| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
RESEARCH ARTICLE |
1 Friedrich-Alexander University of Erlangen–Nuremberg and German Institute of Economic Research (DIW Berlin), Germany.
2 Max Planck Institute for Human Development, Berlin, Germany and University of Virginia, Charlottesville.
3 Berlin University of Technology (TUB) and German Institute for Economic Research (DIW Berlin), Germany.
Address correspondence to Frieder R. Lang, Institute of Psychogerontology, University of Erlangen–Nuremberg, Naegelsbachstrasse 25, 91052 Erlangen, Germany. E-mail: flang{at}geronto.uni-erlangen.de
 |
Abstract |
|---|
 TOP Abstract MethodsStudy 1Study 2ResultsDiscussionReferences |
|---|
How do scientific and medical advances in the understanding of human aging contribute to people's hopes and desires regarding the length of their lives and the conditions of death and dying? Scholarly work in gerontology and the medical sciences has reported major improvements in conditions of human aging (Baltes & Mayer, 1999
), an extended life expectancy (Robine, Crimmins, Horiuchi, & Zeng, 2007), increased "healthy active lifetime" (World Health Organization, 2006), and much physical, cognitive, and neurological plasticity in old age (Baltes, Reuter-Lorenz, & Rösler, 2006; Ball et al., 2002). However, it is not yet well understood how such advances contribute to personal wishes and expectations regarding the length of life. For example, although most people are well aware that the chances of living a long life have greatly improved, it is not yet clear—at least not based on sound empirical data—whether the facts on aging strengthen people's desire to live that long. Indeed, it does not go without saying that people want to live as long as scientific and medical progress makes possible, especially because the dysfunctional state of many of the oldest-old people is widely known and may have adverse motivational consequences.
The aim here is to lay the groundwork for studying the determinants of lifetime and end-of-life desires, and the prevalence of such desires. Given the relatively limited amount of past research in this area, we concede that a sound theoretical foundation for the sources of people's expectations about aging is lacking. To begin building such a foundation, we propose a dual-source information model of the nature of human aging.
The model contends that personal expectations are based on two major sources of information. The first, an experiential source, pertains to personal experiences with aging, such as an individual's evaluation of his or her current health. The second, a scientific-societal source, comes from scientific findings as reported in the mass media and as communicated by education, and from professional organizations dealing with aging issues. The influence of these scientific-societal sources on an individual depends on advances in the aging sciences, but also on the individual's access to knowledge, for example, the extent to which an individual has educational goals in life or interest in science.
As a first test of this model, we used relatively global indicators of aging expectations, such as desired lifetime and a sense of control over one's own death. The main objective here was to enrich our current knowledge about the individual-science-society interface by collecting information on three psychological facets of age-related expectations. First, what are people's individual lifetime preferences and how do they relate to experiential or scientific sources of aging-related information? Second, to what extent do people want to control the process of death and dying? Third, how do effects of personal experiences on people's desires regarding length of life depend on scientific knowledge? We explored these facets in two ways. In Study 1, prior to performing a telephone interview, we randomly exposed participants to three types of information about recent research findings on aging: good, bad, or no (control) scientific news. In Study 2 we aimed at partly replicating and extending the findings from the first study by using a face-to-face interview method.
Desired Longevity Differs From Expected Longevity
Researchers typically investigate expectations on aging by relying on estimations of perceived probabilities of reaching a certain age or on estimates of remaining years of life. For example, Mirowsky and Ross (2000) observed in a national U.S. sample that respondents, on average, estimated that they would live to the age of 81 years. Moreover, the healthier the participants were, the more support they had, and the better their socioeconomic situation, the longer life expectancy they reported. In recent years, aging researchers have begun to more systematically emphasize the critical role of subjective lifetime expectancy (Smith, Taylor, & Sloan, 2001; Ziegelmann, Lippke & Schwarzer, 2006), as well as expectations about remaining future lifetime (Fung & Carstensen, 2006; Lang, 2000). However, high estimates of personal life expectancy may not necessarily imply the actual desire to live that long.
