s type of t o t a l community loss, as the main r e c i p i e n t s of confidences and providers of feelings of security are gone. The \"bereavement overload\" occasioned by multiple interpersonal losses i n a short period of time, notes Kastenbaum 29 (1969), may negatively impact individuals' psychological health. As i n d i v i d u a l s experience successive deaths of friends and family, a \"pool of g r i e f \" develops which causes each addi t i o n a l death to be a reminder of past ones and influence one's a b i l i t y to adjust or adapt to the loss (Moss & Moss, 1989). Given such a s i g n i f i c a n t loss of one's s o c i a l network, multiple bereaved i n d i v i d u a l s may experience as much, i f not more, negative personal and s o c i a l functioning than spouse bereaved i n d i v i d u a l s . The following hypotheses (H) and research expectations (E) about bereavement reactions to d i f f e r e n t types of interpersonal loss were proposed: HI: Spouse bereaved individuals w i l l report the lowest l e v e l s of perceived health, which w i l l be s i g n i f i c a n t l y lower than the peer bereaved, which w i l l be s i g n i f i c a n t l y lower than the non-bereaved comparison group. E l : Perceived health scores for the multiple bereaved group w i l l be equal to, or even less than, the spouse bereaved group; multiple bereaved i n d i v i d u a l s ' scores also w i l l be s i g n i f i c a n t l y less than those of the peer bereaved and non-bereaved i n d i v i d u a l s . H2: Spouse bereaved individuals w i l l report the highest l e v e l s of negative a f f e c t ( i . e . , low negative a f f e c t scores), which w i l l be s i g n i f i c a n t l y greater than the peer bereaved, which w i l l be s i g n i f i c a n t l y greater than the non-bereaved comparison group. E2: Negative a f f e c t levels for the multiple bereaved group w i l l be equal to, or even greater than, the spouse bereaved group; multiple bereaved i n d i v i d u a l s ' l e v e l s of negative a f f e c t also w i l l be s i g n i f i c a n t l y greater than those of the peer bereaved and non-bereaved in d i v i d u a l s . H3: Spouse bereaved individuals w i l l report the lowest l e v e l s of p o s i t i v e a f f e c t , which w i l l be s i g n i f i c a n t l y lower than the peer bereaved, which w i l l be s i g n i f i c a n t l y lower than the non-bereaved comparison group. E3: Po s i t i v e a f f e c t scores for the multiple bereaved group w i l l be equal to, or even less than, the spouse bereaved group; multiple bereaved i n d i v i d u a l s ' scores also w i l l be s i g n i f i c a n t l y less than those of the peer bereaved and non-bereaved i n d i v i d u a l s . 30 H4: Spouse bereaved individuals w i l l report the lowest l e v e l s of preparation for the future, which w i l l be s i g n i f i c a n t l y lower than the peer bereaved, which w i l l be s i g n i f i c a n t l y lower than the non-bereaved comparison group. E4: Preparation for the future scores for the multiple bereaved group w i l l be equal to, or even less than, the spouse bereaved group; multiple bereaved i n d i v i d u a l s ' scores also w i l l be s i g n i f i c a n t l y less than those of the peer bereaved and non-bereaved i n d i v i d u a l s . H5: Spouse bereaved individuals w i l l report the lowest l e v e l s of s o c i a l functioning, which w i l l be s i g n i f i c a n t l y lower than the peer bereaved, which w i l l be s i g n i f i c a n t l y lower than the non-bereaved comparison group. E5: S o c i a l functioning scores for the multiple bereaved group w i l l be equal to, or even less than, the spouse bereaved group; multiple bereaved i n d i v i d u a l s ' scores also w i l l be s i g n i f i c a n t l y less than those of the peer bereaved and non-bereaved i n d i v i d u a l s . H5A: Spouse bereaved individuals w i l l report s i g n i f i c a n t l y less s o c i a l involvement than the peer bereaved, which w i l l be s i g n i f i c a n t l y lower than the non-bereaved comparison group. H5B: Spouse bereaved individuals w i l l report s i g n i f i c a n t l y lower levels of emotional support than the peer bereaved, which w i l l be s i g n i f i c a n t l y lower than the non-bereaved comparison group. H5C: Spouse bereaved individuals w i l l report s i g n i f i c a n t l y lower l e v e l s of s a t i s f a c t i o n with family r e l a t i o n s h i p s than the peer bereaved, which w i l l be s i g n i f i c a n t l y lower than the non-bereaved comparison group. H5D: Spouse bereaved individuals w i l l report s i g n i f i c a n t l y lower lev e l s of s a t i s f a c t i o n with f r i e n d r e l a t i o n s h i p s than the peer bereaved, which w i l l be s i g n i f i c a n t l y lower than the non-bereaved comparison group. Gender i s s o c i a l l y constructed, and i t s examination allows for placing i n d i v i d u a l s ' personal and s o c i a l functioning i n a c u l t u r a l context. In general (regardless of bereavement status) women report greater levels of depression than men (Gallagher et a l . , 1983). The expressivity factor may account for some of the greater reporting of affe c t by women (both bereaved and non-bereaved) , in that women