| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Neal Krause, Department of Health Behavior and Health Education, School of Public Health, The University of Michigan, 1420 Washington Heights, Ann Arbor, MI 48109-2029. E-mail: nkrause{at}umich.edu
 |
Abstract |
|---|
 TOP Abstract Design and MethodsResultsDiscussionReferences |
|---|
Key Words: Traumatic life events • Cohort differences
Even a moment's reflection reveals that it would be virtually impossible to inventory all the stressful life events that older people have encountered over the course of a lifetime. Fortunately, one way to overcome this problem involves turning to a different type of stressor that is especially well suited for the task—lifetime trauma. Wheaton (1994) defines trauma as events that are "... spectacular, horrifying, and just deeply disturbing experiences" (p. 90). Traumatic events are differentiated from stressful life events by their imputed seriousness. Included among traumatic events are sexual and physical abuse, witnessing a violent crime, the premature loss of a parent, and participation in combat. Several researchers have examined the relationship between lifetime trauma and outcomes such as physical health status (Krause, 1998), but there do not appear to be any studies that focus specifically on the relationship between lifetime trauma and life satisfaction. The first goal of this study is to fill this gap in the knowledge base by assessing the relationship between lifetime trauma and life satisfaction with data provided by a recent nationwide survey of older people.
But looking at the direct effects of trauma on life satisfaction does not go far enough. Researchers have argued for some time that older people turn to social as well as psychological resources in an effort to confront the difficult events that arise in life. One of the most widely studied resources is social support (Krause, 2001). This literature indicates that older adults who are embedded in strong social support networks tend to cope more effectively with the deleterious effects of stress than older people who do not maintain close ties with others. Consistent with this view, the second goal of the present study is to examine the relationships among lifetime trauma, social support, and life satisfaction. Although a number of investigators have examined whether social support buffers the noxious effects of stressful life events, there do not appear to be any studies in the literature that assess whether support from significant others also offsets the deleterious effects of lifetime trauma.
In the process of addressing these issues, an effort is made to expand the scope of inquiry by turning to a final limitation in prior work. By far, most researchers examine the relationship between stress and well-being by pooling all older respondents into a single study sample—typically all individuals age 65 and older. Whether these investigators are aware of it or not, this strategy assumes that all older people respond to stress in the same way. However, this may not be true. The classic work of Elder (1999) suggests that reactions to trauma (in his case the Great Depression) depend, in part, on the age cohort in which an older person is embedded. Although this research adds considerably to our understanding, it deals solely with a single traumatic event. By ignoring cumulative exposure to trauma over the entire life course, this work runs the risk of underestimating the impact of the single event that is evaluated (see Turner & Lloyd, 1995, for an empirical evaluation of this problem). In order to overcome this problem, researchers should work with checklists that capture multiple traumatic events so that the stress context may be more adequately depicted. Based on this rationale, the final goal of the present study is to examine the effects of multiple traumatic events in three age cohorts of older people: the young–old (65–74 years old), the old–old (75–84 years old) and the oldest–old (85 years of age and older). There do not appear to be any studies in the literature that probe for cohort differences in the relationship between multiple traumatic events and life satisfaction with data provided by older adults.
|
Design and Methods |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
The sampling frame consisted of all eligible individuals contained in the Health Care Financing Administration (HCFA) Medicare Beneficiary Eligibility List (HCFA is now called the Centers for Medicare and Medicaid Services—CMS). Three waves of data were collected between 1991 and 1999. All data collection was conducted by Harris Interactive in New York City. A total of 1,103 interviews were completed successfully at the baseline interview in 1992–1993. Of these study participants, 605 were reinterviewed in 1996–1997. A third wave of data was collected in 1998–1999. A total of 530 older people who participated in the earlier waves of data collection were successfully reinterviewed at Wave 3.
