| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
Max Planck Institute for Human Development, Berlin, Germany.
Address correspondence to Jacqui Smith or Alexandra Freund, Center for Lifespan Psychology, Institute for Human Development, Lentzeallee 94, 14195 Berlin, Germany. E-mail: smith{at}mpib-berlin.mpg.de or freund{at}mpib-berlin.mpg.de
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
PSPECIAL SECTION
ossible selves are highly personalized hoped-for and feared images of the self that function as incentives for action (Brandtstädter, 1998; Markus & Herzog, 1991). These images contain information about self-related concerns that guide decisions about what goals to work on, where to expend time and effort, what to avoid or resist, and what to abandon. In this sense, possible selves can be regarded as guides for selection processes (M. M. Baltes & Carstensen, 1991; Freund & Baltes, 2000). Possible selves also reflect an individual's motivation to try to control the direction of her or his future life: motivations such as self-improvement, self-maintenance, or efforts to minimize loss. It has been suggested that individuals change and recalibrate their possible selves in response to changes in life circumstances in order to maintain well-being (Cross & Markus, 1991; Hooker, 1992, 1999; Markus & Herzog, 1991; Ryff, 1991).
Although several cross-sectional studies have included older adults (e.g., Cross & Markus, 1991; Hooker, 1992; Ryff, 1991), most research on the possible selves model of Markus and colleagues has been undertaken with college students and adolescents (e.g., Fraser & Eccles, 1995; Markus & Ruvolo, 1989; Oyserman & Markus, 1990). There are still many questions about the operation of this motivational system oriented to the future in the period of old age. To address some of these questions, we investigated the life domains and motivational orientation reflected in the possible selves of adults aged between 70 and 100+ years who participated in the Berlin Aging Study (BASEBASE: P. B. Baltes & Mayer, 1999). We examined cross-sectional age differences and patterns of stability and change over a period of 4 years.
The literature on the self in old age suggests that there might be differences between the young-old and oldest old in terms of the number, content, and dynamics of possible selves (e.g., Markus & Herzog, 1991). To begin, it is recognized that, for the majority of the oldest old, declining health and losses in life quality are inevitable and that this places strong constraints on achieving new goals (Smith, 2001; Smith & Baltes, 1999). Furthermore, the transition to very old age is generally accompanied by an increased awareness of the closeness of one's own death that influences the types of personal goals selected in late life and the temporal extension of these goals (e.g., Carstensen, Isaacowitz, & Charles, 1999; Johnson & Barer, 1997). The awareness of a limit to the personal time left to live is reflected in the temporal focus of the self-concept: Thoughts about the future may be overshadowed by a focus on the present and by the construction of a self-narrative about the whole of life (Whitbourne, 1985).
Given these proposals, we expected that people in their 70s would report more hopes and fears than would very old individuals (>80 years old) and that a substantial subgroup would report having no hopes or fears. If old age is a period when some possible selves are cast away (Markus & Herzog, 1991), we should also observe a decrease over the 4 years in the number of hopes and fears mentioned and an increase in the proportion of participants who report no future images of the self. We expected these changes to interact with age: Most change should be observed in persons aged over 80 years. Alternatively, it could be that adaptive processes of compensation, selection, and optimization in old age lead not to changes in sheer number but rather to a reduction with increasing age in the range of different domains in which one has hopes and fears (Freund & Baltes, 2000).
Possible selves function as a motivational system primarily within domains. Markus and Ruvolo (1989), for example, proposed that when a domain gains in personal salience, an individual develops countervailing hopes and fears in that domain. Hopes are linked to strategies outlining what to do to attain domain-specific goals, and the matched fears offset or balance these hopes with images of what could happen if the desired states were not realized (Oyserman & Markus, 1990). Matched hopes and fears in a domain indicate a high level of motivational control in that domain and the availability of means and strategies for working on future goals (Markus & Ruvolo, 1989). Knowing in which domains older adults have hopes and fears and in which domains matched possible selves are constructed can therefore tell us much about the operation of the system.