The paucity of empirical work on individual desires regarding longevity and the end of life is particularly puzzling when one considers that disciplines such as psychology, gerontology, and economics often build on motivational concepts related to death anxiety (Tomer, 2000), denial of death (Slemrod, 2003), subjective survival probability (Hurd & McGarry, 2002), mortality risk perception (Gan, Gong, Hurd, & McFadden, 2004), mortality salience (Cicirelli, 2006), and subjective proximity to death (Lang, 2000). However, such concepts reflect expectancies rather than values or preferences. Few studies in the aging literature explicitly address individuals' longevity desires. Cicirelli found that older adults, on average, wanted to live longer than they actually expected to live. This discrepancy increased with age. It is not clear, however, to what extent age-related differences in longevity desires are rooted in differences in personal experiences or by science-based sources of information on aging.
Scientific-Societal Information on Human Aging
In general, two contradictory predictions can be made in relation to the desire for greater longevity and higher levels of vitality. On the one hand, in recent history, society has witnessed increasing vitality and activity in old age, which can be expected to generate positive expectations about longevity and the quality of life in old age. However, much of the good news from recent aging research relates to the "young-old" or "third age" (Baltes & Smith, 2003; Laslett, 1991). This phase of life has shown significant positive changes in life quality and in physical as well as mental fitness since the beginning of the 20th century.
On the other hand, aging is not always a positive experience. Recent findings about the calamities encountered in the "fourth age" (Baltes & Smith, 2003) suggest drastic losses of cognitive and mental fitness. People's understanding of the last phase of life is likely to entail an awareness of such findings. Here the high prevalence of dementia, physical frailty, and multimorbidity may possibly outweigh people's positive expectations about long life and very old age. In fact, there is much evidence suggesting that aging also implies sources of dissatisfaction, which are often based on personal experiences and decreased vitality (Baltes & Mayer, 1999). Such findings may reflect a sense of hesitation to welcome demographers' findings of increasing longevity in modern times as a "gift" of longer life.
The Related Topic of Death and Dying
Attitudes toward one's own death and dying have been explored in relation to concepts such as death anxiety or mortality salience (Tomer, 2000). For example, having better socioeconomic resources is related to a stronger will to maintain control over the circumstances of one's own death (Cicirelli, 1997). Ditto, Druley, Smucker, Moore, and Danks (1996) explored preferences of living over dying with hypothetical scenarios of health impairments: Young and old adults rated whether they would prefer to either live or die if confronted with several health impairments (e.g., coma, confinement to bed). Response patterns were nearly identical across age. Young and old adults expressed strong preferences for not wanting to live in situations of severe mental or physical incapacitation. Another critical issue is how scientific advances in the medical sciences may trigger concerns about the continuation of medical treatment when one is unable to give informed consent. An indication of the personal desire to control the conditions of one's death and dying may thus be expressed in a living will containing instructions to relatives and doctors regarding medical treatment in case of an incapability to decide.
In the present study, we explored effects of experiential sources and scientific-societal sources of aging-related information on desired longevity ("what age would you like to reach?"), desire to control when and how to die, and on depositing a living will. We expect a greater susceptibility of lifetime desires to personal-experiential sources of aging information, and a relative robustness against the science-based sources of information. Thus, we conceive of desired longevity as a dependent variable rather than as a predictor of adaptation processes across adulthood; this is in contrast to other research.
In Study 1, we randomly divided participants into three groups before the interview: the first group received good news on aging; the second group received bad news on aging; and the third, neutral (control) group received no information. We also added measures dealing with several facets of experiential (e.g., preference of high life quality or high longevity) and science-based (scientific interest) sources of information on human aging. Whereas Study 1 was based on a telephone interview, Study 2 aimed at partly replicating and extending the findings in a face-to-face computer-assisted interview (CAPI) situation.
|
METHODS |
|---|
|
TOPAbstractMethodsStudy 1Study 2ResultsDiscussionReferences |
|---|
). |
STUDY 1 |
|---|
|
TOPAbstractMethodsStudy 1Study 2ResultsDiscussionReferences |
|---|
|
Desired Lifetime Duration
Participants responded to this question: "What age would you like to reach?" (The wording of the German item of desired lifetime duration is "Wenn Sie einmal ueber die Laenge ihres Lebens nachdenken, Wie alt wuerden Sie gerne werden?") We accepted all given answers and sorted them online into a 14-point scale, with 1 as "below 60 years," 2 as "60 to 64 years," and so on, up to 14 as "120 years or older." We transformed responses on the basis of midcategory centers of each age bracket (1 = 60 years, 2 = 62 years, 3 = 67 years, ... 14 = 120 years). In addition, participants rated whether social policy and research efforts should aim at increasing the quality of life, increasing longevity, or both. We used preference for life-quality enhancement over longevity as a dichotomous variable (yes = 1, no = 0).
