UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Bereavement and spouse sanctification Tweed, Roger Gordon 1997

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

Item Metadata


831-ubc_1997-0371.pdf [ 1.93MB ]
JSON: 831-1.0087769.json
JSON-LD: 831-1.0087769-ld.json
RDF/XML (Pretty): 831-1.0087769-rdf.xml
RDF/JSON: 831-1.0087769-rdf.json
Turtle: 831-1.0087769-turtle.txt
N-Triples: 831-1.0087769-rdf-ntriples.txt
Original Record: 831-1.0087769-source.json
Full Text

Full Text

B E R E A V E M E N T A N D SPOUSE SANCTIFICATION by ROGER GORDON TWEED B . A . Simon Fraser University, 1995 A THESIS SUBMITTED I N P A R T I A L F U L F I L L M E N T OF T H E REQUIREMENTS F O R T H E D E G R E E OF M A S T E R OF  ARTS in  T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Psychology)  We accent this thesis as conforming to the required standard  T H E U N I V E R S I T Y OF BRITISH C O L U M B I A July 1997 © Roger Gordon Tweed, 1997  In presenting this thesis degree  at the  in partial fulfilment  for an advanced  and study. I further agree that permission for extensive  copying of this thesis for scholarly purposes or  requirements  University of British Columbia, I agree that the Library shall make it  freely available for reference department  of the  by  his  or  her  may be granted  representatives.  It  is  by the  understood  that  head of my copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada  DE-6 (2/88)  11  Abstract Prospective and follow-up data from bereaved (n = 169) and nonbereaved (n = 211) women revealed that the bereaved tended to sanctify (rate more favorably) their husband and the marriage after the husband's death. Correlational analyses revealed that those who sanctified their husband more also tended to report more intense grief. A cluster analysis suggested three subgroups of widows: (a) satisfied sanctifiers, who tended to give their husband high ratings before bereavement and even higher ratings after bereavement, (b) dissatisfied sanctifiers, who tended to give their husband low ratings before bereavement, but higher ratings after bereavement, and (c) dissatisfied nonsanctifiers, who tended to give their husband low ratings both before and after bereavement. Possible mechanisms underlying sanctification are discussed.  Table of Contents Abstract  ii  List of Tables  v  List of Figures  vi  Acknowledgments Introduction  vii 1  Previous Findings on Sanctification  1  Sanctification Occurs  1  Retrospective Spousal Ratings Correlate Positively with Depression  1  Additional Questions  3  Do Widows Sanctify Themselves as a Marital Partner Too?  3  Do All Widows Sanctify?  4  Does Husband Sanctification Correlate with Concurrent Adjustment?  4  Do Personality and Demographics Predict Husband Sanctification?  5  Does Husband Sanctification Predict Recovery?  6  Method  6  Participants  6  Measures  7  Testing the Presence and Breadth of Sanctification Among Widows  11  Individual Differences in Sanctification of Husband, Self, and Marriage  12  Associations Between Sanctification and Adjustment  13  Personality as Predictor of Sanctification  13  iv Results  14 Sanctification Occurred for Widows  14  Widows Sanctified Themselves as a Marital Partner Too  14  Predictors of and Sanctification  ,  15  Sanctification as Predictor of Subsequent Recovery  16  Sanctification, Concurrent Outcome, and Benefits to Bereavement  16  Individual Differences in Sanctification of Husband, Self, and Marriage  17  Discussion  19  Sanctification and Adjustment  20  Mechanisms Underlying Spouse Sanctification  21  Footnotes  :  37  Bibliography  38  Appendix I: Items in Processing and Grief Scales  41  List of Tables Table 1: Widow and Control Ratings of Spouse, Self, and Marital Conflict: Means, Group by Time Interactions, and Between Group T Tests  26  Table 2: Personality and Sanctification: Correlations Of Prebereavement Personality Measures With Sanctification of Marriage and Spouse  28  Table 3: Early Sanctification as Predictor of Later Recovery (Marital and Husband Sanctification at Wave 1 Correlated with Adjustment at Wave 2 Controlling For Adjustment at Wave 1)... 29 Table 4: Concurrent Correlations with Sanctification: Correlations Of Postloss Adjustment Measures (6-12 months postloss) with Sanctification of Marriage and Spouse. 30 Table 5: Ratings of Spouse, Self, and Marital Conflict Among Bereaved Clusters: Dissatisfied Sanctifiers, Satisfied Sanctifiers, and Nonsanctifiers  32  Table 6: Distress and Grief Among the Dissatisfied Sanctifying (N=75) Satisfied Sanctifying (N=68), and Dissatisfied Nonsanctifying (N=22) Clusters  34  List of Figures Figure 1: Dendrogram for Clusters on Spouse, Self, and Conflict Rating at Baseline and Wave 1 Using Ward's Method  36  V l l  Acknowledgments I thank Dr. Damn Lehman for his time and consideration in guiding me through this thesis and for his kindness throughout our working relationship. I thank Dr. Anita DeLongis and Dr. Mark Schaller for their willingness to serve as members of my thesis committee and for their considered evaluation and feedback. I thank Dr. Camille B. Wortman who kindly provided the data for this project and also provided valuable feedback and a friendly voice. I also wish to thank Michelle Haring for her assistance in preparing the data for analysis and for assistance with conceptualization of the project and manuscript preparation. Most of all I thank Candace, my wife, and Cara and Shale, my daughters for providing a fun and loving environment at home throughout my work on this thesis.  1 Bereavement and Spouse Sanctification The word "sanctification" commonly refers to a religious act of removing sin from a person or object. In the bereavement literature, however, no religious meaning is implied when the term is used. Rather, in this context, spouse sanctification refers to the act of idealizing one's deceased partner (e.g., Futterman, Gallagher, Thompson, Lovett, & Gilewski, 1990; Lopata, 1981). Speaking of the same process, Harvey, Weber, and Orbuch (1990) used the label "canonization," which means to declare someone a saint. Previous Findings on Sanctification Sanctification occurs. Lopata (1981) was the first to provide suggestive evidence of husband sanctification when she reported data from a sample of Chicago area widows who gave unrealistically positive descriptions of their deceased husband. Although this study lacked a nonbereaved control sample, it did reveal that 3/4 of the bereaved respondents placed their husband at the most extreme positive end of scales assessing goodness, usefulness, honesty, kindness, friendliness, and warmth. Futterman et al.'s (1990), study which included a control group, was able to provide more compelling evidence for spousal sanctification. They gathered retrospective assessments of marital adjustment from the bereaved and current assessments from a comparison sample. On average, the bereaved evaluated their marriages in more positive terms than did controls. Futterman et al. included both men and women in their study and did not find any significant gender differences in the tendency to sanctify one's marriage following bereavement. Retrospective spousal ratings correlate positively with depression. Futterman et al. (1990) found that postloss marital ratings were correlated with depression such that those that rated their  2  marriage most positively had the highest depression scores, but the strength of this relation diminished over time. For the nonbereaved the relation was reversed; those that rated their marriage most negatively had the highest depression scores. Futterman et al. suggested two possible explanations for the relation between depression and positive retrospective marital ratings among the bereaved: First, they suggested that postbereavement depression may lead to more positive marital ratings (unfortunately, however, the authors did not specify the mechanism by which depression would lead to more positive recall). Second, Futterman et al. suggested that the postbereavement marital ratings may accurately reflect the quality of the marriage and the resulting magnitude of the loss due to bereavement. According to this second explanation, those widows whose husbands provided the most love, care, and support in fact lost more than others and as a result tended to give accurately high marital ratings and also to become the most depressed. With these preliminary findings as a guide, in the present study we first anticipated that women would sanctify following bereavement. We also hypothesized that sanctification would correlate positively with distress. The current dataset contains marital ratings from both before and after bereavement. The unique prospective nature of this dataset allowed direct examination of sanctification (change in spouse rating) unlike previous studies which have only been able to examine postbereavement marital and spouse ratings (these previous studies obviously have not controlled for prebereavement ratings). In all correlational analyses of sanctification in the present study, postloss marital ratings were examined controlling for preloss ratings.  