Whereas concerns about the social, interpersonal, and occupational self are prominent domains in the future scenarios of young and middle-aged adults, in later life the domain of health is seen as being more salient (Hooker, 1992; 1999). Cross and Markus (1991), however, reported that although the fears of 60 year olds were about health and lifestyle constraints (e.g., becoming physically dependent), their hoped-for selves were about personal characteristics. Similarly, in interviews with men and women aged over 85 years, Troll and Skaff (1997) found that perceived changes in personality characteristics (and especially positive changes) were mentioned more often than changes in physical health, everyday competence, and lifestyle. Other researchers suggest that strivings for satisfying interpersonal relationships are important in old age (e.g., Antonucci, 1990; Carstensen et al., 1999). In addition to health, we expected that concerns about social relationships and personal characteristics would be the focal domains of the possible selves of individuals aged between 70 and 100+ years.
A further question concerns the motivational orientations of possible selves in old age: Do they indicate desires for self-improvement (gain), self-maintenance, or efforts to minimize or prevent losses? Bearon (1989) suggested that whereas middle-aged adults focus on achieving hoped-for selves (e.g., acquiring material possessions), older adults are more concerned with ensuring predictablity (i.e., maintenance) and preventing or avoiding feared selves such as illness and dependency. We expected that, across all domains, the central motivational orientation of hopes would be maintenance and of fears it would be loss prevention or minimization.
Possible selves are portrayed as being acquired, maintained, transformed, and given up over time (Markus & Herzog, 1991). In this regard, we expected to observe both stability and change in the profile of domains over the 4 years. Whereas self-concerns about long-term projects (e.g., health maintenance) may be more stable, hopes and fears about short-term or time-specific concerns may change considerably over time. Given the increased likelihood of declining health in very old age, we expected health to become the focus of the possible selves of most individuals over time. Impaired physical and sensory functioning could play a role in giving up possible selves linked to activities and interests, and the loss of social partners may be associated with a reduction in hopes and fears about social relationships. Time may also play a role in changing the overall motivational orientation from striving for maintenance to a focus on the management of loss (Heckhausen, 1998; Staudinger, Marsiske, & Baltes, 1995).
A further aspect relates to the extent to which the structure of possible selves in old age involves matched hopes and fears within domains. Is there, for instance, a trend toward more integrated or compartmentalized possible selves? A trend toward a highly integrated (dedifferentiated) structure might involve the addition or maintenance of matched hopes and fears over the 4 years or a restriction of matched possible selves to highly salient late-life domains like health. A trend toward compartmentalization (differentiation), in contrast, might involve the addition or maintenance of more fears than hopes or the increased domination of possible selves by fears. This latter prediction is consistent with proposals that very old age is a period when self-regulation strategies designed to maintain a sense of well-being are pushed to their limits (Smith & Baltes, 1997, 1999). Although the literature is clear that the dynamics of possible selves reflect changing life circumstances, proposals about specific connections and the time frame of these connections are rare. For this reason, we report exploratory analyses of the relationships between possible self-characteristics and individual differences in declining health, changes in life contexts, and well-being.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Measures
Possible selves task and scoring.
At baseline and at follow-up, the BASE possible selves task was part of an individual face-to-face interview session assessing self and personality constructs (Smith & Baltes, 1999). We adapted the present open-end task from ones described previously (e.g., Cross & Markus, 1991; Hooker, 1992; Ryff, 1991) so that it would be suitable for interviews with people aged 70 to 100+ years and practical given our time constraints. Participants were asked to generate at least two important hopes and fears about their future. Instruction was the same at both measurement occasions. The task began with the collection of hoped-for selves with the following instruction: "We are interested in your personal wishes and hopes. Who would you like to become someday? For example: What sort of a person would you like to be? What experiences would you like to have? What feelings would you like to have? What are your personal wishes and hopes? I would like to..." The instruction for feared selves was as follows: "You have just described who you would like to become someday. Sometimes people also have images of themselves that they would dread or fear becoming. Please describe two images of yourself that you fear or dread. For example: What sort of a person would you not like to become? What experiences would you not like to have? What feelings would you not like to have? What images of yourself do you fear or dread? I do not want to..." Participants' spontaneous responses were tape-recorded and transcribed verbatim.