tend to report greater degrees of 31 a f f e c t . This may also be due to women's greater l e v e l of intimacy and connectedness with relationships (e.g., G i l l i g a n , 1982) . Consistent with other studies that have found women to indicate poorer health than men (e.g., Thompson et a l . , 1984), women are predicted to have lower perceived health than men. Men are expected to report lower levels of s o c i a l functioning than women (Stroebe & Stroebe, 1983): decreased s o c i a l involvement, less emotional investment i n relationships, and less s a t i s f a c t i o n both with friendships and relationships with family. The following hypotheses (H) about women's and men's e f f e c t i v e personal and s o c i a l functioning were proposed: H6: Men w i l l have s i g n i f i c a n t l y higher l e v e l s of perceived health than women. H7: Men w i l l have s i g n i f i c a n t l y lower l e v e l s of negative a f f e c t than women. H8: Women w i l l have s i g n i f i c a n t l y higher l e v e l s of p o s i t i v e a f f e c t than men. H9: Women w i l l have s i g n i f i c a n t l y higher l e v e l s of preparation for the future than men. H10: Women w i l l have s i g n i f i c a n t l y higher l e v e l s of s o c i a l functioning than men. H10A: Women w i l l have s i g n i f i c a n t l y higher l e v e l s of s o c i a l involvement than men. H10B: Women w i l l have s i g n i f i c a n t l y higher l e v e l s of emotional investment than men. H10C: Women w i l l have s i g n i f i c a n t l y higher l e v e l s of s a t i s f a c t i o n with family r e l a t i o n s h i p s than men. H10D: Women w i l l have s i g n i f i c a n t l y higher l e v e l s of s a t i s f a c t i o n with f r i e n d r e l a t i o n s h i p s than men. Individuals' reactions to close interpersonal loss are expected to vary by gender. Drawing from the widowhood l i t e r a t u r e (e.g., Jacobs et a l , 1989; Thompson et a l , 1984), bereaved women tend to report greater levels of psychological 32 d i s t r e s s and poorer physical health than widowed men. The more normative nature (and greater expectation) of becoming a widow as opposed to a widower also suggests that spouse bereaved women w i l l report greater le v e l s of preparation for the future than widowed men. While i t i s expected that women's and men's reactions to s p e c i f i c interpersonal losses w i l l be consistent with widowhood empirical findings and conceptual l i t e r a t u r e \u2014 women reporting poorer health and greater d i s t r e s s , while men reporting poorer s o c i a l f u n c t i o n i n g \u2014 t h e examination of the in t e r a c t i o n between gender and d i f f e r e n t bereavement statuses are more exploratory. Given the limited empirical attention to l a t e r l i f e peer bereavement and the absence of l i t e r a t u r e on multiple bereaved older individuals, the following research questions were examined: Ql: Does gender interact with bereavement status to influence levels of perceived health? Q2: Does gender interact with bereavement status to influence levels of negative affect? Q3: Does gender interact with bereavement status to influence levels of posit i v e affect? Q4: Does gender interact with bereavement status to influence levels of preparation for the future? Q5A: Does gender interact with bereavement status to influence levels of s o c i a l involvement? Q5B: Does gender interact with bereavement status to influence levels of emotional investment? Q5C: Does gender interact with bereavement status to influence levels of family s a t i s f a c t i o n ? Q5D: Does gender interact with bereavement status to influence levels of friend s a t i s f a c t i o n ? Further exploratory analyses were conducted within the peer bereaved group of individuals. While t h i s was outside the 33 context of the proposed comparison study, the lack of attention peer bereavement has received ( p a r t i c u l a r l y i n contrast to widowhood) combined with the large sample si z e of peer bereaved in d i v i d u a l s , c a l l s attention to t h i s understudied large group of i n d i v i d u a l s . Given the growing body of l a t e r l i f e f riendship l i t e r a t u r e that sheds l i g h t on the value of close interpersonal r e l a t i o n s , research expectations were drawn about ind i v i d u a l s * reactions to the death of a close friend. Exploratory analyses within the peer bereaved group were conducted using gender, as well as close friendship number, close family r e l a t i o n s h i p number, and age, as independent varia b l e s . Thus, i t was anticipated that the potential interplay between gender, close friendship network size, family network siz e , and age would be elucidated and shed l i g h t on the understanding of reactions to l a t e r l i f e friendship loss. A sampling of the research questions follows. The number of friends an individual maintains i s related to a host of c h a r a c t e r i s t i c s including greater well-being, l i f e s a t i s f a c t i o n , and health (Crohan & Antonucci, 1989) . Would los i n g one of many friends or many of a l l friends d i f f e r e n t i a l l y impact an individual? It i s expected that losing a l l of one's friendship