A fourth wave of interviews was conducted in 2002–2003. However, the sampling strategy was complex. Two groups of older people were interviewed at this point. The first consisted of all older adults who participated in Waves 1–3. A total of 269 of these individuals were reinterviewed at Wave 4. This group was supplemented with a sample of older people who had not been involved in the project previously. The CMS files were once again used as a sampling frame for identifying elders in the supplementary sample. However, in this case, the sample was selected so that when it was combined with those who had participated in the study previously, there would be approximately equal numbers of people in each of the following age cohorts: 65–57 (n = 491); 75–84 (n = 515); and 85 and older (n = 509). Altogether, the Wave 4 sample consisted of 1,518 older people. The overall response rate for both groups in the Wave 4 sample was 54%. The response rate is somewhat lower than what might be found in the typical survey of older people because a larger amount of nonresponse was encountered among people age 85 and older (see Rodgers & Herzog, 1992, for a discussion of this persistent problem).
Because data on lifetime trauma were not obtained until the Wave 4 survey, all analyses in this study are based on this round of interviews only. As discussed below, several sets of analyses are performed in this study. After using listwise deletion of cases containing item nonresponse, the sample sizes in the analyses ranged from 1,397 older people to 457 older study participants. Based on the sample of 1,397 respondents, preliminary analyses revealed that the average age was 74.7 years (SD = 7.4 years), approximately 42% were men, 58% were married at the time the interview took place, and 89% were White. The descriptive analyses further revealed that the average number of years of completed schooling was 12.0 years (SD = 3.3 years). Based on the Census Bureau's most recent Current Population Survey (CPS) estimates, these descriptive data were weighted by age, gender, education, and race. All analyses that follow are based on weighted data as well.
Measures
The measures used in this study are listed in Table 1. The procedures used to code the survey indicators are provided in the footnotes of this table.
|
It is important to briefly examine why life satisfaction is an especially good outcome measure to use in this study. As discussed earlier, life satisfaction arises from an assessment of the fit between desired goals in life and actual life outcomes. Although these evaluations are made throughout the life course, they may become especially important in late life. Evidence of this may be found in the work of Erikson's (1959) widely cited theory of aging. He argues that the life span is divided into eight stages and that each stage poses a unique developmental challenge. The final stage, which arises in late life, is characterized by the crisis of integrity versus despair. This is a time of deep introspection where an individual begins to accept the kind of person he or she has become over the years. This is accomplished by reconciling what one set out to do in life with what has actually been accomplished. If this crisis is resolved successfully, elderly people are thought to derive a deep sense of meaning in life. But if it is not resolved successfully, they are likely to slip into despair. Traumatic events are likely to interfere with the process of developing a sense of integrity because it is often very difficult to make sense of them and to see their larger purpose in life. Moreover, the strong negative emotions created by traumatic events may make it difficult to think objectively about them, and as a result, it may be hard for older people to reconcile what has happened to them and to be at peace with the past.
Lifetime trauma
The older people in this study were presented with a list of 22 traumatic life events. This checklist was assembled from several sources, including the work of Wheaton, Roszell, and Hall (1997), Turner and Lloyd (1995), and the traumatic events listed in the Diagnostic and Statistical Manual for Mental Disorders (DSM–IV; American Psychiatric Association, 1994, p. 424). The study participants were asked whether they ever experienced any of these events in their lifetime. If they indicated they had been exposed to an event, they were asked to report their age when they first encountered the stressor. A simple unweighted sum of the total number of traumatic events was computed from these data. Consistent with the work of Norris (1992), preliminary analysis revealed that exposure to trauma was not uncommon and that the older people in the present study experienced an average of 2.7 traumatic events (SD = 2.1) over the course of their lives.
Social support
Four indicators are used in this study to assess how often family members and close friends provided emotional support to older study participants in the year prior to the interview. These items, which were taken from the work of Krause (1995), are coded so that a high score represents more emotional support. The reliability of this brief composite measure is.858.