There were two steps in the procedure used to code the responses about hopes and fears. The first step involved coding the number of hopes and fears and the content domains mentioned. Transcribed protocols were initially segmented into meaningful idea units using a coding system developed to quantify aspects of self-definition in old age (Freund & Smith, 1999). The system specified segment categories (e.g., statements about domains such as family, friends, health, or comments about not having hopes or fears). A segment was defined as beginning with a new thematic content category and ending with the start of a different category. Two pairs of independent raters (one pair per measurement period) subsequently coded these segments into 1 of 24 content categories, using a computer-based randomized procedure in order to avoid context effects. Transcripts of hopes and fears were coded separately, as were the transcripts from the two measurement occasions. Because the content categories mentioned in transcripts were coded at different times in the study and by different pairs of coders, we tried to calibrate the scores by training the coders of the follow-up transcripts until their ratings of a subsample of the baseline transcripts matched those of the initial coders. Intercoder agreement on each occasion was satisfactory: baseline 
= .82, and follow-up = .82.
These data were aggregated to provide the following scores for each participant for the baseline and follow-up assessment: total number of hopes and fears, number of different domains of possible selves (maximum = 6), presence of hopes and fears in six life domains (0 = no, 1 = yes), presence of matched hopes and fears within each of the six domains (0 = no, 1 = yes), and total number of matched domains (maximum = 6). The six aggregated domains were as follows: personal characteristics (e.g., personality dispositions, emotions, physical appearance, body image, political and religious attitudes), health and functional capacity (e.g., illness, physical mobility, vision, hearing), interests and activities (e.g., indoor and outdoor hobbies, daily routines), social relationships (e.g., contacts with partner, children, friends), life events (e.g., thoughts about death and dying, moving), and cognition (e.g., comments about memory, dementia). Examples of hopes and fears in these six domains are included in Table 1.
|
Measures of life changes: health, life events, and well-being.
Exploratory analyses examined the links between changes in possible selves and changed life circumstances. For these analyses we used a subjective and objective measure of health, a question about significant life events, and measures of two dimensions of subjective well-being. The objective measure of health (number of illnesses) was based on physician-observed ICD-9 diagnoses of illnesses weighted by severity. The diagnoses were determined in a clinical examination at baseline and at follow-up and supported by additional blood and saliva laboratory assessments (see P. B. Baltes & Mayer, 1999). Self-rated health was measured on a 5-point response scale, ranging from 1 very poor to 5 excellent, with the standard single-item question, "How would you rate your health at the present time?"
Life events during the period between baseline and the follow-up were assessed in several ways in BASE. Comparison of residential address and marital status information indicated changes for 20 individuals (10%): Six people moved into a home, two moved out of a home into a private residence, 10 people were widowed, and two were married. The BASE protocol also included a self-report question about significant intervening life events. Fifty people (24%) reported that an event had occurred that they perceived to have made a significant change in their life. Because these subgroups overlapped, we combined them for analyses.
Factor scores for two dimensions of subjective well-being, satisfaction with aging (five items) and life satisfaction (four items), were obtained with a German translation of the Philadelphia Geriatric Center Morale Scale (Lawton, 1975; Smith, 2001). Each item was read aloud by an interviewer and the participant's response on a 5-point scale was recorded.
For the measures of health and well-being we computed time-adjusted difference scores as measures of change appropriate for this two-wave longitudinal study (Rogosa, 1995). Difference scores (expressed in t scores) between T1 and T3 were adjusted for time because the interval between measurements was not the same for all participants (M = 3.77 years, range = 2.37–5.18 years) and because these interval variations may be differentially important for the younger and older participants. Scores were standardized to the larger core BASE sample (N = 516). Mean difference scores (standardized as t score units per year) were as follows: subjective health, M = -0.62, SD = 2.41; life satisfaction, M = -0.54, SD = 2.53; and aging satisfaction, M = -0.83, SD = 2.40.
Procedure
All testing was carried out in the participant's place of residence (e.g., private home, nursing home, hospital) by trained research assistants. Responses about possible selves were collected as the second task in a 90-min interview session about self and personality (see Smith & Baltes, 1999, for details about this session). The first task was also open-ended: "Who am I?" self-definitions at T1 (Freund & Smith, 1999) and "What I like/dislike about my age" at T3. All responses were spoken by the participants and recorded by the interviewer (as well as tape-recorded). Measures of health, functional capacity, and well-being were obtained in different sessions of the BASE Intensive Protocol (see P. B. Baltes & Mayer, 1999).