network ( i . e . , losing one's only close friend) w i l l negatively impact individuals more than those who have one or several remaining friends. Would the e f f e c t s be the same for women and men? On the one hand, women tend to report greater involvement and intimacy in relationships ( G i l l i g a n , 1982), p a r t i c u l a r l y outside of marriage; thus, the reactions to the loss of a close f r i e n d may be greater for women than men. On the 34 other hand, women's embeddedness i n a more interconnected s o c i a l network may o f f e r the opportunity for greater support than men's close r e l a t i o n s h i p s . Regardless of gender, though, the number of close family relationships i s expected to moderate the e f f e c t s of friendship loss. The normativeness of loss i s associated with age; increased losses are expected with increasing age. However, the normative nature of loss for the old-old (age 75 and older) as compared to the young-old (age 65 - 74) may not warrant the assumption that the death of a close f r i e n d evokes greater negative psycho-social reactions i n the old-old. The combined losses associated with having more experiences may cause greater d i f f i c u l t i e s . Thus, i s age associated with levels of negative reactions to the loss of a close friend? And w i l l the size of one's close r e l a t i o n s h i p network, either friend, family, or both, inter a c t with age to account f o r differences i n peer bereavement reactions? W i l l t h i s be further moderated by gender? 35 Chapter II Method Sample The sample on which the following analyses are based was drawn from a large representative Canadian study e n t i t l e d Survey on Ageing and Independence (SAI; S t a t i s t i c s Canada, 1991). Twenty thousand Canadians aged 4 5 and older responded to interviews that were about 3 0 minutes i n length, with approximately 90% of them being conducted over the telephone and the r e s t conducted in respondents' homes. The content of the questionnaire was diverse, addressing aspects of health, a c t i v i t y , independence, and retirement. A response rate of 81% was obtained for the t o t a l sample of individuals aged 4 5 and over: those aged 65-79 had a response rate between 81%-84%, and i n d i v i d u a l s aged 80 or older had a response rate of 74%. Individuals excluded from the SAI included people l i v i n g in the Yukon and Northwest T e r r i t o r i e s , on F i r s t Nation's reservations, as well as Armed Force's members and i n s t i t u t i o n a l i z e d adults ( i . e . , an estimated 8% of those aged 65 to 79 and 24% of in d i v i d u a l s aged 80 and older, S t a t i s t i c s Canada, 1991). Given the focus of t h i s study, only women and men aged 65 and older (N = 10,059) were included due to the normative nature of interpersonal losses i n l a t e r l i f e : age i s associated both with mortality rates and the experience of being bereft of a spouse or f r i e n d . Data then were compiled on i n d i v i d u a l s who had experienced the death of a spouse, the death of a close f r i e n d , or both, i n the preceding twelve months. As well, data were compiled on non-bereaved individuals ( i . e . , no death of a spouse 36 or close f r i e n d i n the preceding twelve months); i n d i v i d u a l s who responded that they \"did not know\" i f they had experienced the p a r t i c u l a r type of loss were excluded from the study due to the suspected u n r e l i a b i l i t y of the data. The non-bereaved comparison group also excluded individuals whose marital status had changed i n four or fewer years (e.g., no individuals widowed f o r fewer than f i v e years); empirical evidence suggests that adjustment to (or p a r t i a l return to ordinary levels of functioning) following such losses as widowhood and divorce may be accomplished within such a time frame (e.g., Lund et a l . , 1993). The t o t a l study sample consisted of 8,878 adults aged 65 and older who experienced the death of a close f r i e n d , the death of a spouse, or both the death of a close f r i e n d and a spouse, within the l a s t year as well as non-bereaved individuals who had no change i n marital status for f i v e or more years. Approximately 43% of the sample were men and 57% women. The non-bereaved were the si n g l e largest group, comprising approximately 60% of the t o t a l study sample, n = 5358 (2275 men, 3083 women). Each of the bereaved groups were mutually exclusive: 3 6.0% of the pa r t i c i p a n t s recently experienced the death of a close f r i e n d and no death of a spouse i n the l a s t four years ( i . e . , peer bereaved), n = 3198 (1468 men, 1730 women); 2.3% of the par t i c i p a n t s recently experienced the death of a spouse ( i . e . , spouse bereaved), n = 204 (64 men, 140 women); and 1.3% of the pa r t i c i p a n t s recently experienced the death of both a spouse and a close f r i e n d ( i . e . , multiple bereaved), n = 118 (35 men, 83 women). 