3 Additional Questions A series of important, additional questions about spousal sanctification remain unexplored. In part, this is because previous studies lacked prospective data. The present study allows an examination of these significant research questions, which we describe below. Do widows sanctify themselves as a marital partner too? Previous research (Futterman et al., 1990; Lopata, 1981) has examined retrospective assessments of the spouse and marriage, but to our knowledge, no research has addressed widows' sanctification of themselves. "Selfsanctification" as used here refers to a widow's idealization of her past self specifically in regard to the marriage. To be consistent with our use and Lopata's use (1981, 1986) of the term "spouse sanctification," we have chosen to use the term "self-sanctification" to refer to this idealization of one's past self, even though this use stretches somewhat the formal definition of "sanctification." If the sanctification process relies upon the death of the person being sanctified, then only the husband will be sanctified, but if the process instead relies upon a more generalized idealization of the pre-bereavement state, then it would be more likely that positive ratings of their husband and marriage would generalize to other aspects of the marriage, for example, their own contribution to the marriage. If one truly comes to view one's deceased husband and one's marriage in idealized terms, it stands to reason that one would feel that one should have made significant contributions to such a marriage. The likely result is that one would distort the positivity of one's own contribution as well. The bereaved in the current study rated not only their spouse and their spouse's contribution to the marriage, but also the frequency of marital conflict and their own contribution to the marriage, thus allowing an examination of sanctification in each of these domains.  4  Do all widows sanctify? Though previous research has suggested that bereaved spouses typically sanctify (Futterman et al., 1990; Lopata, 1981), we were interested in the possibility that a significant subgroup may follow an opposite pattern and vilify their spouse and other aspects of the marriage. Wortman and Silver (1989) suggested that many people do not follow "typical" patterns of coping with irrevocable loss. The typical pattern of grief, according to clinical lore, calls for depression (grief) followed by recovery, but many, according to Wortman and Silver, never become depressed, or become depressed and fail to recover. Vilification, like failure to show depression or recovery, may represent an atypical, but nonetheless common response to bereavement. While speculative, a number of reasons can be suggested for some widows vilifying their husbands. For example, the husbands may have become less appealing as they became ill and died, the wives may have found it easier to dislike their husband when he was absent, or the husbands may have left their wife in financial difficulty, to name a few. Previous researchers (Futterman et al., 1990; Lopata, 1981) were unable to determine whether vilification had occurred for any given respondent. This is because they lacked prebereavement measures and so could not examine change in marital ratings. The prospective data in the current study allowed us to detect vilification of spouse, self, and marital conflict from Baseline (prior to the loss of a spouse) to Wave 1 (6-12 months following the death of a spouse) and Wave 2 (18-24 months following the death of a spouse). For a subset of respondents, we were also able to do this for Wave 3 (36-48 months postloss). Does husband sanctification correlate with concurrent adjustment? Futterman et al. (1990) found that depression was associated with positive retrospective marital ratings among the recently bereaved. The current, prospective dataset allowed a direct examination of the relations  5  between adjustment and sanctification (post-bereavement marital rating controlling for prebereavement rating). Besides including a measure of depressed mood as did Futterman et al., we included measures of positive affect, anxiety, grief (comprised of seven subscales), processing of the loss, and self-reported benefits from the loss. We also generated a general distress index composed of depression, anxiety, and negatively weighted well-being. This use of multiple outcome measures allows a more precise examination of the relations between sanctification and adjustment than would use of only a single outcome measure. Do personality and demographics predict husband sanctification? Prior studies, lacking prebereavement spousal ratings by the bereaved, were unable to distinguish sanctifiers from vilifiers, and consequently could not adequately address predictors of sanctification. We were interested in the role that individual differences might play in sanctification. Axe widows with certain constellations of personality traits more likely to sanctify than others? To assess these relations, we used the Big Five dimensions of personality, as conceptualized by Costa and McCrae, to predict sanctification among widows. Previous research has suggested that these personality traits are stable after age thirty (Costa & McCrae, 1988; Costa & Metter, 1994) and are better predictors of well-being than are life circumstances (Costa & McCrae, 1987). Widows in this study completed a measure of the Big Five personality dimensions before they were bereaved. An additional set of exploratory variables was assessed as potential predictors of sanctification. These included age of the bereaved, family income, personal income, education, caregiving prior to the death, and illness of the spouse at baseline.  6  Does husband sanctification predict recovery? Lopata (1981; see also Nahmani et al., 1991) suggested that sanctification can help the bereaved detach from the deceased, in part because a sanctified spouse becomes more unreal and distant, and that this detachment helps the bereaved complete the process of grief work. However, a number of theorists view spouse sanctification much less positively. Rando (1984, 1993; see also Rubin, 1985), for example, has argued that spousal sanctification inhibits recovery because a composite recollection of a partly good and partly bad marital relationship is necessary to legitimize any ambivalent or negative feelings about the deceased, thereby allowing the bereaved to work through them. Rando recommended that caregivers intervene to overcome sanctification by discussing both positive and negative aspects of the deceased. Bowlby-West (1983) took a similarly cautious view of sanctification, suggesting that sanctification by the bereaved is a homeostatic adjustment to being stuck in one of the phases of grief and that sanctification has dysfunctional consequences. Lieberman (1978) categorized sanctification of a grief object as a form of morbid grief. These suggestions in the literature that sanctification is either adaptive or maladaptive were explored in the current dataset. In particular, the relation between sanctification and recovery was examined by testing whether sanctification predicts subsequent recovery. Method Participants One thousand and fifty eight women were initially interviewed in 1987 (hereafter referred to as the Baseline interview) as part of the Changing Lives of Older Couples (CLOC) study.  