Desire to Control Death and Dying
Participants indicated on a 7-point scale (from 1, absolutely don't agree to 7, fully agree) their agreement with this statement: "I want to determine on my own, when and how to die" (M = 5.7, SD = 2.0).
Subjective Health
Participants responded to three items assessing their physical and mental health on a 7-point rating scale (e.g., "How healthy do you feel, physically and mentally?"). The alpha value of the three items was 
= 0.67. Higher values indicate better health.
We included indicators of socioeconomic resources and personality as covariates, that is, education (years of schooling to completion of highest educational degree), size of household, and income (average weighted household income), plus two ultrashort personality constructs: neuroticism (three items, = 0.59) and openness to experience (four items, = 0.62). We tested both personality scales in an earlier study (Lang, 2005), which yielded acceptable coefficients of 6-week retest reliability (r >.75) and converging validity with established scales.
Manipulation Check
As a check of the informational conditions, we had participants rate the proportion of things improving with age (i.e., "When evaluating aging with 100 points in total, how many points reflect how much is getting better with age, and how many reflect how much is getting worse?"). Responses had to add up to 100 points. On average, participants expected 46% (SD = 19.1) of aging changes pertaining to things getting better. As intended by the experimental conditions, there was a statistically significant effect in the expected direction of the information groups on perceptions of how much is improving with age. Respondents who were given bad news on aging expected more negative age changes (M = 44.1, SD = 19.9) than those who received good news (M = 47.8, SD = 20.1) or no news (M = 46.2, SD = 17.0); F(2; 1,093) = 3.4, p <.05, 
2 =.01. Although the manipulation effect appears small, it is of substance in this telephone study. The effect serves to show that, on the whole, participants were responsive to the statements of positive or negative information on recent scientific advancements.
|
STUDY 2 |
|---|
|
TOPAbstractMethodsStudy 1Study 2ResultsDiscussionReferences |
|---|
Desired Lifetime Duration
We added this preliminary statement to the item from Study 1: "When thinking about your total lifetime, what age would you like to reach?" (See the Desired Lifetime Duration section for the original wording.) Respondents produced a number in years. We accepted all given answers of respondents, and we recoded them as in Study 1. Moreover, we added one new response format ("I don't care"); 101 participants, or 24.2%, opted for this response.
Two additional items dealt with expectations and the determination of desired longevity: First, subjective probability of desired lifetime ("How likely is it that you will live that long [to the age you want to reach]?") was rated on an 11-point scale ranging from 0, absolutely unlikely, to 10, very likely. (The wording of the German item on subjective probability is "Fuer wie wahrscheinlich halten Sie es, dass Sie tatsaechlich so alt werden?") Second, determination of desired lifetime duration ("How desirable would it be for you to reach the age [you want to] if you were confronted with a health problem resulting in frailty or the need for health care?") was rated with a 4-point scale ranging from 1, not at all desirable, to 4, very desirable. For preparation of a living will, participants stated whether they had signed a legal document giving instructions for medical treatment in a case of incapacity to decide ("Have you prepared a personal living will providing instructions for medical treatment in case of severe injury or disease?" Yes = 1, no = 0.)
Finally, participants rated one item assessing their interest in science and technology ("How interested are you in issues related to science and technology?"). Participants rated the item on a 5-point scale (ranging from 1, not at all interested, to 5, very strongly interested).
We assessed subjective health with a single item with a 5-point scale ("How would you rate your current health?"). Covariates were education (years), income, occupational status (occupied = 1, not occupied = 0), marital status, household size, living with children in the same household, and—as in Study 1—neuroticism ( = 0.53) and openness to experience ( = 0.76).
Missing Cases Treatment (Both Studies)
In Study 1, with respect to the focal criterion variable of desired lifetime, a total of 6.3% (n = 75) of participants did not respond. A missing value analysis revealed that missing cases are nonrandom and nonignorable (Little's MCAR, or missing completely at random, test; 
2 = 70.0, df = 29, p <.001). Separate variance t tests and logistic regression showed that nonresponders as compared with responders were older (t = 3.8, df = 81.7, p <.01; odds ratio or OR = 1.03, p <.01) and more likely to be unemployed (t = –2.1, df = 80.1, p <.05; OR = 2.27, p <.01). Although findings remain unchanged when we included missing-imputed values based on the expectation maximization method, we decided to report analyses in what follows after we dropped these 75 cases for reasons of space. Estimates of age and occupation on desired lifetime may nevertheless be biased as a result of nonrandom sample selectivity.