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Table 2 summarizes the findings for the characteristics of possible selves collected at T1 and T3 and includes, for comparison, baseline information for the subgroups of the core BASE sample who did not continue in the longitudinal study (n = 203 individuals who were deceased and n = 107 people who were alive at T3 but did not complete the task then). Significant age- and time-related differences are indicated in Table 2.
|
As predicted, at baseline and at follow-up there were age-related differences in the number of hopes generated. Younger participants (aged 70–79 years) on average generated significantly more hopes (baseline M = 2.48; follow-up M = 2.49) than the older group (aged 80–100+ years: M baseline = 2.25, and M follow-up = 2.12), F(1,204) = 5.54, p < .05. There were no age-related differences in the number of fears mentioned.
Whereas there were no significant time-related differences in the number of hopes generated, the average number of fears decreased over time (T1 M = 2.24, T3 M = 1,87), F(1,204) = 17.73, p < .001. We had also expected that the number of different domains covered by the possible selves scenarios would show age/cohort differences. Across hopes and fears, 95% of subjects nominated at least two different domains on both occasions (T1 M = 2.92; T3 M = 2.84). At baseline, only 7 (3%) participants focused on a single domain, and at follow-up it was 5%. The age comparison for number of different domains was not significant (p = .113). At both baseline and follow-up, 35% of participants had no matched possible selves. The majority had a matched possible self in one domain (50% at T1 and 55% at T3).
Content Domains of Hoped-For, Feared, and Matched Possible Selves
Because theory suggests that possible selves primarily function within domains, we ran separate analyses in the six domains of possible selves coded. Comparisons between the longitudinal sample and the two subgroups of noncontinuers at baseline revealed differences in the likelihood of mentioning two domains of hopes, one domain of fears, and in two domains of matched possible selves. Proportionately fewer of the nonsurvivors mentioned hopes about personal characteristics, 
2(2, N = 516) = 11.3, p < .01, and fears about health, 2(2, N = 516) = 6.3, p < .05. More of the nonsurvivors than the longitudinal sample or the surviving nonparticipants mentioned hopes about life events, 2(2, N = 516) = 6.9, p < .05, especially death and dying. Nonsurvivors were less likely to have matched possible selves about personal characteristics, 2(2, N = 516) = 6.68, p < .05, and social relationships, 2(2, N = 516) = 8.68, p < .01, and overall they had fewer domains that were matched in their possible selves, 2(2, N = 516) = 4.68, p < .01). Again, these differences were a function of the advanced age of the nonsurvivor subgroup: When age was added as a covariate, group differences were no longer significant.
The most prominent domains of hopes and fears mentioned at both measurement occasions by the longitudinal sample were personal characteristics and health (see Table 2). The domains of health and personal characteristics were also the focus of matched possible selves (e.g., 27% of the sample had matched possible selves for health at T1 and 32% at T3). The majority of participants had at least one matched possible self: On both measurement occasions, 35% of participants had no matched domains in their possible selves.
We had expected that hopes and fears about family and social relationships would be mentioned as often as possible selves about health and personal characteristics, but this was not the case. Social relationships also was the only domain for which there was an age/cohort difference: At follow-up, participants aged over 80 years reported fewer hopes in this domain compared to the 70–79-year-olds 2(1, N = 206) = 7.75, p < .01, and fewer older participants had matched possible selves about social relationships, 2(1, N = 206) = 5.75, p < .05. In current self descriptions, Freund and Smith (1999) found that BASE participants frequently mentioned their activities and interests. In the possible selves scenarios of the same participants, however, this domain was less prominent and indeed was rarely mentioned as a feared self. Given the realistic constraints associated with declines in health, activities and interests may be things that people are prepared to abandon (or to leave out of their future self images) because they can gain esteem from memories of past acts and present achievements. Although concerns about cognitive functioning (memory, dementia) are often thought to occupy the minds of older adults, this domain was mentioned rarely in the possible self images of BASE participants.