1  Persons were eligible for the study if they were married and the husband of the couple was at least 65 years of age. Respondents whose husband died after the Baseline interview were contacted for  7  follow-up interviews 6-12 months (Wave 1), 18-24 months (Wave 2), and 36-48 months (Wave 3) following loss. A subsample of the nonbereaved respondents (n = 211) was also contacted for follow-up interviews in order to obtain nonbereaved control data. The Baseline respondents had an average age of 68 years (range: 38-87) and had been married an average of 41 years at Baseline. They had an average of 11.7 years of education (range: 0-17). Nine hundred and forty four of the baseline respondents were White and 114 were African American. Of the 1058 women interviewed at Baseline, 169 widowed women were included in the current study. This represents all widows with data at Baseline and the first two postbereavement interviews. These widows had an average age at baseline of 69 years (range: 48-87) and had been 2  married an average of 42 years. The widows had an average of 11.5 years of education (range: 417). One hundred and forty three were White and 26 were African American. Two hundred and eleven female nonbereaved controls were included in the study. They had an average age at baseline of 69 years (range: 47-85) and had been married an average of 42 years. They had an average of 12 years of education (range: 5-17). One hundred and eighty one were White and 30 were African American. Measures As part of a more extensive interview, respondents rated aspects of their marriage including their treatment of their husband, their husband's treatment of them, and the frequency of unpleasant conflict in the marriage. Ratings from this series of questions represent the key dependent variables for this study. In particular, spouse ratings were assessed with a composite of four items (alpha=.77) asking how much the husband cared for the respondent, listened to the respondent, made too many demands on the respondent (reversed), and treated the respondent as  8  well as she deserved. Self ratings were measured with three items (alpha=.62) asking how much the respondent made the spouse feel cared for, how much the respondent listened to the spouse, and the extent to which the respondent treated the spouse as well as he deserved. Frequency of marital conflict came from a single item. All of these eight aforementioned items were entered into a principal components analysis. All items loaded over .55 or higher on the first principal component. A composite measure of marital evaluation (alpha=. 82) was constructed from these eight items. Item-level data were standardized before being summed into marital scales and the scales were converted to T score metric based on the 1058 women interviewed at Baseline. T scores have a mean of 50 and standard deviation of 10 in a reference population (the 1058 women at Baseline in this case), and thus in many cases are easier to interpret than raw scores. For widows, but not for controls, the Wave 1 and Wave 2 marriage-related ratings were necessarily retrospective to the time before the loss. The retrospective nature of the widow ratings and the contemporaneous nature of the control respondent ratings could potentially create a confound when comparing these groups if, for example, people universally glorify relationships when reporting retrospectively, but we know of no evidence for such a possibility. To overcome this possible confound, control respondents could have been asked to retrospectively rate their current spouse, but we would expect control participants' retrospective ratings to differ little from, and be largely inferred from, their current ratings of their partner as suggested by the work of McFarland and Ross (1987). If sanctification does result from a simple retrospection effect, then the effect should become stronger as more time passes after the husband's death. This possibility that a sanctification effect would become stronger over time was examined in the current study.  9  A significant ceiling problem existed with the marital evaluation scales. At the Baseline timepoint, 6% of the to-be-widowed women gave the highest possible rating on the marital evaluation scale, so sanctification after this time point could not be assessed for these people. The ceiling problem was more severe for the subscales (16%, 29%, and 17% of the sample reported the maximum positive values for the husband, self, and conflict subscales respectively at baseline). We chose a conservative approach and included all available participants in each reported analysis of these scales, even though this necessarily reduced the chance to observe sanctification. Of these subscales, only the husband subscale was entered into correlational analyses. As a precaution, each of the correlational analyses with this subscale were conducted both with and without women giving the maximal rating at Baseline. All significant effects increased in magnitude beyond those reported in this paper when this portion of the sample was dropped from these analyses, so we concluded that the findings were not attributable to a ceiling effect. The frequency and intensity with which the widows were cognitively processing their loss at Wave 1 were assessed with eight items (alpha=.74) including "At times, I want or need to think about my husband or his death" and "During the past month, how often did you actually talk about your husband or his death with anyone?". All eight items are displayed in Appendix A. Personality was measured with a 43-item experimental form of the Five Factor Inventory, designed with Paul Costa (see Costa & McCrae, 1992). The dimensions measured included Neuroticism, Extroversion, Openness to Experience, Agreeableness, and Conscientiousness. Alphas in this sample were low for Extroversion (.49) and Openness to Experience (.43) and only somewhat higher for Neuroticism (.66), Agreeableness (.58), and Conscientiousness (.73). Prior  10  to the study, a decision was made to broadly sample the domain of each construct. This strategy was pursued to maximize the content validity even though it reduced internal consistency. Depressed mood was measured with an 11-item version of the Center for Epidemiological Studies Depression Scale. The 20-item CES-D Scale (CES-D; Radloff, 1977) has been shown to meaningfully differentiate depressed participants from controls (Boyd & Weissman, 1982; Craig & Van Natta, 1979). Kohut, Berkman, Evans, and Cornoni-Huntley (1983) reported that the 11item short form used here produced an alpha of .81 and correlated .95 with the full 20-item scale. As mentioned previously, other adjustment constructs, including positive affect, anxiety, and grief were also assessed. Positive affect was measured with the Positive Affect Subscale (PAS; alpha=.79) of Bradburn's Affect Balance Scale (Bradburn, 1969). Each of the 5 PAS items describes a positive affect, and respondents reported how much they had experienced each affect in the past week. Representative items include "I felt on top of the world" and "I was particularly excited or interested in something." The Bradburn items provide a measure that is less stable over time than personality, but more stable than measures of current mood (for reviews see McCauley & Bremer, 1991; Andrews & Robinson, 1991). The PAS has produced low correlations (i.e., below .10) with measured negative affect when the response format asks for raw counts of times each positive affect has been recently experienced, but significantly correlated with negative affect when as in the current study the response options assess the proportion of recent time during which these positive affects have been experienced (Warr, Barter, & Brownbridge, 1983). Anxiety was measured with 10 items from the Symptom Check List 90 Revised (SCL-90R; Derogatis, 1977). The alpha for the anxiety items in this sample was .76. Items included "How much have you been bothered by the feeling that something bad is going to happen to  11  you?" and "How much have you been bothered by feeling tense and keyed up in the past seven days?" Principal components analysis of the CES-D, SCL-R anxiety, and PAS scales with the 1058 women at Baseline suggested a single component with an eigenvalue greater than one and with all variables loading .68 or higher. Likewise, principal components analyses for widows at Wave 1, Wave 2, and Wave 3 each produced only