37 The average age of the t o t a l study sample was 74.77 years, with no s i g n i f i c a n t age differences between bereavement statuses, F(3,8877) = 1.34, p_ = .26. Most participants had some secondary schooling, with values ranging from eight or less years of schooling through a university degree. Bereavement status differences on leve l s of schooling, F(3,8877) = 3.14, p. < .05, were not s i g n i f i c a n t between groups on follow-up analyses using Scheffe, p_ < .05. The number of other losses experienced i n the preceding year (e.g., loss of residence, i l l n e s s \/ i n j u r y of respondent) d i f f e r e d between bereavement statuses, F(3,8777) = 95.96, p_ < .05, ranging from 0 to 4, with an average of .46. Follow-up analyses with Scheffe, p_ < .05, revealed that the non-bereaved reported the fewest number of other losses compared to the peer, spouse, and multiple bereaved. As well, the peer bereaved reported fewer losses than the multiple bereaved. While respondents tended not to have any health l i m i t a t i o n s that hampered t h e i r involvement i n physical a c t i v i t i e s , differences between the bereavement statuses existed, F(3,8877) = 5.67, p_ < .001. Scheffe, p_ < .05, follow-up analyses revealed that the peer bereaved tended to experience health l i m i t a t i o n s more than the non-bereaved. Participants reported on average three to four close family relationships and three to four close friends. Peer bereaved and multiple bereaved individuals i d e n t i f i e d maintaining more close family relationships than the spouse bereaved and non-bereaved. The peer bereaved also reported having a greater number of close friends than the non-bereaved. See Table 1, for socio-demographic c h a r a c t e r i s t i c s of the sample. Table 1 Ch a r a c t e r i s t i c s Of Bereaved and Non-Bereaved Adults 38 Multiple Spouse Peer Non-Bereaved Bereaved Bereaved bereaved n 118 204 3198 5358 Gender Men 35 64 1468 2275 Women 83 140 1730 3083 a Age M 75. 56 a 75. 36 a 74 . 80 a 74 . 72 a SD 6. 70 6. 23 6. 34 6. 45 Education M 2. 29 a 2 . 07 a 2 . 34 a 2 . 25 a SD 1. 54 1. 41 1. 56 1. 54 No. Close Family 0 M 4 . 14 a 3 . 04 b 3 . 79 a 3 . 21b SD 3 . 73 3 . 32 3 . 49 3 . 19 No. Close Friends M 3 . 80 ab 3 . 48ab 4 . 28 a 2 . 95 b SD 5. 08 5. 59 7. 27 5. 89 Other Losses M 85 a 64 ab 59 b 37 SD \u2022 84 \u2022 73 \u2022 70 \u2022 61 . . . f Health Limitation M 3 lab 32ab 39 a 35 b SD \u2022 47 \u2022 47 \u2022 49 \u2022 48 Note. Any row means that do not share subscripts d i f f e r at p_ < .05, with Scheffe procedure. Ranges 65-102; aged 80 and older given mean of 84. Ranges from 1 = (8 years or l e s s ) , 2 = (some high school), through 6 = c d (university degree). Ranges 0-30. Ranges 0-98. e f Ranges 0-4. 0 = presence of health l i m i t a t i o n , 1 = no health l i m i t a t i o n . Measures Dependent Variables The items for the seven dependent measures are included i n Appendix A. Drawing on Weiss' (1993) a r t i c u l a t i o n of the multidimensional c r i t e r i a for e f f e c t i v e personal functioning, a perceived health index, a negative a f f e c t scale, and a p o s i t i v e a f f e c t scale were used to p a r a l l e l Weiss' (1993) f i r s t three dimensions of e f f e c t i v e f u n c t i o n i n g \u2014 a b i l i t y to give energy to everyday l i f e , freedom from disturbing thoughts and fe e l i n g s , a b i l i t y to experience g r a t i f i c a t i o n , respectively. A preparation fo r the future index was guided by Weiss' (1993) fourth dimension of e f f e c t i v e functioning\u2014hopefulness for the f u t u r e \u2014 b u t was not assessed. Items intended to measure preparation for the future f a i l e d to form a unidimensional index and did not approximate Weiss' (1993) related dimension of hopefulness for the future as c l o s e l y as did the other dependent measures and t h e i r respective dimensions of e f f e c t i v e functioning. S p e c i f i c a l l y , the preparation for the future items assessed more long-term f i n a n c i a l planning and a c t i v i t y levels than did Weiss' (1993) concepts of \"being able to plan and care about plans\" (p. 278) and maintain a hopefulness for desirable things to come. Social functioning, approximating Weiss' (1993) f i f t h dimension of e f f e c t i v e functioning, was assessed with indices of s o c i a l involvement, emotional investment, family s a t i s f a c t i o n , and f r i e n d s a t i s f a c t i o n . A p r i n c i p a l components factor analysis with varimax ro t a t i o n was conducted on a l l individual scale\/index items. High factor loadings were evident on a l l scales\/indices (see Appendix B, 40 Table 1). There were no high cross loadings among the variables; the factor analysis revealed a l l scales\/indices to be d i s t i n c t . A d d i t i o n a l l y , low intercorrelations among dependent measures evidenced the need to keep the conceptually d i s t i n c t factors of e f f e c t i v e functioning independent of one another (see Appendix B, Table 2). 