Although the group profile of possible self domains were very similar at baseline and at follow-up, there were interesting differences in the rank orders of the domains of hoped-for and feared selves at the two occasions. Personal characteristics was the most frequently mentioned domain of hopes at baseline (53%) but at follow-up the most prominent domain of hopes was health (55%). The odds ratio (OR) that a hoped-for possible self related to personal characteristics was mentioned was 1.24 greater at T1 than at T3 (53% vs 41%; confidence interval [CI] = 1.39–1.09) and was similar in the two age/cohort groups. In contrast, the odds of mentioning hopes about health were greater at T3 than at T1 (OR T3:T1 = 1.62; CI = 1.7–1.54). This time-related difference was significant in the 70–79 year age group (OR = 2.02, CI= 2.19–1.85) but not in the older group (>80 years): 60% of the younger participants mentioned hopes about health at T3 compared with only 43% at T1.
The rank order of the domains of feared selves in the domains of health and personal characteristics did not differ at the two measurement occasions. The most prominent fears were about health (57% at T1 and 54% at T3). However, all BASE participants were more likely to mention fears about life events at baseline (34%) than at follow-up (24%, OR = 1.63). People aged 80 years and older were more likely to mention fears about social relationships at baseline than at follow-up (OR T1:T3 = 1.43), whereas people aged 70–79 years were more likely to mention fears in this domain at follow-up (OR T3:T1 = 1.35).
The odds of a matched possible self about personal characteristics was also greater at T1 than at T3 (31% vs 19%, OR = 1.91, CI = 2.13–1.69) and did not differ across age groups. For people aged 80 years and older, the likelihood of having a matched possible self about health was greater at T3 than at T1 (OR = 1.6, CI = 1.78–1.42). For the younger group, matched possible selves about health were equally likely on both occasions. Older participants were also more likely to have matched possible selves about social relationships at T1 compared with T3 (OR = 2.2, CI = 2.5–1.9). There were no time- or age-related differences in the other domains of hopes, fears, or matched possible selves.
Motivational Orientation
We had expected that the central motivational orientation underlying hopes would be maintenance and that underlying fears would be avoidance of loss. However, as reported in Table 2, the dominant motivational orientation in hoped-for selves was improvement (e.g., hopes to attain, achieve, or re-experience something). For fears, the dominant orientation was, as expected, avoidance. There were no age/cohort differences in these measures in the longitudinal sample.
At baseline, compared with the longitudinal sample and the surviving nonparticipants, proportionately fewer of the nonsurvivors (n = 203) had statements about maintainence in their hoped-for selves, 2(2, N = 516) = 6.79, p < .05, and fewer had statements about avoiding loss in their feared selves, 2(2, N = 516) = 7.92, p < .05. More people included some statements about maintenance in their hoped-for selves, however, at follow-up (60%) than at baseline (43%, OR T3:T1 = 1.99, CI = 1.8–2.2).
Intraindividual Dynamics Over Time
To gain some insight into the extent of change and stability of the possible selves of the BASE participants at the individual level, we cross-tabulated the presence or absence of hopes, fears, and matched possible selves at T1 and T3 for each domain. This allowed us to identify those individuals who remained stable over time (either by consistently mentioning or not mentioning the domain, i.e., the diagonal of the 2 x 2 matrix) and those individuals who had changed over time either by adding or deleting the domain at follow-up. The results of this domain-specific classification are given in Table 3. We also aggregated across domains, classifying individuals according to whether, in general, their numbers of hopes, fears, and matched possible selves were stable or changed over time. Because the time interval between measurement was not the same for each person, we computed time-adjusted difference scores for continuous variables (e.g., number of hopes) as a further measure of change.
|
The percentage of individuals showing stability and change differed by domain (see Table 3). Analyses at the domain level suggest more intraindividual stability than change. Highest frequencies of change occurred in the domains of personal characteristics (49% changed their hoped-for selves, 47% changed their feared images) and health (48% changed hoped-for selves, 43% changed feared images). In the domain of personal characteristics and life events there was a trend for more people to delete hopes and fears than to add them. In the domain of health, however, more people added (30%) than deleted (18%) hoped-for self images. The distribution across the categories of stability and change did not differ by age/cohort group.