one component with an eigenvalue greater than one. Based on the association observed between these three adjustment measures, a distress index was constructed by unit weighting standardized scores on the CES-D, SCL-R anxiety, and PAS (negatively weighted). Grief was measured with a 23-item scale constructed for this study (alpha = .88; see Appendix A). Items were chosen to broadly sample the conceptual domain of grief. Seven subscales of the grief scale included anxiety, anger, despair, negative responses to reminders, longing, rumination, and shock. The average alpha for these subscales was .69 with a minimum alpha of .59 for the anger subscale and a maximum alpha of .75 for shock. Testing the Presence and Breadth of Sanctification Among Widows Repeated measure ANOVA analyses were conducted to check for group (widow vs. control) by time (pre- vs. postloss) interactions. A significant group by time interaction would indicate that the bereaved and nonbereaved followed different patterns over time in their ratings of spouse, marital conflict, and/or self. We controlled for income and education at Baseline in the repeated measure analyses because the bereaved reported a lower income at Baseline than did the nonbereaved, t(379)=3.41, £=.001, and a lower level of education (11.5 years on average for bereaved, 12 years for controls), t(379)=2.04, p_=042. Following the test for an interaction, t  12  tests were conducted to compare widows to controls at each time point to clarify how the patterns of ratings differed between the bereaved and nonbereaved. A significant constraint limited the analyses: Only a minority of the nonbereaved control respondents were interviewed at both Waves 1 and 2. Due to funding limitations, only 73 female control respondents were interviewed at Wave 1. A higher number of female control respondents were interviewed at Wave 2 (n=171), but only 33 of these had been interviewed at Wave 1. To address this constraint, within-subject analyses with the bereaved and controls were made from Baseline to Wave 1 and Baseline to Wave 2, but never from Wave 1 to Wave 2 because including both Wave 1 and Wave 2 would have reduced the number of available control respondents to 33. Individual Differences in Sanctification of Husband, Self, and Marriage To differentiate subgroups based on their pre- and postbereavement marital ratings, we cluster analyzed the bereaved sample on the ratings of their husband, self, and marital conflict. Cluster analysis merges cases into subgroups based on any number of chosen input variables. Within any particular subgroup (called a cluster) resulting from this procedure, cases will have somewhat similar profiles on the chosen input variables. We used Ward's method of cluster analysis combined with a squared euclidian distance measure (some other common methods of cluster analysis, such as the average distance between groups method or the complete linkage method, tend to more frequently produce clusters too small for inferential analysis). Because we expected that most of the change in ratings and resulting profile differences would occur between Baseline and Wave 1, we limited our input variables to ratings of husband, self, and marital conflict at Baseline and Wave 1.  13  Associations Between Sanctification And Adjustment We assessed relations between sanctification and concurrent adjustment among the widows by calculating correlations between adjustment and sanctification as measured by the overall marital evaluation and the husband subscale (the self and conflict subscales were not the primary variables of interest, so to reduce the number of analyses they were not examined separately in correlational analyses). Adjustment measures for these analyses included the distress index, depressed mood, anxiety, grief, and positive affect. The relations between sanctification and adjustment were further examined by conducting ANOVA's comparing the levels of distress and grief among the clusters of women who did and did not sanctify. We also evaluated whether sanctification predicted subsequent recovery by calculating semipartial correlations between sanctification at Wave 1 and adjustment at Wave 2 (controlling for adjustment at Wave 1). Widows were also asked to what extent, reported on a 5-point scale, there were benefits to being bereaved. Correlations between reporting benefits and sanctifying one's marriage and spouse were also examined. Sanctification for these and all other correlational analyses reported in this study was defined as postbereavement marital rating (overall scale and spouse subscale) controlling for prebereavement marital rating. Personality as Predictor of Sanctification The power of personality to predict sanctification among the widows was assessed by examining correlations of the Big Five personality dimensions measured at Baseline with sanctification as assessed by the overall marriage evaluation and the husband subscale.  14  Results Sanctification Occurred For Widows Marital ratings by bereaved and nonbereaved respondents are listed in Table 1. The bereaved produced a different pattern of overall marital evaluations over time than did controls (Baseline to Wave 1: F(l, 240)=14.26, rj<001; Baseline to Wave 2: F(l, 338)=18.79, p<.001). Simple effect analyses revealed that the groups provided similar marital evaluations at Baseline (t(378)=.27, p_=787), but the widows provided more positive marital evaluations after their loss (Wave 1, t(241)=3.84, p_<001 and Wave 2, t(339)=4.56, p_<001). A similar pattern emerged for the spousal evaluation subscale. Group by time interactions were significant (Baseline to Wave 1: F(l, 240)=13.52, p_<001; Baseline to Wave 2: F(l, 338)=14.99, p<001) and simple effect analyses revealed further that although bereaved women tended to give higher ratings of their spouse after bereavement than did controls (Wave 1, t(241)=3.14, p=.002, and Wave 2, t(339)=3.48, p=.001), they did not do so before the death (t(378)=.19, p_=849). The pattern also replicated for the marital conflict subscale with widows reporting less conflict than controls at Wave 1, t(241)=4.62, p<001 and Wave 2, t(334)=4.61, p<.001, but not at Baseline, t(372)=.22, p=.827. Widows Sanctified Themselves as a Marital Partner Too The bereaved also appeared to have sanctified their own role in the marriage. Selfsanctification refers here to an idealization of the prebereavement self with particular reference to orientations toward the husband. The group by time interaction was nonsignificant, F(l,240)=2.77, p=.098, in the Baseline to Wave 1 analysis, the analysis for which the number of female controls was lowest (control n=73 at wave 1 versus n=171 at wave 2), but significant, F(l,  15  338)=6.80, p_=.010, in the Baseline to Wave 2 analysis where power was higher. T-tests clarified the group differences by revealing that the widows produced self-ratings similar to those of nonbereaved women, t(378)=.84, p=.400, at Baseline, but generated higher self-ratings at Wave 1, t(241)=2.97, p=. 003 and Wave 2, t(339)=4.21, p_<001, than did the nonbereaved women interviewed at those times. Taken together, the findings suggest that the sanctification effects were relatively stable for the first two waves of data collection. The means for the third wave (36-48 months postloss) of data collection are not displayed in the table, but on each subscale the mean ratings moved marginally back toward their baseline levels. If mere nostalgia explained sanctification, then one would expect the marital ratings to become more and more favorable as time passed. This opposite pattern goes against the suggestion that sanctification merely reflects an increasingly positive nostalgic feel as the past becomes more distant. Predictors of Sanctification The relations between the Big 5 personality dimensions and sanctification as measured by the overall marital evaluation scale and the spouse subscale are reported in Table 2. Extroversion was positively correlated with both overall marital sanctification and spouse sanctification, but the relations were modest in both cases (r[169]=.17, p_<05, and r[169]=.16, p_< 05, respectively). Demographics (family income, personal income, and education) and other factors including caregiving prior to the death and illness of spouse at baseline were also examined as possible predictors of spouse and marriage sanctification. Family income (r[169]=.20, p_<.05) and personal income (r[169]=17, p_>. 