In Study 2, there were 23 missing cases (4.6% for the question of desired lifetime duration, in addition to the 101 responses of "I don't care." This means that there were 376 cases with valid numerical answers to this question. Participants who gave the "I don't care" response did not differ significantly from other participants with respect to central variables such as age, gender, education, income, death of a parent, or subjective health. In order to test whether the missing responses had an effect on the observed pattern of associations, we entered all variables used in the subsequent analyses in a multivariate analysis: a nonsignificant Little's MCAR test (2 = 82.2, df = 71, p =.17; n = 500). This means that excluding the nonresponding participants from the analyses does not significantly alter the observed associations among variables.
Weighting
We did not weight data from both studies to correct for sample bias. One reason is that the focus of the present research is on relationships within individual response patterns rather than on the estimation of a population distribution.
|
RESULTS |
|---|
|
TOPAbstractMethodsStudy 1Study 2ResultsDiscussionReferences |
|---|
Expectations about Longevity and End of Life
Table 1 displays means for the desired lifetime duration across the three information groups (good news, bad news, and no information on aging) in Study 1, together with results for Study 2. To our surprise, in Study 1 there were no considerable main effects of the preassessment informational groups. Moreover, desired lifetime duration did not differ significantly between the two study samples (t = 1.94, df = 1499, ns, Cohen's d = 0.12). This is an important check of robustness because the two studies relied on two different interview methods (telephone and face-to-face).
The findings of Study 1 were especially surprising to us. In general, giving different information on the positive or negative direction of recent scientific findings on aging did not alter preferences for desired lifetime. There were only a few significant differences in the correlation pattern between the three information groups: when bad news about aging is heard, a stronger preference for enhancement of life quality over longevity is associated with a greater desire for control of death (r =.17, p <.001). We did not observe such associations in the other groups. In sum, the results point mainly to the generality and modesty of expectations about desired length-of-life and end-of-life issues.
Consistently across both studies, the response patterns on desired lifetime were weakly associated with chronological age (Study 1: r =.07, p <.05, N = 1,125; Study 2: r =.17, p <.01, N = 376; after the exclusion of adults >80 years, r =.13, p <.01; n = 364). Figure 1(a) shows the percentage distribution of responses across the four age groups in Study 1. As one can see in Figure 1(a), the distribution of responses was fairly similar across age groups. Older adults (>65 years), on average, expressed a modest desire to live longer than did young and middle-aged adults. Figure 1(b) shows the distribution of desired lifetime in the face-to-face-interview of Study 2, which serves to replicate the findings from the telephone interview study. Only the few participants who were older than 80 years (n = 12, M = 84.8 years, SD = 3.8) deviated from the pattern for obvious reasons, with an average desired lifetime of 93.3 years (SD = 5.8).
|
|
|
We expected that older age groups—as a result of their greater temporal proximity to death—would be more likely to want to control the circumstances surrounding death. This assumption was not borne out. As Table 3 shows, the desire to control one's end of life and the completion of a living will was not related to chronological age. In Study 1, a preference for life quality over longevity and gender were predictive of a desire to control death and dying. Men were less likely to want to control death and dying than women were (OR = 0.73, SE = 0.14, p <.05). Preference of quality of life over quantity was associated with a greater likelihood to desire control over death and dying (OR = 1.50, SE = 0.20, p <.05). In Study 2, only one experiential predictor and none of the science-societal predictors predicted the completion of a living will: When respondents had experienced the death of a parent, they were more likely to have a living will. There was no difference between men and women.
|
|
|
DISCUSSION |
|---|
|
TOPAbstractMethodsStudy 1Study 2ResultsDiscussionReferences |
|---|
There is very little comparable data from other historical epochs or cultural contexts that would shed further light on the central question of what kinds of secular shifts are occurring in late life and longevity expectations in modern societies. In a sense, then, the present study can be seen as providing an impetus for further historical and culturally comparative research. In laying the groundwork for a dual-source information model of human aging, we find that the desire for high longevity depends to some extent on experiential sources of aging information but is relatively robust to scientific-societal information sources. Consistent with the work of Cicirelli (2006), we find that desired longevity exceeds the subjective life expectancy reported in other studies (Mirowsky & Ross, 2000), and that desired lifetime duration depends heavily on current health ratings and other experiential factors. In general, we were surprised by three findings: first, the relative modesty of individual longevity desires; second, the relatively strong preference for active control over one's own death and dying; and third, the relatively low impact of scientific sources of information on aging. Apparently, most Germans in this heterogeneous population sample were more alike than different with respect to lifetime expectations and end-of-life desires. However, the differences observed are of theoretical and practical relevance.
