| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Department of Metamedica, Philosophy and Medical Ethics Section, Vrije Universiteit, Amsterdam, The Netherlands
Berna van Baarsen, Department of Metamedica, Philosophy and Medical Ethics Section, Vrije Universiteit Amsterdam, Van der Boechorststraat 7, 1081 BT Amsterdam, The Netherlands E-mail: b.vbaarsen.metamedica{at}med.vu.nl.
Decision Editor: Fredric D. Wolinsky, PhD
 |
Abstract |
|---|
 TOP Abstract General and Specific Theories...Coping Resources and Adjustment...Research HypothesesMethodsResultsDiscussionReferences |
|---|
Methods. In total, 101 participants, aged 55–89 years, were interviewed before and after their partners' deaths.
Results. Findings were ambiguous with regard to both theories. As hypothesized, partner loss lowered self-esteem, resulting in higher emotional loneliness and social loneliness, that is, perception of less support. Supportive personal relations reduced emotional loneliness. The presence of close friends, however, seemed to increase emotional and social loneliness (i.e., decrease perceived support) in the long term, particularly among bereaved participants with lower self-esteem.
Discussion. The findings highlight the need to integrate theoretical concepts. In explaining adjustment to a partner's death, attention should be paid to underlying mechanisms relevant to the restoration process (e.g., identity change) and the ways in which the adjustment process can be improved (e.g., intimate relationships) or impeded (e.g., dependency-sustaining relationships).
IN recent decades, researchers have developed several theories to clarify the factors that explain the rate and patterns of adjustment to bereavement. Is coping with loss better explained by a general theory or a specific theory? Dykstra and De Jong Gierveld 1994
(p. 235) noted that "processes of reorganization are not unique to widowhood." Yet different types of events pose different demands on the individual's capacity to cope and seem to elicit different coping responses (Billings and Moos 1981; Folkman and Lazarus 1980).
In this longitudinal study we examined adjustment to widowhood among elderly participants in terms of loneliness. Data were gathered before and during the 2.5-year period following bereavement. The focus was on two main issues: the role of social support and that of self-esteem in adjustment to loss (Thoits 1995). Although the impact of these two coping resources has been examined extensively, the results are inconclusive. Moreover, theoretically, there is no agreement about the role of social support in adjustment to widowhood (Stroebe, Stroebe, Abakoumkin, and Schut 1996). In explaining how bereaved elderly people adjust to loneliness, we compared and evaluated assumptions based on a general theory of coping—the theory of mental incongruity (Munch 1972)—and a specific theory of loss and recovery—the theory of relational loneliness (Weiss 1973). These two theories have different conceptual approaches toward explaining how people cope with loss experiences. We did not assess the coping process in terms of how individuals respond cognitively or behaviorally to stressful demands; instead we aimed to clarify the role of resources as part of the adjustment process.
|
General and Specific Theories of Coping With Loss |
|---|
|
TOPAbstractGeneral and Specific Theories...Coping Resources and Adjustment...Research HypothesesMethodsResultsDiscussionReferences |
|---|
; Tazelaar and Wippler 1985; see also Dykstra 1995). According to the theory, the behavioral disposition of an individual is formed by interrelations between standards (i.e., desires) and cognitions (i.e., actual experiences), the so-called mental system. Incongruity refers to an imbalance in the system and is expected to be greater when conditions for change (e.g., income, support) are less favorable. By introducing conditional elements in the explanation of incongruity, the theory explains why individuals do not always behave in keeping with their standards or attitudes. The loneliness experienced by bereaved people is interpreted as an incongruity in relationships, a "discrepancy between the relationships one has and the relationships one wants" (Dykstra and De Jong Gierveld 1994, p. 245). Relationships with others may be a source of incongruity when expectations of support fall short of the actual support received. Bereaved individuals with more social support or higher self-esteem have more favorable conditions for change. For this reason, they are expected to feel less lonely after the loss of a partner than those with less support or lower self-esteem. In accordance with the theoretical framework, the relationships within a personal network are assumed to have about the same meaning; hence, loneliness after a partner's death can be assuaged by any relationship other than that with the partner.