1. Perceived health; \"to give energy to everyday l i f e \" . Weiss (1993) describes e f f e c t i v e functioning as the a b i l i t y to give energy to everyday l i f e . A measure of perceived health was created to approximate t h i s dimension. I t i s believed that i n d i v i d u a l s who view themselves as healthy and possessing adequate physical energy w i l l have the a b i l i t y to meet the challenges of d a i l y l i v i n g and be capable of giving energy to everyday l i f e . Caserta et a l . (1990) found that older adults* ratings of t h e i r own health closely represented the l e v e l of the physical symptoms and physical health of the i n d i v i d u a l s . A composite measure of one 4-point and two 3-point Likert-type scales was used. The measure of respondents des c r i p t i o n of t h e i r own health was condensed from a 4-point L i k e r t response (1 = poor, 2 = f a i r , 3 = good, 4 = excellent) to a 3-point L i k e r t scale (1 = poor\/fair, 2 = good, 3 = excellent). This 3-point range p a r a l l e l e d the response scale of the other two perceived health items: (a) respondents rated t h e i r physical a c t i v i t y l e v e l as compared to others of the same age (1 = less active, 2 = as active, 3 = more active) and (b) respondents rated t h e i r health compared to others of the same age (1 = worse, 2 = same, 3 = b e t t e r ) . Scores were summed and averaged, y i e l d i n g a range from 1 to 3. Higher scores r e f l e c t e d greater lev e l s of perceived 41 health. The i n t e r n a l consistency of the perceived health index was moderately high (Cronbach's alpha = .74) for the t o t a l study sample. 2. Negative a f f e c t ; \"freedom from disturbing thoughts and f e e l i n g s \" . Weiss (1993) notes that e f f e c t i v e functioning requires, among other things, freedom from disturbing thoughts and f e e l i n g s . A measure of negative a f f e c t , that i s , an indicator of individuals' psychological discomfort and d i s t r e s s , was used to approximate t h i s dimension. The negative a f f e c t scale was composed of f i v e 3-point Likert-type items. Four items came from the Bradburn Affect Scale-Negative (1969). Individuals rated the r e l a t i v e presence or absence of four s p e c i f i c feelings and emotions during the past few weeks (1 = often, 2 = sometimes, 3 = never): loneliness, sadness, boredom, and restlessness. On the f i f t h item, participants rated the stressfulness of t h e i r l i f e (1 = very s t r e s s f u l , 2 = not very s t r e s s f u l , 3 = not at a l l s t r e s s f u l ) . Scores were summed and averaged, y i e l d i n g a range from 1 to 3; higher scores represented a greater absence of negative a f f e c t ( i . e . , less negative a f f e c t ) . Analyses of the t o t a l study sample revealed a Cronbach's alpha of .69. 3. P o s i t i v e a f f e c t ; \"to experience g r a t i f i c a t i o n \" . Weiss (1993) also i d e n t i f i e s an individual's a b i l i t y to experience g r a t i f i c a t i o n as an indicator of e f f e c t i v e functioning following bereavement. Closely approximating the degree to which in d i v i d u a l s experience g r a t i f i c a t i o n and pleasure was a measure of p o s i t i v e a f f e c t . This was assessed with a composite of four 3-point Likert-type scales taken from the Bradburn A f f e c t Scale-P o s i t i v e (1969). Individuals rated the r e l a t i v e presence or 42 absence of s p e c i f i c feelings and emotions during the preceding few weeks (1 = never, 2 = sometimes, 3 = often): top of the world, excited, pleased, proud. The four items were summed and averaged, y i e l d i n g a range from 1 to 3. Higher scores represented greater levels of posit i v e a f f e c t . Examination of the t o t a l study sample revealed the scale to be i n t e r n a l l y consistent, with a Cronbach's alpha of .68. 4. So c i a l involvement: \"to function with reasonable adequacy i n s o c i a l r o l e s \" . Weiss (1993) i d e n t i f i e s the \" a b i l i t y to function with reasonable adequacy in s o c i a l r o l e s \" as a dimension of e f f e c t i v e functioning. Guided by t h i s dimension, a composite index of s o c i a l involvement was used to assess s o c i a l functioning. Four 3-point Likert-type scales assessed the frequency an in d i v i d u a l goes to v i s i t f r i e n d s \/ r e l a t i v e s , t a l k s with others on the telephone, has family and friends over to t h e i r residence for a v i s i t , and plays cards or other games. Responses ranged from 1 'never 1, 2 'sometimes', and 3 'often'. Scores were summed and averaged, y i e l d i n g a range from 1 to 3, with higher scores indicating a greater l e v e l of s o c i a l investment. The internal consistency of the index for the t o t a l study sample, as measured by Cronbach's alpha, was .55. 5. Emotional investment: \"to function with reasonable adequacy i n s o c i a l r o l e s \" . Weiss (1993) also i d e n t i f i e s the dimension of functioning in s o c i a l roles as including \"meeting s o c i a l expectations well enough to maintain emotionally s i g n i f i c a n t relationships\" (p. 278). Guided by t h i s dimension, a measure of emotional investment was constructed to assess the l e v e l of emotional support received and provided i n close 43 interpersonal relationships. Two items comprised the measure of emotional investment. The f i r s t item assessed whether or not an i n d i v i d u a l received emotional support in the previous twelve months. The second item assessed whether or not an i n d i v i d u a l provided emotional support in the previous twelve months. These two dichotomous items were summed and then averaged, y i e l d i n g values of 0, .5, and 1; higher scores indicated greater l e v e l s of emotional investment. The internal consistency for the measure of emotional investment, as measured by Cronbach's alpha, was .49. 