It is important to discuss why current levels of emotional support might offset the effects of trauma that arose earlier in the life course. Initially, it might appear as though support that was available at the time the event took place may be more important than current support. Although support at the time the trauma was encountered is likely to be important, there are two reasons why current support also is useful. First, the outcome measure in this study captures current levels of life satisfaction. As a result, current support and feedback from social network members is likely to come into play as older adults engage in the process of introspection that determines current levels of life satisfaction. Second, it would not be feasible to measure emotional support at the time each traumatic event took place because the interview would quickly become cumbersome and repetitive for those who were exposed to multiple traumatic events. This is important because, as noted above, the older people in this study experienced an average of nearly 3 (i.e., 2.7) traumatic events in life.
But beyond these issues, current emotional support may play a vitally important role in the late-life developmental challenges discussed by Erikson (1959). In order to attain a sense of integrity, older people must step back and view their own past behavior more objectively, and they must find a way to manage troubling emotions, such as shame, anger, and anxiety. As Staudinger (2001) points out, this process is greatly enhanced if the life-review process is conducted with a confidant. These trusted others may help an elder process troubling emotions and point out blind spots in their self-perceptions as well as self-serving biases in the autobiographical reconstruction that takes place during this developmental stage. These helping behaviors are the hallmark of emotional support.
Demographic control measures
The relationships among lifetime trauma, emotional support, and life satisfaction were evaluated after the effects of age, gender, marital status, education, and race were controlled statistically. Age is scored continuously in years, and education reflects the total number of years of completed schooling. In contrast, gender (1 = men; 0 = women), marital status (1 = presently married; 0 = otherwise), and race (1 = White; 0 = all other racial and ethnic groups) are coded in a binary format.
Data Analysis Strategy
The analyses for this study are performed in three main segments that move from more general issues to progressively more refined specifications. First, the data provided by all three age cohorts are pooled, and a set of analyses are performed to see if emotional support offsets the noxious effects of lifetime trauma on feelings of life satisfaction for the sample as a whole. Following this, the relationships among lifetime trauma, emotional support, and life satisfaction are examined within each age cohort taken separately. Finally, a set of supplementary analyses are performed that further clarify and refine the findings that have emerged from the data.
Ordinary least squares multiple regression analyses are used to assess the relationships among lifetime trauma, emotional support, and life satisfaction. In essence, tests for the stress-buffering effects of emotional support call for the estimation of a statistical interaction effect between lifetime trauma and emotional support on life satisfaction. These tests are performed in two hierarchical steps. First, the additive effects of trauma, emotional support, and the demographic control variables are added in Step 1. Then a multiplicative term is added to the model in Step 2 to evaluate the joint effects of emotional support and trauma on life satisfaction. If the coefficient associated with the multiplicative term is statistically significant, then additional hand calculations are needed to make sure the interaction effect is in the proposed direction. If the hypothesized stress-buffering effects of emotional support are present, the impact of trauma on life satisfaction should become progressively weaker at increasingly higher levels of emotional support. This will be illustrated in the analyses that follow by using the formulas provided by Aiken and West (1991) to assess the effects of lifetime trauma at select levels of emotional support.