The theory of relational loneliness is a specific theory of loss and recovery that focuses on the psychological and relational mechanisms underlying the adjustment process. According to Weiss 1973, feelings of loneliness may result from a deficit in one or more relational functions, such as attachment, social integration, and reassurance of worth. Two types of loneliness are identified: loneliness through social isolation and loneliness through emotional isolation. In the former, loneliness is caused by a lack of social integration. In the latter, loneliness is caused by the absence of a reliable attachment figure, such as a partner. Loss of an attachment figure leads to identity impairment (Weiss 1973). An essential process in the adjustment to loss is the development of a new self-concept. In addition, Weiss argued that the loss of an attachment figure can only be substituted by another close and intimate bond. Other supportive friendships cannot compensate for the loss.
The two theories agree and disagree, respectively, about the roles of self-esteem and social support in diminishing loneliness after a partner's death. The theory of mental incongruity offers a general conceptual framework in which self-esteem, as well as social support, is one of many possible opportunities for change. The theory does not distinguish between the differing roles of personal and social resources or opportunities to change an undesirable event. On the other hand, the theory of relational loneliness works on specific assumptions about the role of self-esteem and relates self-esteem to the process of identity change. With regard to social support, the theory of mental incongruity predicts that supportive relationships can compensate for the lost support from the partner, because support is seen as a general favorable opportunity for change. The theory does not distinguish between different relation functions. By contrast, the theory of relational loneliness interprets the role of social support on the basis of one relational function, that is, relationships of attachment, and rejects the idea that supportive others can protect the individual against the emotional impact of a partner's death.
|
Coping Resources and Adjustment to Loss |
|---|
|
TOPAbstractGeneral and Specific Theories...Coping Resources and Adjustment...Research HypothesesMethodsResultsDiscussionReferences |
|---|
; Stylianos and Vachon 1993). Research has stressed that social support may serve as a buffer against the adverse effects of partner loss, yet results remain inconclusive (Stroebe et al. 1996). Results from bereavement studies examining the theoretical relevance of the two theories are also equivocal. Much like Weiss 1973, Weiss 1993, Stroebe and colleagues 1996 found no buffering effect for social support, indicating that support cannot compensate for loss of support from a partner. Studies (Dykstra 1995; Dykstra and De Jong Gierveld 1994) in which the theory of mental incongruity was applied revealed that support from friendships offers protection against loneliness after a partner's death. Others have stressed the significance of intimate relationships with well-trusted individuals (e.g., Thomas, DiGiulio, and Sheehan 1988). Close and long-term contacts are important and give individuals a sense of stability in their past and anticipated future, especially among elderly persons (Carstensen, Graff, Levenson, and Gottman 1996). Because older people are confronted with limited time and a sense of mortality, they may confine their social involvement to fewer but more intimate relationships. This process of socioemotional selectivity (Carstensen, Isaacowitz, and Charles 1999) emphasizes how resilient elderly persons can be in regulating their emotions. Older people tend to enhance self-protective coping strategies, for instance, through "dampening" negative affect experiences (Lawton 1996), and as their sources of stress change (e.g., health problems), they may change their coping strategies to remain involved in life (Lazarus 1996).
In addition to social support, a second important coping resource is self-esteem (Thoits 1995). Many studies have examined the impact of self-esteem on the effect of bereavement (e.g., Lund, Caserta, and Dimond 1993). Self-esteem, an evaluative aspect of the self-concept (Greenwald, Bellezza, and Banaji 1988), has been defined as a personal attitude that individuals maintain regarding their own worth and importance (Rosenberg 1965). "With low self-esteem there is little motivation, confidence, and skill to change the circumstances" (Lund et al. 1993, p. 253). High self-esteem is associated with favorable bereavement outcomes in terms of loneliness (Dykstra 1995), and it functions as a buffer against the emotional consequences of stress (Thoits 1995).
Conceptions of self have received a great deal of attention in studies among elderly persons. Findings regarding stability and changes in self-concept (e.g., self-esteem) with aging are inconclusive. Many studies have found relative stability in evaluations of self-esteem, or even an increase with age (e.g., Dietz 1996). Others have found that self-esteem decreases with age because of the loss of roles associated with social identities (Mitsch Bush and Simmons 1981). Widowhood, for instance, may change conceptions of the self and can lead to identity impairment: Bereaved individuals face the task of becoming accustomed to being single instead of one of a couple (Lopata 1993).