05) predicted overall marriage sanctification and also predicted (r[169]=. 20, p_<05, and r[169]=. 16, p_<05, respectively) spousal sanctification. All these  16  "significant" personality and demographic correlations deserve to be treated tentatively because the relations were all modest and because even though 20 correlations were calculated, none of the p_ values were below .01. Sanctification as Predictor of Subsequent Recovery We assessed whether sanctification could predict subsequent recovery (see Table 3). More specifically, we calculated semipartial correlations of sanctification with adjustment at Wave 2 controlling for adjustment at Wave 1. Adjustment measures included measures of grief, depression, anxiety, positive affect, and a composite distress index (depression plus anxiety minus positive affect). As in all correlational analyses described herein, sanctification was defined as postbereavement marital rating controlling for prebereavement marital rating. None of the correlations between sanctification and subsequent recovery were significant. Sanctification. Concurrent Outcome, and Benefits to Bereavement Concurrent correlations with sanctification are displayed in Table 4. All measures of poor adjustment had positive relations with sanctification and the measure of positive affect had negative relations with sanctification. Though most of these relations were not statistically significant on their own, the likelihood of this pattern occurring by chance was less than .001 based on the binomial distribution. These data suggest that sanctification is negatively related to a general construct which could loosely be labeled "adjustment" with widows who sanctified more also tending to report poorer adjustment. The highest correlations were observed with the grief scale and subscales. Correlation coefficients for grief were r(169)=.23 (p_<01) with overall marriage sanctification and r(169)=. 26 (p< 001) with spouse sanctification. Of the grief subscales, the longing subscale produced the  17  highest positive correlation with both overall marriage (r(169)=.27, rj<001) and spouse (r(169)=.27, rj<001) sanctification. In line with this set of correlations, widows who sanctified more tended to report fewer benefits from bereavement (r(169)=-.37, p<.001 for overall marriage scale, and r(169)=-.40, rj<001 for spouse subscale) than other widows and were more likely to report intense cognitive processing of the loss (r(169)=.16, p<05 for overall scale, r(169)=.21, P_<.01 for spouse subscale). Individual Differences in Sanctification of Husband. Self, and Marriage Change scores revealed that many widows (66%) gave more positive evaluations of their marriage after the death than before on the overall marital evaluation scale. Widows increased their marital evaluation scores by up to three standard deviations. Likewise, many widows gave more favorable ratings after the death on the spouse (60%), self (50%), and conflict (42%) subscales. Not all widows, however, sanctified in all these areas. A significant subset of widows (23%) gave more negative overall marital evaluations after the death than before, producing decreases of up to 2 standard deviations in their marital evaluation. Similarly, a subset gave less favorable ratings of spouse (21%), self (18%), and conflict (14%) after the death than before. Widows decreased their spouse rating by up to 3 standard deviations, their self rating by up to 2.5 standard deviations, and increased their marital conflict rating by up to 3.5 standard deviations by the time of the Wave 1 interview. These individual differences in sanctification were explored with a cluster analysis. After clustering the bereaved sample on ratings of spouse, self, and conflict at Baseline and Wave 1, we examined the means for the clusters on the input variables and examined the dendrogram (Figure 1), a graphic illustration of multivariate distances at which cases were combined to create clusters.  18  The diagram can be read from left to right, and longer horizontal lines in the dendrogram indicate more stable clusters. We determined that a three cluster solution provided the best clarification of the data and was also most acceptable according to the dendrogram. Means for the clusters are displayed in Table 5. Repeated measures ANOVAs confirmed that the groups followed different patterns over time in ratings of marriage, spouse, self, and marital conflict (group by time interactions, which indicate differing patterns of change, were significant at p_<001 for all four variables). After examining the marital ratings of each group, we labeled the groups "Dissatisfied Sanctifiers" (n=75), "Satisfied Sanctifiers" (n=68), and "Dissatisfied Nonsanctifiers" (n=22). The total number of widows included in the cluster analysis was 165 because only widows with complete data on the three relevant variables were included. The centroid for the Dissatisfied Sanctifiers indicated that they tended to produce less favorable than average ratings of spouse (a T score of 46.92 with metric based on the 1058 women interviewed at Baseline), self (47.89), and conflict (54.06) at Baseline, but sanctified on all three dimensions by Wave 1 (54.37, 52.58, 46.93). We labeled them dissatisfied because they seemed somewhat dissatisfied with their marriage at Baseline. Satisfied Sanctifiers tended to start with more favorable than average ratings of their spouse (55.66), self (56.71), and conflict frequency (43.17). By Wave 1 these already satisfied respondents nonetheless tended to sanctify their spouse (59.59), self (59.97), and frequency of marital conflict (40.94). We expect that the centroid for this group would have shifted even more between Baseline and Wave 1 had many of them not already been at or near the most extremely favorable points on the scales at Baseline. Dissatisfied Nonsanctifiers tended to start with much lower than average ratings of their spouse (35.54) and of themselves (40.81). At Wave 1 they again gave low ratings of their deceased spouse (32.10), but higher ratings of self  19  (45.10). The Dissatisfied Nonsanctifiers reported a higher than average frequency of marital conflict at Baseline (56.20), and their reported frequency of conflict changed little over time. The adjustment reported by widows in these clusters was examined with one-way ANOVA's and Tukey HSD follow-up tests as shown in Table 6. The results suggest that the groups differed on their reported level of grief at Wave 1, F(2, 162)=4.22, p=.016), with the Dissatisfied Nonsanctifiers reporting the least grief. On the distress index, however, the groups only differed prior to the loss, F(2, 160)=4.85, p_=009), with the Dissatisfied Nonsanctifiers reporting significantly more distress prior to the loss than the Satisfied Sanctifiers. The education, age, and income of the three groups were also compared in one way ANOVA's with Tukey HSD follow-up tests. The results indicated that the Dissatisfied Nonsanctifiers reported lower income both before, F(2, 162)=4.81, p=.009), and after, F(2, 151)=6.03, p=.003), the death than did either of the two sanctifying clusters. Discussion Data from this first prospective study of sanctification indicated that, in general, widows sanctified their marriage and their husband. These data also revealed that the process underlying sanctification tends to have considerable breadth such that widows also increased their rating of their own role in the marriage. Thus sanctification appears to be a reasonably broad process transforming numerous aspects of memory for life together with the now-deceased spouse. It proved important to pursue an idiographic approach as well because we discovered that sanctification was not universal. A significant subset of women failed to sanctify or even vilified their spouse, and to our knowledge this result has not been previously reported. Women who sanctified little or not at all tended to report less grief, less cognitive processing of the loss, and  20  more benefits to widowhood. A modest correlation suggested that these women may have also had lower incomes. The cluster of women who failed to sanctify, and were labeled Dissatisfied Nonsanctifiers, produced a pattern