|
Results |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
|
Model 2 reveals that a statistically significant interaction effect is present between lifetime trauma and emotional support on feelings of life satisfaction (b =.039; p <.001; unstandardized estimates are discussed when reviewing interaction effects because standardized effects are meaningless in this context). As discussed above, formulas provided by Aiken and West (1991) are used to clarify the nature of the relationship among these constructs. These hand calculations involve computing the relationship between trauma and life satisfaction at select levels of emotional support. Although any emotional support scores can be used for this purpose, the following values were selected because they reflect the full range of emotional support scores: the lowest emotional support score, –1 standard deviation below the mean, the mean of emotional support, 1 standard deviation above the mean, and the highest observed emotional support value. The findings provided by the hand calculations (not shown in Table 2) indicate that the noxious effects of lifetime trauma become progressively weaker as we move from the lowest to the highest emotional support scores. More specifically, the data indicate that trauma has a substantial deleterious effect on feelings of life satisfaction for older people with the lowest observed emotional support score (ß = –.420; b = –.537; p <.001). This is important because 71 study participants had the lowest observed emotional support value. The data further reveal that at 1 standard deviation below the mean emotional support score, traumatic events are still significantly associated with life satisfaction, but the size of the relationship is beginning to taper off (ß = –.312; b = –.399; p <.001). At mean levels of emotional support, a significant relationship between lifetime trauma and life satisfaction is still present, but the effect size is about half as large as the estimate observed for those with the lowest support scores (ß = –.202; b = –.258; p <.001). The impact of lifetime trauma is reduced even further for older adults with emotional support scores at 1 standard deviation above the mean (ß = –.092; b = –.117; p <.01). Finally, the hand calculations reveal that the relationship between trauma and life satisfaction is not statistically significant for older people with the highest observed emotional support values (ß = –.054; b = –.068; NS). This is noteworthy because a total of 254 older adults had the highest observed emotional support scores. Looking across the full range of estimates, the data suggest that emotional support tends to reduce the relationship between lifetime trauma and life satisfaction by about 87%. This is a powerful illustration of the beneficial influence of informal social support in late life.
Assessing the Influence of Age Cohort
Findings from the analyses that were designed to evaluate the relationships among trauma, emotional support, and life satisfaction within the young–old, old–old, and oldest–old cohorts are presented in Table 3. In order to make the results easier to grasp, coefficients are reported only for the core measures of trauma and emotional support. However, it should be emphasized that these estimates were obtained after the effects of age, gender, education, marital status, and race had been controlled statistically.
|
The data in the middle panel of Table 3 come from the analysis of the relationships among trauma, emotional support, and life satisfaction among members of the old–old cohort (ages 75–84). Once again, the findings suggest that more lifetime trauma is associated with less life satisfaction (ß = –.168; p <.001). In addition, more emotional support also appears to be associated with greater life satisfaction (ß =.130; p <.01). Finally, the data reveal that emotional support tends to reduce the pernicious effects of lifetime trauma on life satisfaction (b =.032; p <.05), but the additional hand calculations (not shown in Table 3) indicate that the stress-buffering effects of emotional support are fairly modest.
The final set of cohort analyses probe the relationships among trauma, emotional support, and life satisfaction among the oldest–old (age 85 and older). Model 1 indicates that more lifetime trauma is associated with less life satisfaction (ß = –.140; p <.01). In contrast, the findings reveal that older people with more emotional support from family and friends do not appear to be more satisfied with life (ß =.081; NS). However, unlike the findings from the other age cohort groups, the results in Table 3 suggest that emotional support plays a major role in helping members of the oldest–old cohort deal with the traumatic events that have arisen over the course of their lives. More specifically, the statistical interaction effect between trauma and emotional support on life satisfaction is significant (b =.075; p <.001). The hand calculations discussed earlier were performed to illustrate the relationship among these constructs more clearly. Estimates of the relationship between lifetime trauma and life satisfaction were observed at the following emotional support values: the lowest emotional support score (ß = –.521; b = –.752; p <.001), –1 standard deviation below the mean (ß = –.337; b = –.486; p <.001), emotional support mean (ß = –.150; b = –.216; p <.01), +1 standard deviation above the mean (ß =.037; b =.054; NS), and the highest emotional support score (ß =.102; b =.148; NS). The impact of trauma on life satisfaction for the oldest–old cohort members with the lowest support value (ß = –.521) is unusually large by social and behavioral science standards.
Supplementary Analyses
The findings that have emerged so far suggest that exposure to trauma tends to erode feelings of life satisfaction. Moreover, the data indicate that these noxious effects are offset by emotional support, and this stress-buffering effect is especially evident in the oldest–old cohort. Four sets of additional analyses were performed to rule out competing explanations for these results and to further clarify the nature of the findings.