Because supporting friends or relatives may "bolster self-esteem or a sense of identity" (Thoits 1995, p. 65) and, conversely, people with high self-esteem are probably more likely to feel in control when interacting with others and to have an adequate network for support (Dykstra 1995), interaction effects between personal (e.g., self-esteem) and social (e.g., network support) coping resources are of interest.
|
Research Hypotheses |
|---|
|
TOPAbstractGeneral and Specific Theories...Coping Resources and Adjustment...Research HypothesesMethodsResultsDiscussionReferences |
|---|
According to the theory of relational loneliness, a partner's death has a greater impact on emotional loneliness than it does on social loneliness. For this reason, we formulated hypotheses with regard to emotional loneliness only. We expected that a partner's death would lower self-esteem, which would impair adjustment to emotional loneliness. Social support would have no impact on emotional loneliness. The hypotheses derived from the theory of relational loneliness were as follows:
In keeping with the theory of mental incongruity, we expected that incongruity after a partner's death would be lower (i.e., weaker loneliness) when conditions were more favorable (higher self-esteem, more social support). We expected that the presence of more favorable conditions would increase the individual's chances of adapting to both types of loneliness. The hypotheses that follow were based on the theory of mental incongruity.
|
Methods |
|---|
|
TOPAbstractGeneral and Specific Theories...Coping Resources and Adjustment...Research HypothesesMethodsResultsDiscussionReferences |
|---|
). The respondents with partners (N = 2,606) formed the baseline group for the present study. They received a checklist every 6 months in which we inquired about major life events, such as the death of a partner. During this monitoring period, 239 respondents (9%) lost their partners, and thus became potential participants for our follow-up study. In total, 197 of those bereaved individuals were approached because 34 participants had been widowed longer than 7 months and 8 had died (first-step ineligibles).
Procedure.
Potential participants received an introductory letter informing them about WALS. Shortly after, they were contacted by an interviewer who asked for their cooperation. Interviews were held by trained lay interviewers at the respondents' homes.
Response.
Of the 197 bereaved individuals, 111 completed the entire sequence of interviews (T0–T5). During the course of the study, 19 widow(er)s died and 26 became unable to participate because of mental or physical illness (second-step ineligibles). In addition, 41 widow(er)s refused to take part in the study. The response rate, corrected for first- and second-step ineligibles, was 62%.
Selectivity.
Respondents who took part in the entire study (N = 111) were significantly younger at T0 (Ma = 71.3) and physically more healthy (Mh = 17.7, instrumental activities of daily living [IADL] measure, range = 4–20) than those who died (Ma = 79.3, Mh = 13.5) and those who were too ill to participate (Ma = 79.9, Mh = 14.5). A significantly higher number of men died during the course of the study (63% of the deceased were men) as compared to the number of men who completed their participation (39% of the participants were men). Much like the deceased and ailing individuals, the respondents who refused to cooperate were older (Ma = 76.8) than the participating widow(er)s. We found no significant differences in education, income, or loneliness. The selectivity of the sample indicates a need for caution in generalizing the findings to the population of bereaved elderly individuals.
Sample.
At the time of their partners' deaths, the respondents who fully participated (N = 111) were approximately 73 years old (SD = 8.4). The women (n = 68) were widowed at an earlier age than the men (n = 43), Mwomen = 70.6, Mmen = 76.9; t(109) = 4.14, p < .001. The average length of widowhood at T5 was 2.8 years (SD = 0.2).
Measures
Emotional and social loneliness.
Our measures of emotional and social loneliness were based on a Dutch loneliness scale by De Jong Gierveld and Kamphuis 1985. Table 1 presents the items that describe the two types of loneliness. Mokken scale analysis carried out on the NESTOR data yielded two subscales or factors: Emotional Loneliness (Loevinger's H = 0.48, reliability 
= 0.84) and Social Loneliness (H = 0.43, = 0.77; Van Baarsen, Smit, Snijders, and Knipscheer 1999). Because the distinction between emotional and social loneliness might have been an artifact due to the division between the negatively and positively formulated items, additional analyses were carried out. Item response theory analysis (Andersen test, Martin-Löf test) and an external validity study based on the association of the items with theoretically relevant background, personality, and social variables showed that the two subscales represented two different concepts (Van Baarsen, Snijders, Smit, and Van Duijn 2001). Emotional loneliness refers to the feeling of lacking reliable attachments to others. Social loneliness describes loneliness experienced because of lack of social integration. The resemblance of the social loneliness items with items from other studies (e.g., Stroebe et al. 1996) suggests that the social loneliness scale can be viewed as a measure of perceived social support (Van Baarsen 2001). The correlation between the emotional loneliness scale and the social loneliness/perceived support scale, corrected for attenuation, was .41.