of results suggesting a somewhat more difficult life prior to the loss than was evident among the other two clusters. The Dissatisfied Nonsanctifiers reported higher distress prior to the loss than did the Satisfied Sanctifiers, lower marital satisfaction than either of the other groups, and lower income both before and after the loss than either of the other two clusters. These reports of pre-loss distress and limited income could be interpreted as suggesting that these women were engaged in caring for ill spouses prior to the loss, but further examination of this possibility revealed no differences among the clusters in proportion of women providing care at Baseline. Therefore, their higher reports of distress at Baseline cannot simply be attributed to caregiving. While speculative, a number of reasons can be suggested for why this cluster failed to show the sanctification pattern. For example, the husbands may have become less appealing as they became ill and died, the wives may have found it easier to dislike their husband when he was absent, memories of a poor marriage may have restrained their ability to sanctify, or the husbands may have left their wife in financial difficulty. In agreement with the latter possibility, the Dissatisfied Nonsanctifying widows reported lower income both before and after the death than either of the other two clusters. However, it should be noted that these possible interpretations are speculative at this point. Sanctification and Adjustment Past and present research cannot directly address sanctification's causal potency for aiding or hindering recovery. Nonetheless, we can examine how sanctification fits into the pattern of coping with grief. Rando (1984, 1993) and Futterman et al. (1990) suggested that excessively  21  positive memories of one's deceased spouse will hinder recovery. Lieberman (1978) argued that sanctification of the deceased represents a form of morbid grief. In contrast, Lopata (1981) suggested that postbereavement spouse sanctification benefits the bereaved by enabling detachment from the increasingly unreal sanctified other. The current data suggest that although sanctification is associated with more severe grief in the period shortly after the loss, sanctification does not predict poor future adjustment. The present findings must be treated tentatively because null results may resultfromeither absence of an effect or failure to detect an effect, but the results are nonetheless important given that some therapists may assume that they should work to combat sanctification by helping the bereaved develop a more accurate (i.e., less positive) memory of their spouse (e.g., Parkes & Weiss, 1983; Rando, 1993; Worden, 1991). Our correlational evidence suggests that in some cases, encouraging a client to develop a more realistic appraisal of their deceased spouse may waste valuable therapy time. Clearly, when a client has become preoccupied with an idealized notion of the deceased that is interfering with other relationships (see for example Rando, 1984; Rubin, 1985), working toward a more realistic memory may be required. However, our findings suggest that "positive illusions" of a deceased spouse, like positive illusions of the self (Taylor, 1989), may not be as maladaptive as some have assumed. Mechanisms Underlying Spouse Sanctification Mechanisms underlying sanctification remain something of a mystery because of the correlational nature of the data, but some mechanisms that may have contributed to sanctification deserve comment. These include that (a) a simple retrospection effect increased the widow's ratings of their spouses, (b) widows, in response to societal pressure, desired to speak well of the  22  deceased regardless of their true feelings towards him (Harvey, Weber, & Orbuch, 1990; Lopata, 1981), (c) widows spoke especially highly of the deceased as a way of masking suppressed anger toward him (Worden, 1991), (d) widows no longer experienced negative interactions with the spouse that could disconfirm high evaluations (Rubin, 1991), and (e) widows reconstructed memories of their marriage with reference to their current dysphoric state. The first suggestion that a simple retrospection effect caused sanctification would lead us to expect evidence for such an effect outside of the bereavement context, but we know of no evidence for such an effect. The effect, if linear, would lead to a continued increase in marital ratings as more time passes after the death, but this pattern was not evident in the present data. Also, we would not expect sanctification to be correlated with poor adjustment if sanctification were due solely to a simple retrospection effect. We conclude that sanctification is likely not due merely to retrospection. The second suggested mechanism for sanctification, that widows responded to societal pressure to speak well of the deceased (Lopata, 1981), while possible, cannot fully account for the sanctification pattern observed in the current analysis. Widows sanctified not only memories of their husbands, but also memories of themselves as marital partners. This finding of selfsanctification cannot be explained by a simple desire to speak well of the deceased. The third suggestion, that some widows who sanctify are suppressing anger toward the deceased, may be true in particular cases as Worden (1991) suggests. This anger could have existed prior to the death, or may have developed subsequent to the death. If, however, sanctification consistently indicated suppressed anger, then one would expect to find that sanctification would predict poor subsequent outcome (Worden, 1991). In the current study, no  23  significant relation was observed between sanctification and subsequent recovery. This null effect fails to support this universal suppressed anger hypothesis, but as with most null results, should not be taken as evidence strongly refuting the hypothesis. A fourth possibility suggested by Rubin (1991) is that widows sanctify because their husband is no longer present to disconfirm the widow's idealized view of him and the marriage. This mechanism presupposes that women generally gravitate towards favorable evaluations of their husband and marriage, but that inevitable negative interactions with a living partner work against this force. Once the husband passes away and negative interactions no longer occur, the tendency toward positive evaluations becomes more likely unless the process is restrained by unusually prominent memories of negative interactions. It seems to us that this presumption has intuitive appeal. This fourth possible contributor to sanctification need not be limited to the bereavement context. The same effect could occur following other involuntary relational separations such as when one party in a relationship must relocate for medical treatment or when one party is sent to prison or to war. As afifthpossibility, memory reconstruction influenced by a simple contrast effect may have played an important role in generating sanctified memories. Lopata (1986) suggested that grief work requires that the bereaved reconstruct the past tofitthe present. Husband sanctification may represent one example of this type of reconstruction. The idea that memories are reconstructed is not new. Bartlett (1932/1967) suggested over half a century ago that memories are strongly constructive, usually inaccurate, and frequently altered to make them more satisfying. More recent research suggests that people infer autobiographical and other memories from current knowledge, beliefs, and experiences and that these inferences are frequently in error  24  (e.g., Conway, 1990; McFarland & Ross, 1987; McFarland, Ross, & Giltrow 1992). Sanctification may result from this type of inference in which the present functions as a benchmark for reconstructing the past because the present is salient and available in memory (Harvey, Weber, & Orbuch, 1990). The reconstruction process for memories of the spouse, self, and marital conflict may operate as follows: As part of the grief process, the surviving spouse may mentally contrast the current dysphoric state (which may be caused by a constellation of factors) with the previous married state. Memory reconstruction may be influenced by the widow's belief that the previous marital experience was very different from