The first set of additional analyses probe the relationship between emotional support and lifetime trauma. A number of researchers claim that exposure to trauma tends to inhibit a person's ability to form meaningful relationships with others (e.g., Brown & Harris, 1978). If this is true, and the correlation between trauma and emotional support is large, then there is a risk that the statistical interaction effect between trauma and emotional support on life satisfaction may be confounded with the relationship between trauma and emotional support. Subsequent analysis revealed that the correlation between lifetime trauma and emotional support (.092; p <.01) is statistically significant. However, two points should be made about this relationship. First, the sign of this relationship is positive, suggesting that elders who experience more traumatic events tend to get more emotional support from significant others, not less, as some suggest. Second, and more important, the size of the relationship is quite modest. Consequently, it is very unlikely that statistical confounding influenced the findings in this study.
Life-course themes figure prominently in the rationale that was developed for this study. But the measure of lifetime trauma that was used combines all events arising at all points in the life course into a single summary score. This makes it impossible to distinguish between the effects of traumatic events that arose decades ago and the effects of trauma encountered only recently. If the results reported in the previous sections can be attributed solely to recent traumatic events, then the life course perspective is of relatively little value for research in this field. In order to rule out this possibility, a set of additional analyses were conducted to examine the interaction among emotional support, life satisfaction, and trauma encountered at specific points in the life course. The following categories were created to represent the age when traumatic events were first encountered: Trauma prior to age 18, traumatic events that arose between age 18 and age 64, and trauma that was encountered at age 65 or older. Because the strongest interaction effects between trauma and emotional support were found in the oldest–old cohort, this additional set of analyses was conducted only within this group. The findings (not shown here) reveal that the strongest statistical interaction effects emerged between emotional support and trauma arising between age 18 and 64 (b =.108; p <.001). Emotional support also appears to offset the effects of trauma arising prior to age 18, but the size of the relationship is not as large (b =.068; p <.05). Finally, and perhaps most importantly, the data reveal that emotional support does not significantly reduce the effects of traumatic events that were encountered after age 64 (b =.006; NS). Viewed broadly, these results underscore the importance of assuming a life course perspective because they show that current emotional support only offsets events that arose at least 20 years prior to the time the interview took place.
A third potential explanation for the findings from this study is that emotional support buffers the effects of lifetime trauma among the oldest–old simply because members of this cohort are able to get more emotional support in the first place. This possibility was evaluated with a univariate analysis of covariance. The demographic variables identified above were used as covariates. The findings reveal that mean levels of emotional support do not differ significantly across the three age cohorts (p =.875). Viewed in more general terms, these data indicate that the sheer quantity of emotional support may not explain the effects that were observed in this study and that the findings instead have something to do with either the quality of emotional support received by the oldest–old or the way in which members of this cohort utilize the assistance provided to them.
The final set of supplementary analyses has to do with cohort differences in exposure to lifetime trauma. Because the members of the oldest–old cohort have lived longer than the other study participants, it's possible they have been exposed to more traumatic events than their younger counterparts. If this is true, then the findings observed in the previous section may be due in part to greater lifetime exposure to trauma among the oldest–old. Once again a simple univariate analysis of covariance was performed to see if there are significant differences in the amount of trauma that was reported by members of the three age cohorts. The findings reveal that significant differences exist across cohorts (p =.005). After adjusting for the covariates, the following mean levels of exposure to trauma were found in each group: young–old (
= 2.275), old–old ( = 2.879), oldest–old ( = 2.831). A series of contrasts revealed that only the difference between the young–old and old–old was statistically significant (p <.01) and that the mean level of exposure for the oldest–old did not differ significantly from that of the young–old (p =.152) or old–old (p =.823).