|
, is comparable to the Rosenberg Self-Esteem Scale (1965). This scale has been used in various research projects in the Netherlands (e.g., Dykstra 1995). People with higher self-esteem (i.e., higher scores) are supposed to have a more positive view of their identity (Greenwald et al. 1988). Cronbach's alpha across measurements varied between .79 and .82.
Social support.
Two measures were used. Network support was based on two items. One of these was instrumental support: "How many times over the last (half) year has ... helped you with daily tasks (in and around the house, such as preparing meals, cleaning, transportation, and minor repairs)?" The other was emotional support: "How many times over the last (half) year has ... given you attention and affection?" This question applied to the 11 most important members of the participants' networks, other than the partner, with whom their contact frequency was at least once a month. We identified the core network by means of the domain-specific approach, using seven domains of relationships such as children, neighbors, and work-related contacts (Van Tilburg 1995). We assessed the presence of a confidant by asking respondents whether they had a best (fe)male friend (1 = "no," 2 = "yes").
Background variables.
General health (T0) was measured with one item: "How is your health in general?" Answers ranged from 1 ("poor") to 5 ("very good"). High correlation with measures of physical disability (activities of daily living: r = -.58; IADL: r = -.41) indicated that the one-item scale was a valid measure of general health. Lost support from a partner was based on instrumental support (one item) and emotional support (one item) received during the marriage (T0; same procedure as network support). We measured support standards by asking respondents whether they believed that a widow(er) needs someone to "give them attention and affection," who "helps them with daily chores in and around the house," and who "calls up occasionally just to chat." Contact standards measured the respondents' views concerning the importance of having contacts with family (e.g., "A widow(er) should keep in touch with her/his children") and people outside the family (e.g., "Neighbors are extremely important when you have just lost your partner"). Items were based on examples from Stevens 1989 and Dykstra 1990. Cronbach's alpha for support standards (three items) was .52. The alphas for contact standards regarding family (three items) and nonfamily (four items) were .52 and .62, respectively.
Methods of Analysis
We tested hypotheses by means of multiple regression analysis. To increase the clarity of the findings, we used three postloss measurements: 6 months (T1), 1.5 years (T3), and 2.5 years (T5) after the partner's death. In the first step, we entered background variables as covariates: gender, age at widowhood, general health, and time interval between T0 and T1. Also, we entered lost support from the partner and support and contact standards to control for (changed) cognitions and expectations of contacts and support following the partner's death. We monitored the effect of the partner's death on loneliness by entering preloss levels of emotional loneliness or social loneliness/perceived support. To test Hypotheses 2–5, we included self-esteem and the social support variables in the analyses. Changes in self-esteem or network support after the partner's death were considered to have an impact when postloss measures (T1) influenced loneliness once we checked preloss measures (T0). In the second step, we entered interaction terms with gender (control variables) and two-way interaction terms of self-esteem with network support or a best (fe)male friend (Hypothesis 5). We computed interaction terms by multiplying the centered values of the variables. Because some interaction terms were strongly correlated (r > .50, p < .001), nonsignificant interaction terms were omitted (one by one, p > .05). Table 2 and Table 3 present the results of the multiple regression analyses on emotional loneliness and social loneliness/perceived support, respectively.
|
|
|
Results |
|---|
|
TOPAbstractGeneral and Specific Theories...Coping Resources and Adjustment...Research HypothesesMethodsResultsDiscussionReferences |
|---|
|
Explaining Differences in Emotional Loneliness and Social Loneliness/Perceived Support Among Bereaved Persons
Control variables.