the current dysphoric state. The act of reconstructing memories based partly on the belief that the past marital experience contrasted with current dysphoria could lead to sanctification of numerous aspects of the prebereavement state including sanctification of the deceased spouse, the self, and the frequency of marital conflict. This suggested mechanism concurs with the finding that sanctification affects not only memories of the spouse, but also memories of other aspects of the marriage including marital conflict and the self. Further, we would expect that if dysphoric experience in large part drives sanctification, then sanctification would be most evident among those reporting the most dysphoric outcomes and that as dysphoria dissipates with the passing of time, the sanctification effect would weaken as well. All of these expectations are in line with the present data. In sum, we have suggested four processes that may contribute to postbereavement sanctification: a response to societal pressure to speak well of the deceased, a suppression of anger toward the deceased, a new evaluation of the marriage owing to an absence of negative marital interactions with a living spouse, and a reconstruction of marriage memories with reference to a belief that marital experience contrasted with the experience of post-bereavement  25  dysphoria. The memory reconstruction process bestfitsthe current data, but all these factors may play a role in sanctification. Clearly, further research is needed to clarify the mechanisms underlying sanctification.  26  vi C  <N  o on  •c  o o V  o o  o o V'  o o  roo  ON  ed  I*  o U a, o  6  "S en cd  oo  ffl  a C  _ mo o Cd  t-i  ts I—I  u  P  B o  o o V  (L> Vi  a  ffl  a  H >%  o  o o V  O  CD  o  cd  v"  m  m  ON CN  00 ON  in  IT)  ON  o  >/->  m >n  ON Vi  VO  ON  Vi  3  e  o o,  cd  O  c  '•3 ed  ON ON  (N  OS  ON  H  o  2  3  cd >  w  13 -•-»  •c  vo ON  ON  O  c o U  u  Vi  O T3  Vi  "cd O Vi  X>  o U  3  t« <L>  O  a.  27 00  o o V  X  tO=  o o V  CD  CN  co  o o V  at  T3  o o  > CD  c  CD  co  CN  S o  CD  o o  CN  00  00  m K3  o o XI  CD CD  & <D UH  <D  o o  o o  X  1-1  co CD co  13  ON O  CD  $ c o  *o CD  co aj X>  CD  '£  Kl  X  CD CD  CD  °  s  co X  CN  "S  CN  j£  O  00  o  c  cn  o o  '£  CD VH  O  o  VO  co  CO  CD CD  CD  a> CO  in  ON IT)  vo  d  -cr' m  in  IT)  m © m  O  C  vq in  ON  ON  m  in  in  o in  X  CD  .S  —  CD CD  co  +3  a ON  t-c  c  CD 00  CD  '3  -t-> co  <D X CD XI CD  a,  o vo 00  B  ON  ON VO  ON  o o  o  m  in  oo ON  •«a-  CD  O  U  o  "c3  o CO  -a O  ccS  "S, o o  U  X C4H  "CD  CO  a  o 13  CD  O  U  ccj 'CD  o  o CD  %  o o o  CD  co  CO co  X  O  "S  CO  •g co  O0 C  <D  CD  'co  CD  C O  e  co  CD CD  O CD  u, CD X Ui  o c  CD  ccj  CD  X  Table 2 Personality and Sanctification: Correlations Of Prebereavement Personality Measures With Sanctification of Marriage and Spouse  Sanctification at Wave 1 (6-12 months after bereavement)  Marital  Spouse  Evaluation  Subscale  Neuroticism  .01  -.01  Extroversion  .17*  Openness to Experience  .00  -.06  Agreeableness  .03  .06  -.07  -.04  Conscientiousness  Note: * p<=.05  .16*  * p<=. 01  Note: Sanctification is defined as postloss marital rating controlling for preloss rating.  Table 3 Early Sanctification as Predictor of Later Recovery (Marital and Husband Sanctification at Wave 1 Correlated with Adjustment at Wave 2 Controlling For Adjustment at Wave 1)  Overall Marital Sanctification Distress Index  -.03  Grief  -.08  CES-Depression  .03  Anxiety  .04  Positive Affect  .12 Spouse Sanctification  Distress Index  -.05  Grief  -.06  CES-Depression  .02  Anxiety  .00  Positive Affect  . 11  Note: None of these correlations were significant at £<=. 05. Sanctification is defined as Wave 1 marital rating controlling for preloss rating. Wave 1 and Wave 2 were both postloss (6-12 months and 18-24 months respectively).  30  Table 4 Concurrent Correlations with Sanctification: Correlations Of Postloss Adjustment Measures (612 months postloss) with Sanctification of Marriage and Spouse  Sanctification  Overall Marital Sanctification  Spouse Sanctification  Distress Index  .10  14  CES-Depression  .06  .12  Anxiety  .06  .06  Grief  .23**  .26***  Anxiety  .15  .19*  Anger  .15  .16*  Despair  .21**  .25**  Negative Response to Reminders  .15*  .16*  Longing  .27***  .28***  Ruminations  .09  .14  Shock  .10  .13  .16*  .21**  Grief Subscales  Processing Positive Affect  -.10  -.12  31  Benefits to Bereavement  Note: * p_<=05  ** p<=. 01  _37***  _40***  *** p_<=001  Note: Distress index is a composite of depression, anxiety, and negatively weighted positive affect. Sanctification is defined as postloss marital rating controlling for pre-loss rating.  32 CN  o o V  o  a,  <D  o o v'  =3 O  cS C  <D CD  PQ  "3 co  o o V  ed PQ  CN  o cd CD  CD  5 CD  X OH  3  o  c5  CD _g  "3  o o v'  Ol  PQ  CN  1—1  CO  oo  CD  © VO  CO  m  00 •—I  00 CO  CN CN  CO  rr in  ,V— i O  VO CO  m  in in  CN ON  ON  CN  CO  cd  m  CD  IT)  o oo  m CO  e CD  S3 cd CD  vb  *3  CO  cd PQ  e 0 cd  t> C cd co  C  o  ON  vb  vb  co  -cr  c? in  CD  in  1 2  SI  oo  CN  vo  S  -a  CD O  <u CO CO CO  co  X  CO  CO  3 00  '"cd  O  co  CD  on  cd  C  CD  3 O  OH  T3 <D  to co  "3 CO CO  33  o o  o o  o o  o o  o o  o o  o o  o o  o o  V  V  V  V'  v'  V  V  V'  v'  CD  .5 co cd .£>  ts CD  o  o o  oo m  o  V  U  G cu  V  V  s o  oo ON  00 m  vo  ON  co-  >—<  m  i—i  in  in  in  00  in vo  ON  00 m  o  CD  c o  in  T3 0) co cd  ON  m ON  m  o  >—i  ON  ON  CN  VO  o  o  00 m CN  CN vo  >n  m  t>  o  ON  in  ON  m  >—<  X> o  •c  e cd •G  vo vo VO  m  vo o  m  in m  in  tm  o CN  ON  00  r-'  VO  in  T3  rvd in  00 o  CU  O o CO  cu cu  cu  -*->  HH •»^»  T3 CU CO  %-» cd C/3  H3-  o  O  G  G cd G O  o U cd  -y  cd  G cd  OJ -a  u G cd  CO CO  Q  "O CU  G cd  o  G  3  CO  CO  cd  O  C/3  C/3  CU O  CO  CU  <G  '•t3  <u tG '"cd  CU  CO  Cy  00  tG  CO  '3  '•*3 o  CO  G cd C/3 T3  cu tG  co  CO  CU  "i  C/2  60 G  o  G cd G O  CU -*-»  o  34  a cd  oo vo  >> CD  H OJ  a ST o  c  s  cd CvQ  Q  CD  A  V  vi  CN  CN  X  2  ON  o  o o  o  T3  CO  CO  VO VO  rm ON  VO  CO  T—1  CN  o  00  cd  C/3  CO  m  1 3 <D  CO  9  CD  S3  o  CD  m  CO  in  a cd c  r—1  in  vo  T—1  ON  m  m  o  ,—<  ,—i  >—i  co' 'cr  o  in  CO  O  o  c cd  &0 1 3 CD  co cd CO  co  CO  13  -iCO ->  CJH CO  t-l CD  o CN CN  CD  CD  S3  CN  o  a  00  in  m  CO  vo in  C «§'  cd CM <D  CD  c  CO  o o in  CD  a  co  CO  — , I  <-i  CN  CD  CD  m  CN  oo o m  m  in  CN CD  cd  CO  m  CD  CO CO  CO ON  cd  CD  CD  vo  C/J  c  13  cd CO CO  CO CO  CD  S 'O3  CD  C  13  CO  O  5  in  CN in  73  cd CO  c  o  cd  CD  -4-*  CO  VI  ta  <D  m  in  CD  5  CD  35  cu -a c  Q  X  cd  H->  o  •s cu  >  0 OH  CU  op  T 3  CU Si O O  HI CO  cd  s .s "3 CO  00  cu  a C  cd  Id cu  1 cd  C _o 'co  CO  CU u<  OH  cu  -a  C+H  CU o  CO  o  O  cd X  cu -o .c co co CU  cu  s  o 00  a  V  cu  td  a o  cu O  1 3  Ol  cd o  •a 00  CU  c2 cd  UH  cu  Id  <U  cu  1—I  CU  a o  o  •c J*  +->  cu  1a CU  a CU  £ CU Ui  cd  O o c  cu T3 CU H-» O  3 C  CU  T3  H->  o  X> co  O OH  T3 CU o  O  T3  o  CU o  o  co  C  o co  -4-»  3  CU  OH  B o  3  ON VO  cd cu  H  CO  CU  cu >  cu  Xi  o  cu UH  o  cfa o H-»  cu  CU  cu CO  >, C  "cd cd  36  Figure 1 Dendrogram for Clusters on Spouse, Self, and Conflict Rating at Baseline and Wave 1 Using Ward's Method 0  5  10  15  20  25  --+-  --+  satisfied sanctifiers  dissatisfied "sahcfifiers ~  nonsanctifiers  37  Footnotes *Four hundred and seventy four men were also interviewed at Baseline, but men were not included in this study because only thirtyfivebereaved men participated in one or more postbereavement interviews. Data from the third postbereavement interview (36-48 months after loss) were excluded from inferential analyses because (a) the number of bereaved respondents dropped from 169 to 91 and (b) control participants were not asked the relevant questions for this study at the third wave.  38  Bibliography Andrews, F. M., & Robinson, J. P. (1991). Measures of subjective well-being. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes: Vol. 1. Measures of social psychological attitudes, (pp. 61-114). New York: Harcourt Brace Jovanovich. Bartlett, F. C. (1967). Remembering: A study in experimental and social psychology. Cambridge: Cambridge University Press. (Original work published 1932) Bowlby-West, L. (1983). The impact of death on the family system. Journal of Family Therapy. 