|
Discussion |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
It is important to reflect carefully on what the findings from this study mean. The data indicate that the additive effects of trauma on life satisfaction are significant in all three age cohorts. This means that regardless of age cohort, all older people are vulnerable to the effects of lifetime trauma. But the findings further reveal that the stress-buffering effects of emotional support are enjoyed only by some older adults (especially the oldest–old), but not by others (i.e., the young–old). So the key theoretical question is why the oldest–old are able to draw more benefits from the emotional assistance provided by family and friends. There are two possibilities here. First, there may be something about the historical time in which members of the oldest–old were raised that influences how they seek out, receive, and utilize emotional support. Unfortunately, the literature has very little to say about the historical experiences of different cohorts of older people even though researchers repeatedly argue that these are important factors. Second, the findings observed in this study may reflect age rather than cohort effects. Clearly, the two cannot be disentangled with cross-sectional data, but providing a few brief comments about the influence of age may be helpful to those wishing to conduct research in this area. The classic work of Erikson (1959) that was discussed earlier provides one way of specifying why there may be age-related effects of trauma. Unfortunately, Erikson does not identify the age at which the final challenge of integrity versus despair is encountered. However, in his later work he notes that it is brought about by a growing awareness of death (see Hoare, 2002). Because the oldest–old are closest to death, perhaps Erikson's insights are most applicable to them. Carstensen's (1992) theory of socioemotional selectivity may also shed some light on why emotional support is so effective for members of the oldest–old cohort. She argues that as people grow older, they become increasingly oriented toward relationships that are emotionally supportive. To the extent this is true, the growing importance of emotional support may affect how the oldest–old utilize this form of assistance, and it may even influence the way in which others provide it.
The findings from this study have potentially important practice implications. There have been numerous calls for interventions that are designed to help people deal with the noxious effects of stress (Gottlieb, 2000). Unfortunately, advocates typically discuss interventions that are either aimed at stressful life events or chronic strains, while less attention has been given to lifetime trauma. This is especially true when it comes to interventions for older people. Correctly specifying the target group is one of the most important tasks in intervention design. If the target group is too broad, those who don't need help are included in the study. In contrast, if the target group is too narrow, then people who need assistance are excluded. There are three ways in which the findings from the present study may be useful for specifying the intervention target group. First, the data suggest that older adults who have been exposed to lifetime trauma are at risk. Second, the findings reveal that those who do not have sufficient emotional support should be included in the target group. This is especially true for members of the oldest–old cohort. Third, when working with members of the oldest–old cohort, the results indicate that elders who encountered traumatic events between ages 18 and 64 appear to be the most vulnerable.
Before conducting interventions along the lines outlined above, it is important to replicate the findings reported in this study. In the process, the limitations in this work should be overcome as well. Two are discussed briefly below.
First, the data used in this study are cross-sectional. This makes it impossible to determine the direction of causality between life satisfaction and traumatic events. Data that have been gathered at more than two points in time are needed to disentangle the temporal ordering of these constructs.
Second, data on the occurrence and age of exposure to traumatic events were obtained through retrospective recall. The accuracy of these reports has been hotly debated in the literature. Some maintain these reports are flawed (Maughan & Rutter, 1997), while others provide data showing that the extent of the problem has been exaggerated greatly (Bernstein et al., 1994). The only way to resolve this issue conclusively is to follow participants over the entire course of their lives. Since this simply isn't feasible, the only option is to rely on retrospective recall. Nevertheless, the potential problems with retrospective recall should be kept in mind as the findings from this study are reviewed.
Some time ago, Pearlin and Skaff (1996) argued that research on stress and the life course made a natural match. The findings from the present study suggest this makes a good deal of sense. However, it has been disappointing to find that few investigators have followed their sound advice. It is hoped that the empirical findings from this study encourage researchers to probe more deeply into the relationship among trauma across the life course and feelings of life satisfaction among our aging population.
|
Footnotes |
|---|
1 School of Public Health and Institute of Gerontology, University of Michigan, Ann Arbor.
Decision Editor: Linda S. Noelker, PhD
Received for publication March 4, 2004. Accepted for publication April 21, 2004.
|
References |
|---|
|
TOPAbstractDesign and MethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