Previous levels of loneliness were highly predictive of the loneliness experienced after the partner's death. Some differences between the two types of loneliness emerged. Whereas predictive values of preloss emotional loneliness decreased over time (b decreased from .438 to .341), the impact of initial social loneliness/perceived support increased (b increased from .280 to .407). Thus, the degree of social loneliness/perceived support before partner loss became increasingly important to postloss social loneliness/perceived support as time elapsed. The impact of support standards and contact standards differed in terms of direction and strength. First, support standards were both significant and consistent in their prediction of emotional loneliness, whereas social loneliness/perceived support was affected by contact standards only. Second, higher contact standards regarding nonfamily seemed to result in higher emotional loneliness but less social loneliness, that is, perception of more support. Poor general health increased the likelihood of emotional loneliness and social loneliness (i.e., perception of less support) among respondents after partner loss. This effect seemed to be constant for emotional loneliness. For social loneliness/perceived support, however, this effect became increasingly important over time. We witnessed an intriguing effect of lost support from the partner. The bereaved elderly individuals who had lost more partner support, compared with those who had lost less, experienced significantly less social loneliness, that is, perceived more support during the 2.5 years following the partner's death. Two main gender effects were found, but these effects became significant only after we entered gender interaction effects.
Main effects of self-esteem, network support, and presence of a confidant.
In keeping with the theory of mental incongruity (Hypothesis 4), higher preloss self-esteem was indicative of lower levels of emotional loneliness. However, the effect was mediated by postloss levels of self-esteem (not in Table 2 ) and was only found at T5. Another finding was consistent with the theory of relational loneliness (Hypothesis 2): A decrease in self-esteem after the death was related to an increase in emotional loneliness. A decrease in self-esteem was also predictive of higher levels of social loneliness, that is, perception of less support at T3 and T5.
Preloss network support (T0) had no effect whatsoever. The failure to predict emotional loneliness from initial network support was consistent with the theory of relational loneliness (Hypothesis 3). However, increases in network support shortly after partner loss alleviated feelings of emotional loneliness—but only at T5. This is incongruent with the theory of relational loneliness but falls in line with the theory of mental incongruity (Hypothesis 5). Changes in network support did not affect social loneliness/perceived support. Moreover, we found opposite effects for the presence of a confidant. In agreement with the theory of mental incongruity (Hypothesis 5), the presence of a best female friend seemed to serve as a predictor of less intense feelings of emotional loneliness. This effect, however, was established only at T1 and became nonsignificant after we entered gender interaction effects. The presence of a best male friend, by contrast, appeared to fall in line with the theory of relational loneliness (Hypothesis 3): It seemed to have no alleviating effect on emotional loneliness and even to be predictive of greater social loneliness, that is, perception of less support later in the bereavement process (T5).
Gender interaction effects.
We observed six gender interaction effects. Our interpretations were based on separate regression analyses. Age at widowhood increased emotional loneliness at T1 more for widowers (beta = .436, p < .05) than for widows (not significant = ns). This suggests that the widowers were more inclined with increasing age to experience emotional loneliness at T1 than widows. Higher support standards increased emotional loneliness at T1 more for widows (beta = .393, p < .01) than for widowers (ns). However, higher contact standards regarding family reduced emotional loneliness and social loneliness/perceived support at T3 more for widowers (betae = -.393, p < .05; betas = -.341, p < .10) than for widows (ns). Finally, higher self-esteem reduced emotional loneliness at T3 and T5 more for widowers (betaT3 = -.534, p < .10; betaT5 = -.702, p < .05) than for widows (ns).
Interaction effects between self-esteem and network support or the presence of a confidant.
The results partly supported the prediction that was based on the theory of mental incongruity (Hypothesis 6). As expected, a (fe)male confidant reduced loneliness more for those who had higher self-esteem. Nonetheless, separate regression analyses showed widow(er)s with lower self-esteem who had confidants to be the most vulnerable: Lower self-esteem increased emotional loneliness at T5 more for respondents with a best female friend (beta = -.131, ns) than those without one (beta = .119, ns). Lower self-esteem increased social loneliness/perceived support at T3 more for respondents with a best male friend (beta = -.569, p < .01) than those without one (beta = -.046, ns). The results did not support the hypothesis regarding network support: This support reduced social loneliness/perceived support at T5 more for widow(er)s with lower self-esteem than for those with higher self-esteem.
|
Discussion |
|---|
|
TOPAbstractGeneral and Specific Theories...Coping Resources and Adjustment...Research HypothesesMethodsResultsDiscussionReferences |
|---|
Results regarding the role of self-esteem support the theory of relational loneliness in that partner loss can lead to a loss of identity (i.e., lower self-esteem). Moreover, lowered self-esteem increases emotional loneliness over time. We also find support for the prediction based on the theory of mental incongruity that incongruity is weaker (i.e., less loneliness) with favorable conditions for change (i.e., higher self-esteem). The results apply to emotional loneliness as well as social loneliness/perceived support.