6. S a t i s f a c t i o n with family rela t i o n s h i p s : \"to function with reasonable adequacy i n s o c i a l r o l e s \" . Also guided by Weiss' (1993) dimension of s o c i a l functioning, a measure was used to assess i n d i v i d u a l s ' o v e r a l l s a t i s f a c t i o n with the kind and number of family relationships. Frequency of i n t e r a c t i o n and s a t i s f a c t i o n with relationships are conventional methods of assessing relationships (Adams, 1989; de Vries, 1996). The degree to which one adequately functions in s o c i a l r o l e s may be r e f l e c t e d i n the lev e l s of s a t i s f a c t i o n or d i s s a t i s f a c t i o n with one's close interpersonal relationships. A six-point Likert-type scale, ranging from very d i s s a t i s f i e d to very s a t i s f i e d , assessed p a r t i c i p a n t s ' s a t i s f a c t i o n with the kind and frequency of contact with family relationships. Due to poor psychometric properties ( i . e . , a trimodal d i s t r i b u t i o n ) , the measure was condensed to a 3-point L i k e r t scale (1 = d i s s a t i s f i e d , 2 = s a t i s f i e d , 3 = very s a t i s f i e d ) . 7. S a t i s f a c t i o n with friends: \"to function with reasonable adequacy i n s o c i a l r o l e s \" . Drawing on Weiss' (1993) dimension of 44 s o c i a l functioning, a conventional measure was used to assess s a t i s f a c t i o n with close friend relationships. A six-point Likert-type scale, ranging from very d i s s a t i s f i e d to very s a t i s f i e d , assessed respondents' s a t i s f a c t i o n with the kind and frequency of contact with friends. Due to poor psychometric properties ( i . e . , a trimodal d i s t r i b u t i o n ) , the measure was condensed to a 3-point Like r t scale (1 = d i s s a t i s f i e d , 2 = s a t i s f i e d , 3 = very s a t i s f i e d ) . Independent Variables The independent variables consisted of bereavement status and gender. Bereavement status was a nominal l e v e l v a r i a ble coded into four l e v e l s : 1 = non-bereaved (individuals who had neither experienced the death of a close f r i e n d or spouse i n the preceding twelve months nor a change i n marital status for over four years), 2 = peer bereaved (individuals who experienced the death of a close friend i n the preceding twelve months and no death of a spouse in the l a s t four years), 3 = spouse bereaved (individuals who experienced the death of a spouse i n the preceding twelve months and no death of a close f r i e n d in the l a s t year), and 4 = multiple bereaved (individuals who experienced the death of both a spouse and a close f r i e n d i n the preceding twelve months). A l l four levels of bereavement status were mutually exclusive. Gender was a nominal l e v e l variable and coded 1 = male, 2 = female. Control Variables Six variables were used as controls: t o t a l number of other losses experienced in the preceding twelve months, health 45 l i m i t a t i o n s , age, education, number of close family r e l a t i o n s h i p s , and number of close friends (see Appendix A). Total number of losses in the preceding 12 months included the change or loss of a job, change i n residence, a person moving into or leaving the respondents' house, serious i l l n e s s or injury of the respondent, or the serious i l l n e s s or injury of a family member or f r i e n d (potential range was 0 - 5) . The t o t a l number of losses were used as a control variable i n order to assess the unique impact of s p e c i f i c losses. Health l i m i t a t i o n s (assessed with a forced choice item indicating whether or not there were any l i m i t a t i o n s i n the amount or kind of a c t i v i t y that one could undertake) was controlled for, as older adults' declines i n physical health may be associated with l i m i t a t i o n s i n s o c i a l interactions and \" l i m i t opportunities for se r v i c i n g friendships and developing new t i e s \" (Allan & Adams, 1989, p. 57). Age was treated as a control variable as there i s known v a r i a b i l i t y across older adults. As well, c o n t r o l l i n g for age i s consistent with the study of bereavement reactions in l a t e r l i f e . Education also was controlled for, as i t p o t e n t i a l l y influences opportunities for interaction with others, friendship networks, and a b i l i t y to afford or obtain services. Education represents the highest l e v e l of formal education an i n d i v i d u a l received, ranging from 1 (eight years or less of schooling) to 6 (a un i v e r s i t y degree). While the education l e v e l s represent ordinal l e v e l data, education was treated as an i n t e r v a l l e v e l variable given that, conceptually, there was an ordering of education from lowest to highest. Both the number of close family members and number of close friends an individual i d e n t i f i e d was cont r o l l e d 46 for as i n d i v i d u a l s with larger networks have a greater l i k e l i h o o d of experiencing the death of a