In the context of hoped-for selves, the motivational orientation of improvement (to attain something) was consistently mentioned by 68% of BASE participants. Twelve percent added this orientation and 16% deleted it over time. More change was observed in mention of a maintenance orientation: Here 28% added this orientation to their hoped-for selves and 17% deleted it. These changes differed by age/cohort group. Compared with people over the age of 80 years, more people aged between 70 and 79 years added the maintenance orientation (37% vs 15%) and did not mention it (35% vs 16%), 2(3, N = 206) = 15.8, p < .001. The orientation of feared selves was highly stable over time.
We also examined the extent to which individuals mentioned matched hopes and fears in specific domains and the dynamics of these matches over time (see Table 3). As can be seen, few individuals consistently mentioned matched hopes and fears over the 4 years of the study in each domain. (This finding may appear to contradict the report above that 28% maintained a matched possible self, but recall that this was aggregated across domains.) The domains of health (11.7%) and personal characteristics (8.7%) showed the highest percentages of stable mention. In these domains, 53% and 59% of participants, respectively, never had matched possible selves (although they may have had separate hopes or fears), whereas others either deleted (15% and 22%) or added matches (20% and 11%) over time.
Links to Changes in Health, Life Events, and Well-Being
Possible selves are said to work within domains (e.g., Markus & Ruvolo, 1989) but there are also suggestions that separate aspects of the general structure of possible selves (e.g., number of hopes, fears, and matched possible selves) may also be linked to general well-being (e.g., Showers, Abramson, & Hogan, 1998). We ran several exploratory analyses to examine links between changes in well-being, health, and life contexts and changes in possible selves. To begin, we focused on the aggregated characteristics. Time-adjusted difference scores (change units per year) were calculated for number of matched possible selves (M = 0.02, SD = 0.25, range = -0.77–0.88), number of hoped-for selves (M = -0.01, SD = 0.39, range = -2.01–1.37), and number of feared self images (M = -0.1, SD = 0.35, range = -.85–1.03). Correlations between these change scores and the similarly constructed change scores for life satisfaction, satisfaction with aging, subjective health, and number of diagnosed illnesses were not significant. The possible selves difference scores were also not related to levels of well-being and health status at T1 or at T3. Aggregate changes in possible selves of the subgroup (n = 50) who had experienced a significant life event between measurements did not differ from the remaining sample.
At the aggregated level, we also examined whether subgroups who were stable (mentioned, not mentioned) or changed (added, deleted) their matched possible selves and motivational orientation toward self-improvement or self-maintenance differed in terms of changes in health and well-being. The four subgroups with different temporal patterns of matched possible selves and self-improvement motivations were not distinguished by different health or well-being changes. However, people (n = 56) who added a maintenance orientation to their hoped-for selves over time declined less in life satisfaction, F(3,200) = 5.09, p < .01: Indeed they increased their life satisfaction by an average of two t scores over the 4 years.
A second set of analyses examined the relationships within specific domains of possible selves. At the domain level, time-adjusted difference scores in feared selves were related to time-adjusted difference scores in well-being, health, and life contexts, but changes in hoped-for selves were not. Significant relationships were mostly in the domain of health: The more fears added about health, the greater the decline in life satisfaction (r = -.18, p < .01), aging satisfaction (r = -.17, p < .01), and subjective health (r = -.17, p < .01), and the more illnesses diagnosed at follow-up (r = .22, p < .01). At baseline, hopes about personal characteristics were positively correlated with subjective health (r = .15, p < .05) and negatively related to number of illnesses (r = -.15, p < .05). Hopes about health at baseline were negatively related to baseline subjective health (r = -.15, p < .05). These concurrent correlations were not significant at follow-up. Domain-specific fears at baseline and follow-up showed no relationship to concurrent levels of well-being or health. Baseline domain-specific hoped-for and feared selves were also not correlated with level of well-being or health at follow-up.