The effects of preloss esteem and changes in self-esteem show certain interesting trends over time. Whereas lowered self-esteem has an emotional impact throughout the bereavement process, the emotional effect of preloss self-esteem (which is mediated by postloss self-esteem) is only found at 2.5 years after the death. Consistent with earlier research (Meuser, Davies, and Marwit 1994), preloss self-esteem may become salient only after the initial shock is dissipated. The somewhat delayed effect of lowered self-esteem on social loneliness/perceived support suggests that the loss of self-esteem may influence feelings of relational competence and personal control (Pearlin, Lieberman, Menaghan, and Mullan 1981). If lowered self-esteem discourages widow(er)s from integrating socially and from finding new relationships, feelings of emotional loneliness and social loneliness/perceived support may become more intense.
We find partial support for the two theoretical hypotheses regarding the role of social support in adjustment to partner loss. Two findings support the theory of relational loneliness: (a) preloss support does not protect bereaved elderly people from emotional loneliness, and (b) increases in network support shortly after the partner's death do not help the individual to recover from emotional loneliness in the first 1.5 years of bereavement. However, the long-term mitigating effect of increased network support after a partner's death supports the theory of mental incongruity in that more favorable conditions reduce incongruity.
Although the supportive role of a confidant was not directly examined in this study, the findings suggest that emotional support from female friends might be helpful in overcoming intense emotional loneliness shortly after partner loss. Research has highlighted the value of strong emotional support and empathy during the first period of intense grief (Bankoff 1983). Later in the transition phase, the individual may experience a greater need for instrumental support from less close relationships. This might explain why network support shows a delayed effect. The relevance of distinguishing between who gives what kind of support (Bankoff 1983) and the important role of specific, qualitative support in adequate functioning while facing stress (Billings and Moos 1981) may explain why Stroebe and colleagues 1996 found no mitigating effect for social support. In keeping with the theory of mental incongruity, we find that other supportive personal contacts can partially fill the gap left by a lost partner. Because a partner's death involves more than the loss of intimacy (Lopata 1993), close ties may not only provide a substitute for intimacy, but may also replace the partner in other areas, such as companionship or instrumental support.
The results of this study stress that intimacy is not more relevant to the problems of widowhood than identity, as was suggested by Thomas and colleagues 1988. On the contrary, support resources and self-esteem may supplement each other. Bereaved elderly individuals with low self-esteem seem to be more vulnerable not only when they receive little support from their network after partner loss, but also when they have a confidant. The negative effect of having a confidant among low-self-esteem respondents on the two types of loneliness conflicts with the theory of mental incongruity. This effect tends to support the theory of relational loneliness, because this theory claims that close friends do not, by definition, alleviate a widow(er)'s sense of being alone.
The negative impact of confidants in the long term points to the issue of dependency. Widow(er)s with low self-esteem, particularly those who felt strongly dependent on their partners, may be more inclined than those with high self-esteem to cling to an intimate friend as a replacement. Close ties bolster feelings of security, and past memories can be shared with intimate friends. Although widow(er)s need to be able to be dependent in periods of intense grief (Bankoff 1983), dependency-sustaining relationships may impede the establishment of a new autonomous lifestyle. Although older people may, in general, benefit from increased involvement with close family and friends (Carstensen et al. 1999), it would be worthwhile for researchers to investigate how this form of emotional control functions later in the bereavement process.