5. 279-294. Boyd, J. H., & Weissman, M. M. (1982). Screening for depression in a community sample. Archives of General Psychiatry. 39. 1195-1200. Bradburn, N. M. (1969). The structure of psychological well-being. Chicago: Aldine. Conway, M. (1990). On bias in autobiographical recall: Retrospective adjustments following disconfirmed expectations. Journal of Social Psychology. 130. 183-189. Costa, P. T. Jr., & McCrae, R. R. (1987). Environmental and dispositional influences on well-being: Longitudinal follow-up of an American national sample. British Journal of Psychology. 78. 299-306. Costa, P. T. Jr., & McCrae, R. R. (1988). Personality in adulthood: A six-year longitudinal study of self reports and spouse ratings on the NEO Personality Inventory. Journal of Personality and Social Psychology. 54. 853-863. Costa, P. T. Jr., & McCrae, R. R., (1992). Revised NEO Personality Inventory (NEO PIRTM) and NEO Five-Factor Inventory (NEO-FFI) Professional Manual. Odessa, FL: Psychological Assessment Resources. Costa, P. T. Jr., & Metter, E. J. (1994). Personality stability and its contribution to successful aging. Journal of Geriatric Psychiatry. 27. 41-59. Craig, T. J., & Van Natta, P. (1979). Influence of demographic characteristics on two measures of depressive symptoms. Archives of General Psychiatry. 36. 149-154. Derogatis, L. R. (1977). SCL-90: Administration, scoring, and procedures manual-I for the revised version. Baltimore, MD: Johns Hopkins University School of Medicine. Futterman, A., Gallagher, D., Thompson, L. W., Lovett, S., & Gilewski, M. (1990). Retrospective assessment of marital adjustment and depression during thefirst2 years of spousal bereavement. Psychology and Aging. 5. 277-283.  39  Harvey, J. H., Weber, A. L., & Orbuch, T. L. (1990). Interpersonal accounts: A social psychological perspective. Cambridge, MA: Basil Blackwell. Kohut, F., Berkman, L., Evans, D., & Cornoni-Huntley, J. (1983). Psychosocial assessment devices adapted for population surveys of the elderly: Two new forms of the CES-D. Paper presented at the Annual Meeting of the American Public Health Association, Dallas, TX. Lieberman, S. (1978). Nineteen cases of morbid grief. British Journal of Psychiatry. 132. 159-163. Lopata, H. Z. (1981). Widowhood and husband sanctification. Journal of Marriage and the Family. 43. 439-450. Lopata, H. Z. (1986). Time in anticipated future and events in memory. American Behavioral Scientist. 29. 695-709. McCauley, C , & Bremer, B. A. (1991). Subjective quality of life measures for evaluating medical intervention. Evaluation and the Health Professions. 14. 371-387. McFarland, C , & Ross, M. (1987). The relation between current impressions and memories of self and dating partners. Personality and Social Psychology Bulletin. 13. 228-238. McFarland, C , Ross, M., & Giltrow, M. (1992). Biased recollections in older adults: The role of implicit theories of aging. Journal of Personality and Social Psychology. 62. 837-850. Nahmani, N., Neeman, E., & Nir, C. (1991). Parallel process in bereavement support groups and supervision groups. The Clinical Supervisor. 9. 77-87. Parkes, C. M., & Weiss, R. S. (1983). Recovery from bereavement. New York: Basic Books. Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement. 1. 385-401. Rando, T. A. (1984). Grief, dying, and death: Clinical interventions for caregivers. Champaign, JL: Research Press. Rando, T. A. (1993). Treatment of complicated mourning. Champaign, JL: Research Press. Rubin, S. S. (1985). The resolution of bereavement: A clinical focus on the relationship to the deceased. Psychotherapy: Theory. Research. Training, and Practice. 22. 231-5.  40  Rubin, S. S. (1991). Adult child loss and the two-track model of bereavement. Omega: Journal of Death & Dying. 24. 183-202. Taylor, S. E. (1989). Positive Illusions: Creative self-deception and the healthy mind. New York: Basic Books. Warr, P., Barter, J., & Brownbridge, G. (1983). On the independence of positive and negative affect. Journal of Personality and Social Psychology. 44. 644-651. Worden, J.W. (1991). Grief counseling and grief therapy (2nd ed.). New York: Springer Publishing. Wortman, C. B., & Silver, R. C. (1989). The myths of coping with loss. Journal of Consulting and Clinical Psychology. 57. 349-357.  41  Appendix I: Items in Processing and Grief Scales Processing Scale (Alpha = .74) 1. At times, I want or need to think about my (husband/wife) or (his/her) death. Is this very true, somewhat, a little, or not true at all? 2. During the past month, how often did you actually talk about your (husband/wife) or (his/her) death with anyone? Would you say several times a day, daily or almost daily, 2 to 3 times a week, about once a week, less than once a week, or never? 3. During the past month, how often have you had thoughts or memories of your (husband/wife)? Would you say several times a day, daily or almost daily, 2 to 3 times a week, about once a week, less than once a week, or never? 4. During the past month, how often did you actually cry about the loss of your (husband/wife)? Would you say several times a day, daily or almost daily, 2 to 3 times a week, about once a week, less than once a week, or never? 5. Please tell me how true the following statements are as they apply to you. At times I find that I want or need to cry. Is this very true of you, somewhat, a little, or not true at all of you? 6. I feel the need to be with people who will let me talk about my (husband/wife) or (his/her) death. (Is this very true of you, somewhat, a little, or not true at all of you?) 7. During the past month, have you felt drawn to places and things that make you feel close to your (husband/wife)? 8. How much have you done things that help you feel close to your (husband/wife), such as going to the cemetery? Grief Scale (Alpha=88 for overall scale; alpha for subscales shown after each subscale title) anxiety subscale (.71)  despair subscale (.64) anger subscale (.59) shock subscale (.75) longing  1. During the past month, have you felt afraid of what lies ahead for you? (Would you say no, never; yes, but rarely; yes, sometimes; or yes, often?) 2. Have you felt extremely anxious and unsettled during the past month? 3. During the past month, have you felt worried about how you would manage your day-to-day affairs? 4. Has life seemed empty? 5. Have you felt empty inside, like an important part of you is missing? 6. In the past month, have you felt that life has lost its meaning? 7. During the past month, have you felt resentful or bitter about your (husband's/wife's) death? 8. Have you felt that the death of your (husband/wife) was unfair? 9. In the past month, have you felt anger toward God? 10. Have you felt as though you were in a state of shock? 11. During the past month, have you felt as though you couldn't believe what was happening? 12. Have you felt emotionally numb? 13. Have you found yourself longing to have (him/her) with you?  42  subscale (.75)  negative responses to reminders subscale (.62)  rumination/ intrusive thoughts subscale (.68)  14. People have different reactions to the death of a husband or wife. During the past month, have you had painful waves of missing your (husband/wife)? 15. During the past month, have you experienced feelings of intense pain or grief over the loss of your (husband/wife)? 16. During the past month, have you experienced feelings of grief, loneliness or missing your (husband/wife)? 17. During the past month, did thoughts or memories of your (husband/wife) make you feel sad or upset? 18. Almost everything I see or hear brings back painful reminders of my (husband/wife). 19. It is upsetting to come into contact with reminders of my (husband/wife)? 20. In general, during the past month, has talking with others about (him/her) made you feel sad or upset? 21. During the past month, have you had difficulty falling asleep because thoughts relating to (him/her) kept coming into your mind? 22. During the past month, did you find that you couldn't get thoughts of your (husband/wife) out of your mind even when you wanted to? 23. During the past month, have you found yourself going over and over what happened?  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items