There were also links across domains: The more fears added about health, the more fears were deleted about personal characteristics (r = -.39, p < .01) over time. Deleting fears about personal characteristics was related to greater decline in life satisfaction (r = .15, p < .05). People who deleted a matched possible self in the domain of health showed less change in subjective health, F(3,200) = 5.09, p < .01, and life satisfaction, F(3,200) = 5.09, p < .01, and those who maintained matched possible selves about personal characteristics showed less change in life satisfaction, F(3,200) = 3.89, p < .01. Indeed, deleting a matched possible self in the domain of health and maintaining one about personal characteristics was associated with maintenance (and small improvements) in life satisfaction.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Before summarizing our conclusions, we want to mention four critical points. First, our longitudinal data are restricted to two measurements and to a 4-year interval. Essentially, we compare two discrete snapshots (rather than a single cross-sectional one) of a phenomenon that is portrayed to be highly dynamic and responsive to changes in life context. Four years may not be the appropriate interval for assessing the dynamics of possible selves. The period may be too long or too short to capture changes and their relationship to other factors. Future work should focus on the issue of selecting between-measurement time intervals.
Second, the possible selves collected on the two occasions in BASE most likely reflect only a selection of the future scenarios that older participants actually have in their repertoires. The instructions asked for only two hopes and two fears. Nevertheless, participants differed with regard to the number of possible selves they generated: some generated fewer, some more. The central finding was that 100% at baseline and 96% at follow-up were at least able to generate one future self image (either a hope or a fear). We do not know what the upper bounds of possible selves would be, or how the distribution of topics across domains would change, if the task did not specify a limit. Theory suggests that when selection is called for individuals report the most salient current concerns of their working self concept (Markus & Herzog, 1991). In this context, although the possible selves in our data possibly represent only a small part of a larger pool, they are likely a significant part.
Third, our examination of changes in possible selves is exploratory. The dimensions coded and our efforts to portray the varied patterns of change (both in relation to content domain and motivation) need to be refined. In particular, we believe that it will be important to examine conjoint changes in dimensions and specific thematic changes in order to observe the full complexity of the dynamic effects of possible selves.
Finally, the present study did not ask participants about their specific goal-directed behaviors over time. In future work on possible selves in old age it will be important to include a measure of behavior or action in analyses of links to antecedents and outcomes, to complement the correlational relationships reported to changes in health status and general well-being.
Given these limitations, what conclusions can be made? Whereas the central late-life task of dealing with prospects of declining health represented a focal domain in the possible selves of the majority of subjects, a wide range of other domains was also evident, especially those associated with identity (personal characteristics) and attachment (social relationships, positive contacts with family and friends). Hoped-for selves were not just expressions about avoiding undesirable outcomes or maintaining the current status: Expressions of desires to experience something new or experience something again predominated. As in younger age groups, the possible selves of many individuals also showed a degree of motivation for change.
Possible selves are a dynamic part of the self-concept in very old age. The large majority of BASE participants showed some aspects of change in their future possible selves. New hopes and fears were added; other aspects were lost over time. We also observed changes in the extent to which hopes and fears within a domain were matched and a progression toward adding matches in the domain of health. Possible selves are thought to be maximally motivationally significant when hopes in a domain are balanced by fears about that domain. Most of the participants reported at least one matched positive and negative possible self. In the present study, matched possible selves were operationalized as the presence of a hope and a fear in the same content domain. Subjectively, participants may not have perceived that their hope countervailed their fear in that domain. It is difficult to code positive and negative selves for equivalent connotation. For instance, the positive picture of oneself as an independent person might entail many dimensions and facets that are not addressed by the image of the dependent person one would not like to become. The absence of something negative does not necessarily imply its positive opposite. Future research should include a rating of equivalence (perceived match) of hopes and fears from the perspective of the individual.
Our operationalization of matched possible selves revealed low stability over 4 years. The changes involved both giving up a fear or a hope in a previously matched domain and arriving at new matches in the hopes and fears. Matches in the health domain were most stable. Contrary to our expectations, giving up a match in the domain of health was related to less decline in subjective health and life satisfaction. Being concerned about one's health in old age may actually be detrimental to well-being rather than beneficial. The maintenance of matched hopes and fears about health, for example, might reflect chronic worry and rumination rather than a motivational force. This interpretation is analogous to a report by Niedenthal, Setterlund, and Wherry (1992). They suggested that a large number of highly interrelated possible selves (matched hopes and fears can be viewed as a special case of interrelated possible selves) can have detrimental affective consequences when individuals are confronted with negative events. Similarly, the compartmentalization of positive and negative self-images (and especially an increased focus on positive images) in the context of stressful life changes has been found to minimize the impact of negative life events on well-being (Showers et al., 1998).