The results are not straightforward in terms of clarifying the explanatory accuracy of the two theories. The theory of mental incongruity provides useful concepts, such as mental incongruity, standards, and conditions for change, whereas the theory of relational loneliness is necessary in interpreting the results. First, emotional loneliness might be the most appropriate concept in assessing mental incongruity due to partner loss. Second, conditions for initiating and maintaining social contacts are not favorable under all circumstances. In studying the relative contribution of the various conditions to bereavement outcomes, interaction effects should be taken into account.
Unlike the theory of mental incongruity, attachment models, such as the theory of relational loneliness, offer more insight into the underlying mechanisms relevant to the restoration process (e.g., identity change) and the ways in which the adjustment process can be improved (e.g., new relationships) or impeded (e.g., dependency; Weiss 1993). However, the findings also emphasize that relationships other than the partner can be important coping resources. Presumably, there are other relevant intra- and interpersonal risk factors in bereavement outcomes, in addition to those inherent to the lost attachment bond (Stroebe and Stroebe 1993).
The findings of this study are related to a situation-specific event, that is, a partner's death. We made no distinction between unexpected and anticipated deaths. In addition, we did not check the possible effects of the daily recurring stress that older people may experience before and during bereavement. What is more, the small group of respondents limited the statistical power of the tests, and attrition indicated that the present study is not representative of the older widowed as a whole, particularly of the groups of frail old bereaved persons (i.e., those who are older and less healthy). Also, the final models that were estimated are not fully comparable, and the relatively short follow-up period restricted our research in that we could only study the first 2.5 years of widowhood. Finally, considering the low Chronbach's alphas, the support standards items and the contact standards items seem to be insufficient measures to cover the constructs of support standards and contact standards, respectively. The present results show that future development of such scales would be worthwhile.
Despite these methodological shortcomings, the present study offers important advantages. The longitudinal character of the study made it possible for us to examine how self-esteem and social support function as risk factors for the loneliness experienced after a partner's death. Moreover, the distinction between emotional loneliness and social loneliness/perceived support yields useful information. When researchers focus on loneliness as a general experience, it is difficult for them to determine the role of social support in alleviating the experienced emotional loneliness and social loneliness/perceived support during bereavement. Finally, our comparison of a general and a specific coping theory reveals that knowledge of the specific processes underlying grief and adaptation to loss is necessary to the proper practical application of theoretical concepts. Moreover, researchers can use other domain-specific (age) theories to make specific predictions about the impact of social relations and time perspectives in adjustment processes in later life. The use of theories in research is important (Bengtson, Rice, and Johnson 1999) because theories offer investigators the opportunity to develop insights into various groups of people, to explain counterintuitive findings, and to outline methods for intervention. The distinction between the two types of loneliness, for instance, may increase the effectiveness of interventions. The predictive value of prebereavement emotional loneliness and social loneliness/perceived support suggests that prevention programs should focus on older persons who experience high levels of loneliness during marriage. Knowledge of risk factors of different types of loneliness may be useful in planning and developing intervention strategies and is important in defining which groups of (older) bereaved persons are most vulnerable. This study sheds new light on the roles of self-esteem and social support in the loneliness experienced by bereaved elderly people. Future research should test the tenability of the interpretations provided here.
|
Acknowledgments |
|---|
We thank Margaret Stroebe, Henk Schut, Jan Smit, and Dorly Deeg for their helpful comments on earlier versions of this article.
Received for publication July 28, 2000. Accepted for publication June 12, 2001.
|
References |
|---|
|
TOPAbstractGeneral and Specific Theories...Coping Resources and Adjustment...Research HypothesesMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. S. Rook Gaps in social support resources in later life: An adaptational challenge in need of further research Journal of Social and Personal Relationships, February 1, 2009; 26(1): 103 - 112. [Abstract] [PDF]
|
|
|
|
 |
|
 R. S. Wilson, K. R. Krueger, S. E. Arnold, J. A. Schneider, J. F. Kelly, L. L. Barnes, Y. Tang, and D. A. Bennett Loneliness and Risk of Alzheimer Disease Arch Gen Psychiatry, February 1, 2007; 64(2): 234 - 240. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. K. Barg, R. Huss-Ashmore, M. N. Wittink, G. F. Murray, H. R. Bogner, and J. J. Gallo A mixed-methods approach to understanding loneliness and depression in older adults. J. Gerontol. B. Psychol. Sci. Soc. Sci., November 1, 2006; 61(6): S329 - S339. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