friend; a d d i t i o n a l l y , having larger networks i s associated with p o s i t i v e psychological health outcomes (e.g., Johnson, 1983). (Appendix B, Tables 3 and 4 present correlations between the control and dependent variables, as well as the i n t e r c o r r e l a t i o n s among the control variables, r e s p e c t i v e l y ) . Analyses A 4 (Bereavement status: Non-bereaved, Peer bereaved, Spouse bereaved, Multiple bereaved) X 2 (Gender: Male, Female) Analysis of Covariance (ANCOVA), covarying t o t a l number of losses experienced i n the preceding twelve months, health l i m i t a t i o n s , age, education, number of close family relationships, and number of close friends, was used to test the r e l a t i o n s h i p s between the independent and dependent variables. Seven ANCOVAs were conducted on the dependent variables assessing perceived health, negative a f f e c t , p o s i t i v e a f f e c t , s o c i a l involvement, emotional investment, family s a t i s f a c t i o n , and f r i e n d s a t i s f a c t i o n . Regression analyses were conducted for two reasons: f i r s t , to i d e n t i f y the s i g n i f i c a n t predictors of functioning across bereavement statuses; and second, to uncover patterns o f socio-demographic c h a r a c t e r i s t i c s within bereavement statuses. Gender, number of other losses, health l i m i t a t i o n s , age, education, number of close family r e l a t i o n s , and number o f close friends were treated as predictor variables and entered simultaneously into a regression equation predicting each outcome measure within each bereavement status. Thus, for each of the four bereavement statuses, regression equations were conducted to predict perceived health, negative af f e c t , p o s i t i v e a f f e c t , s o c i a l involvement, emotional investment, s a t i s f a c t i o n with family r e l a t i o n s h i p s , and s a t i s f a c t i o n with friendships. 48 Chapter III Results Univariate Distributions The d i s t r i b u t i o n s of the dependent measures were examined for normality, skewness, and kurtosis. Although the index for perceived health was s l i g h t l y negatively skewed and leptokurtic, the d i s t r i b u t i o n appeared to be normal. The measure for negative a f f e c t was leptokurtic and negatively skewed, with i n d i v i d u a l s reporting r e l a t i v e l y low levels of negative a f f e c t . Although the measures of p o s i t i v e a f f e c t and s o c i a l involvement were k u r t o t i c , the d i s t r i b u t i o n s appeared to be normal. The d i s t r i b u t i o n for emotional investment was kurtotic and p o s i t i v e l y skewed, with more ind i v i d u a l s reporting no receipt and no provision of emotional support. Consistent with measures of s a t i s f a c t i o n , both the measure of s a t i s f a c t i o n with family r e l a t i o n s h i p s and the measure of s a t i s f a c t i o n with friends were k u r t o t i c and skewed, with individuals tending to report much s a t i s f a c t i o n with the kind and frequency of contact with close interpersonal r e l a t i o n s h i p s . Hypothesis Testing The assumption of homogeneity of vari a n c e \u2014 a n underlying assumption of analysis of covariance\u2014was not met for the measures of negative a f f e c t , emotional investment, family s a t i s f a c t i o n , f r i e n d s a t i s f a c t i o n , and approached meeting the assumption on po s i t i v e a f f e c t (Bartlett's p = .033). Analysis of covariance i s a robust s t a t i s t i c a l procedure even when c e l l sizes are unequal and d i s t r i b u t i o n s non-normal, given there i s homogeneity of variance (Kerlinger, 1986). Even i n cases where 49 there i s heterogeneity of variance, ANCOVA proves to be robust given comparable c e l l sizes (Kerlinger, 1986). Less i s known about the robustness of ANCOVA when c e l l sizes are unequal and both normality and homogeneity assumptions are v i o l a t e d . Consequently, i n an e f f o r t to evidence the robustness of ANCOVA, the r e s u l t s of t h i s study were compared with a manipulated sample i n which the c e l l sizes were forced to be more comparable. A smaller random sample of the peer bereaved, as well as a smaller random sample of the non-bereaved, were drawn and used in conjunction with the o r i g i n a l sample of spouse bereaved and multiple bereaved. As a res u l t , the c e l l sizes of the two drawn subsamples were comparable (250 non-bereaved, 2 50 peer bereaved) with those of the spouse bereaved (n = 2 04) and multiple bereaved (n = 118) groups. Even though, in t h i s smaller sample, the homogeneity of variance assumption was v i o l a t e d on 4 of the 7 dependent measures, Harris (1975) suggests that the robustness of ANOVA i s ensured given the large sample and the f a c t that \"the r a t i o of largest to smallest sample size for groups [is] greater than 4:1...and the variance between the largest and smallest group i s no greater than approximately 20:1\" (as c i t e d i n Tabachnick & F i d e l l , 1989, p. 324). Given the robustness of ANCOVA with comparable c e l l sizes in the smaller sample, there was confidence i n not committing an increased number of Type I errors. The vast majority ( i . e . , 13 of the 14 reported effects) of the larger sample's findings were re p l i c a t e d with the smaller sample. Consequently, analyses were conducted and reported on the more representative, larger sample of i n d i v i d u a l s . 