The significance of these observations, however, needs to be compared with data from younger age groups. Research with adolescents on possible selves in the academic domain, for example, has indicated that few young people consistently include matched hopes and fears in their mentioned future selves (e.g., Fraser & Eccles, 1995; Oyserman & Markus, 1990). Having a variety of domains reflected in one's hopes and fears might represent a broad motivational basis that is less vulnerable to losses. Consistent with this view, we found that individuals who added matched possible selves related to personal characteristics, who added two or more new domains of hope, or who maintained motives to strive for something had a higher likelihood of increasing their life satisfaction over time.
The motivational orientation of old and very old persons expressed in their possible selves was related to well-being. A strategy of becoming more oriented toward maintenance was linked to less decline in life satisfaction over time. This does not mean, however, that most older people were oriented toward maintenance. To the contrary and against our expectations, most hopes reflected approach motivation (self-improvement). whereas most fears reflected avoidance motivation. This finding might be partly due to the specific instructions used in this study ("what would you like to become" for hopes, and "what sort of a person would you not like to become" for fears).
The age and time-related changes in the content and structure of possible selves described above may be associated with individual differences in life context, health, and well-being. Predictions about the specifics of these relationships are difficult to make, however, and for this reason we reported exploratory analyses. Conceptually, declining health, loss of a partner, or a housing relocation, for example, could lead to casting away possible selves, to the construction of matched hopes and fears, or to increased compartmentalization of hopes and fears. At least with regard to declining health, we found that it enhanced the likelihood of developing fears in the domain of health. Conversely, reducing the number of matched possible selves in the domain of health was related to the maintenance of subjective health. We consider this evidence for the context-relatedness of possible selves as suggested by Cross and Markus (1991).
Which characteristics (content and structure) of possible selves are linked to well-being in late life? Cross and Markus (1991) found that at all ages persons who reported low life satisfaction generated more hopes about personal characteristics (i.e., indicating a desire to change present self characteristics), compared with those high in life satisfaction who generated more hopes in the occupational, family, and health domains. Within the present sample of older adults, we found that hopes were less related to well-being than were feared self images. Developing and maintaining balanced possible selves, particularly in old age, might also serve as a resource for well-being.
A working self-concept that includes future guides for bolstering self-esteem and a sense of mastery appears to be an important aspect of the self in later life. Whereas for some life domains, esteem can be gained from past feats (e.g., interests, family), others require continuous input ("I am still a nice person and accepted by others"). Other domains probably acquire increased salience in late life (e.g., health), particularly for those individuals who suffer from debilitating illnesses or lack well-functioning support networks. The dynamic interplay of possible self content in relation to new challenges associated with changes in life circumstances and the relationship between current self-descriptions and future scenarios remains to be investigated. Such studies will further the, as of yet limited, understanding of the role of future-oriented motivational systems in the maintenance of well-being in very old age.
|
Acknowledgments |
|---|
Received for publication July 31, 2001. Accepted for publication July 18, 2002.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. A. Hoppmann, D. Gerstorf, J. Smith, and P. L. Klumb Linking Possible Selves and Behavior: Do Domain-Specific Hopes and Fears Translate Into Daily Activities in Very Old Age? J. Gerontol. B. Psychol. Sci. Soc. Sci., March 1, 2007; 62(2): P104 - P111. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Hooker and D. P. McAdams Personality Reconsidered: A New Agenda for Aging Research J. Gerontol. B. Psychol. Sci. Soc. Sci., November 1, 2003; 58(6): P296 - 304. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Smith, I. Maas, K. U. Mayer, H. Helmchen, E. Steinhagen-Thiessen, and P. B. Baltes Two-Wave Longitudinal Findings From the Berlin Aging Study: Introduction to a Collection of Articles J. Gerontol. B. Psychol. Sci. Soc. Sci., November 1, 2002; 57(6): P471 - 473. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