Our findings are consistent with the dual-source model of processing information about aging expectations. The information content provided by the two sources—scientific-societal expectations and personal experiences—often converge in a positive view of aging for the young-old in the third age of life. However, the situation is different for the oldest-old in the fourth age. In this respect, the two sources of information about the nature of aging produce dissonance, in which personal experience plays the more powerful role. Consistent with this is the finding that subjective ideal longevity is associated with subjective health, irrespective of current age, across a broad and heterogeneous sample of adults from 20 to 80 years of age. Effects of subjective health exceed the impact of good versus. bad news on aging. This suggests that expectations about aging reflect idiosyncratic perceptions of one's present health conditions rather than general insights from the science of aging. Individual preferences to live a modest number of years reflect individual views about life quality. For instance, individuals who preferred to live a good rather than a long life preferred a shorter lifetime.
About two thirds of respondents expressed the desire to decide when and how to die. Women expressed a stronger desire to control death and dying than men. One explanation may be that women are more often in the role of family caregiver and are more likely to be confronted with issues related to dying. Such experiences may enhance the desire to exercise control over death. We did not systematically ask for personal experiences related to the death of relatives. However, in Study 2, death of a parent proved to be the best predictor for preparing a living will.
Findings point to a limited relevance of knowledge-related goals in determining one's desired lifetime. Strong interest in science appears to enhance individual desires to live longer while at the same time improving the sense that such hopes are still unrealistic at this time. Having strong interests in science may cause more sophisticated expectations about what is possible and what is desirable in science. However, in Study 1, the informational conditions of good versus bad news on aging had no effect. Processing news from aging research may depend on the willingness to process such information. Clearly, our single-item measure may not have captured all relevant aspects in this respect. More research is warranted addressing the historical, contextual, and idiosyncratic conditions of desired lifetime and end-of-life desires more explicitly, and including a more in-depth assessment of scientific sources of information about aging.
One caveat has to be added when one is interpreting the findings. Our interview methods rely on self-reporting and may thus be positively or negatively biased. This is, however, a general problem in survey research. Furthermore, other (e.g., visual) sources of sensory input in Study 1 may have been more effective. Generally, the amount of variance accounted for in Study 1 was lower than that in Study 2. This may be due to the different sample sizes but also to the telephone interview method, which may lead to greater noise in the data. Visual sources of information on good or bad news about aging might have been processed more efficiently. Our method of presenting good or bad news may not have been the strongest test for a scientific impact of aging expectations, but our manipulation check produced findings that showed that the information was cognitively processed and did yield different expectations about aging changes. The information manipulation thus appears to have worked to some degree. Specifically, respondents who received good news on aging expected a greater percentage of positive aging changes than respondents who received bad news.
Expectations about longevity and the end of life are known to influence motivational processes throughout adulthood (Fung & Carstensen, 2006; Lang, 2000). The current research suggests that personal perspectives on longevity are fairly robust against scientific sources of information but are tied strongly to personal life experiences. Scientific news is thus only part of the story. Future research will have to take into consideration how people's ideals and desires for aging are associated with other everyday representations of aging as well.
In conclusion, the present study is but a first step in a new path of inquiry. To understand the role of old age in a changing society, we need to better understand not only the positive changes in longevity resulting from medical advances, but also the potentially negative implications resulting from the vulnerabilities of the fourth age—for instance, with increasing risks of dementia—that contrast the positive perspective of a "happy gerontology" (Bobbio, 2001). It seems particularly important to begin monitoring and analyzing the impact of such changes in expectations about aging more systematically. It is likely that future comparative analyses will produce a more complex picture that reflects the varying and differential conditions of aging. For the present, the increased longevity so celebrated by demographers seems to be seen not just as a blessing by most people.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication September 14, 2006. Accepted for publication April 13, 2007.
|
References |
|---|
|
TOPAbstractMethodsStudy 1Study 2ResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