50 Most analyses revealed s i g n i f i c a n t bereavement status main e f f e c t s . The t r a d i t i o n a l multiple comparison approach (e.g., Scheffe) using one-way ANOVAs to discern which bereavement groups s i g n i f i c a n t l y d i f f e r e d from one another was inappropriate. S p e c i f i c a l l y , one-way analysis of variance does not allow for the i n c l u s i o n of covariates; thus, analyses are computed on raw data means rather than adjusted means that have accounted for covariates. Consequently, a series of six ANCOVAs were used to examine pairs of bereavement status ( i . e . , s i x pair-wise comparisons, selecting out two levels of bereavement status for each a n a l y s i s ) . A Bonferroni procedure was used to reduce the l i k e l i h o o d of committing Type I errors; thus, an alpha l e v e l of p_ = .008 ( i . e . , .05\/6) was used to determine s t a t i s t i c a l s i g n i f i c a n c e of te s t r e s u l t s . Analyses of bereaved and non-bereaved adults' e f f e c t i v e personal and s o c i a l functioning are presented i n t h i s section. Results of hypotheses testing are organized by each of the seven dependent measures. Next presented, are r e s u l t s of regression analyses conducted within each bereavement group to determine s i g n i f i c a n t predictors of e f f e c t i v e personal and s o c i a l functioning. The f i n a l section of r e s u l t s presents findings from exploratory analyses of the peer bereaved group. Perceived Health; \"giving energy to everyday l i f e \" Hypothesis 1 predicted that the spouse bereaved would experience the lowest l e v e l of perceived health, followed by the peer bereaved, followed by the non-bereaved. As well, i t was expected (El) that multiple bereaved individuals would experience s i m i l a r or even lower levels of perceived health as compared to 51 the spouse bereaved. Hypothesis 6 predicted that women would express lower lev e l s of perceived health than men. Table 2 presents results of an ANCOVA of perceived health by bereavement status and gender, covarying number of other losses, health l i m i t a t i o n s , age, education, number of close family r e l a t i o n s h i p s , and number of close friends. Table 3 includes adjusted means and standard deviations of the perceived health measure both by bereavement status and gender. A l l s i x covariates s i g n i f i c a n t l y affected levels of perceived health; a fewer number of other losses, the absence of health l i m i t a t i o n s , being older, higher levels of education, a greater number of close family relationships, and a greater number of close friends were associated with greater levels of perceived health. Bereavement status and gender did not interac t to a f f e c t l e v e l s of perceived health. Contrary to the predictions, gender did not s i g n i f i c a n t l y a f f e c t levels of perceived health; there was a trend (p_ = .10), however, for women to report lower l e v e l s of perceived health than men. Although a s i g n i f i c a n t main e f f e c t for bereavement status was found, no bereavement status hypotheses were supported. Surprisingly, the peer bereaved experienced greater levels of perceived health than the non-bereaved. Only the expectation that multiple bereaved i n d i v i d u a l s would experience le v e l s of perceived health comparable to the spouse bereaved was supported. Negative A f f e c t : \"freedom from disturbing thoughts and f e e l i n g s \" Hypothesis 2 predicted that the spouse bereaved would experience the greatest levels of negative a f f e c t , followed by the peer bereaved, followed by the non-bereaved. As well, i t was Table 2 Analysis of Covariance on Perceived Health. Source DF F Covariates Total other losses Health l i m i t a t i o n s Age Education Number close family Number close friends 6 555.192*** 1 10.294** 1 2640.322*** 1 59.309*** 1 354.623*** 1 45.126*** 1 49.596*** Main E f f e c t s 4 6.146*** Bereavement status 3 7.227*** Gender 1 2.3 44 2-Way Interactions 3 .675 Bereavement status X Gender 3 .675 **p_ < .01. ***p_ < .001. 53 (0 rH < > (U a) X* 1 c o c (C T3 0) > CO ^ 0) PQ S-l O tw X! 4-> rH (0 tt) ffi T3 0) > \u2022H 0) O u ft o 10 c o \u2022H -p \u2022H > Q T3 CO T3 C CO -p w -c c (0 to c (0 0) 2 -o CD -P (0 \u2022r < c 0) B O Q SI CI c a) S Q SI CI a (0 -p o SI CI Cn in VO vo in in m in \u2022 \u2022 \u2022 \u2022 \u2022 CO CO 00 o H o H o H \u2022 \u2022 \u2022 \u2022 # CN CM CM CM CM r-~ cn cn t-\u00bb CO H o H vo O cn H CO in vo 00 CO in in in in in \u2022 ' H H H o H \u2022 \u2022 \u2022 \u2022 \u2022 CM CM CM CM CM CO vo in 00 CO CM o CM CM H CM CO H VO in in in in in \u2022 \u2022 \u2022 \u2022 \u2022 0 0 Xi ja XI u in cn CO CO o H o H o H \u2022 \u2022 \u2022 \u2022 \u2022 CM CM CM CM CM CO O in cn H O o H CO H CM H CM vo CO in CO -d 0) > CO tt> tt> rH ^ & 0) d) \u2022H CO X! rH \u2022P 53 1H 1 <0 r-H 0 Q) c -P & tt) 0 0 S3 w ft 53 EH co c o \u2022H \u2022p o CO u d) -p c -H \u2022p c rd O \u2022H