Journal of Gerontology: PSYCHOLOGICAL SCIENCES
1999, Vol. 54B, No. 1, P55-P67
Copyright 1999 by The Gerontological Society of America
Content and Function of the Self-Definition
in Old and Very Old Age
Alexandra M. Freund and Jacqui Smith
Max Planck Institute for Human Development, Berlin, Germany.
Spontaneous self-definition was investigated in a heterogeneous sample o/N = 516 participants of the Berlin Aging
Study, aged between 70 and 103 years. The content of the self-definition revealed that old and very old persons view
themselves as active and present-oriented. The self-definition also reflected an inward orientation, and central
themes of life-review, health, and family. Participants generated more positive than negative evaluations in their
self-definition, but the ratio of positive to negative evaluations was less favorable for the oldest old (> 85 years) than
that of persons aged 70 to 84 years. Older individuals with more health-related constraints reported fewer and less
rich self-defining domains (i.e., a less multifaceted self-definition). Positive emotional well-being was associated
with naming more and richer self-defining domains. Multifacetness, however, did not buffer against the negative effect of low functional capacity on subjective well-being.
T
HIS study examines three aspects of older persons'
self-definitions, namely content, evaluative valence,
and multifacetness. Whereas an analysis of the content of
the self-definition provides insight into the salient life
themes at a given point in a person's life, the evaluation
and the multifacetness of the self-definition are thought to
be related to emotional well-being. To date, the content,
structure, and function of self-definition have been studied
mainly in the field of social psychology (e.g., Epstein,
1990; Greenwald & Pratkanis, 1984; Kihlstrom & Cantor,
1984; Markus, 1983; Swann, 1990). Within developmental
psychology, research has focused on the question of how
children acquire a sense of self as a separate entity (e.g.,
Harter, 1983). Whereas the content and function of the selfdefinition of middle-aged and young-old adults has been
investigated (e.g., Hooker, 1992; Ogilvie, 1987; Thoits,
1983), we know little about the self-definition of the oldest
old (i.e., 85 to 100+ years of age).
The self-definition refers to that part of self-related
knowledge that contains attributes crucial for the definition
of oneself. It involves the subjectively most important
characteristics to which individuals feel committed and
which subjectively distinguish their own person from that
of others (Brandtstadter, 1985; Brandtstadter & Greve,
1994; Cantor & Kihlstrom, 1989; Markus, 1983; McGuire,
1984; Wicklund & Gollwitzer, 1982). In the terminology of
Kelly (1955) and later Epstein (1973, 1990), this part of
self-related knowledge contains the central aspects of an
individual's subjective theory regarding his or her own self,
behavior, goals, values, interests, or personal characteristics
(McAdams, 1996). Accordingly, the self-definition does
not refer to an "objective" description of a person, but
rather to those self-conceptions which individuals believe
to be essential for the definition of their own person. Thus,
the self-definition can be viewed as a window to the central
issues and life domains in which a person is involved (Epstein, 1973; Filipp, 1980; James, 1890).
Content of the Self-Definition in Very Old Age
In very old age, themes that were relatively unimportant
in young and middle adulthood gain in salience (P. B.
Baltes, 1987; 1997; Breytspraak, 1994; Erikson, 1982;
Havighurst, 1963; Neugarten, 1964; Thomae, 1979; 1987).
Erikson (1959), for example, proposed that old age is characterized by a tension between integrity and despair, and
that this tension manifests itself by reflections about the
past and thoughts about death. Peck (1968) added three
themes to this description: (a) coping with physical decline,
(b) loss of professional status, and (c) the process of dying.
L'Ecuyer (1990) suggested that the reorganization of the
self-definition is the major task between ages 60 and 77.
Similarly, Waterman and Archer (1990) argued that in late
adulthood, due to a decrease in professional and family
roles and obligations, there is increased room for new selfdefinitions in such domains as hobbies and personal interests. After age 80, according to L'Ecuyer, the central task is
to conclude the process of self-definition in the form of a
complete Gestalt ("oeuvre"), by means of a life review and
the integration of the different aspects of the self-definition
(Waterman & Archer, 1990). Thus, the self-definition of
very old persons may rely on experiences, achievements,
and events that happened in the past, rather than in the present or the future. Also, the themes "life review" and
"death and dying" should have a particularly high salience
for self-definition in old age.
As Peck (1968) proposed, body image (i.e., physical attractiveness and appearance) should be less important for
defining oneself in old age, in comparison to young adulthood, because physical decline and approaching one's
death might lead to the transcendence of the importance of
good looks. Similarly, financial concerns might be unimportant for the self-definition of older persons because materialistic belongings are likely to lose importance when
one's future is limited and one has to face death. In addition, finances might be less of an issue of concern for GerP55
P56
FREUND AND SMITH
man elderly people because, as a function of the welfare
system, their pensions are relatively fixed.
The domains of health, everyday competence and independent living, and daily living routine should also be very
important to the self-definition in advanced old age, because they are much more vulnerable to change than in previous life stages (Atchley, 1991; M. M. Baltes, Maas,
Wilms, Borchelt, & Little, in press; M. M. Baltes & Silverberg, 1994; Whitbourne, 1985). Empirical evidence for the
salience of health in older adults' self-concept stems from
Hooker's (1992) work on possible selves, which showed
that, compared with college-students, older adults aged 52
to 89 years were more likely to report possible selves in the
domain of health.
Outdoor activities which require physical mobility are
likely to be of minor importance for old and very old people, because they are at a high risk for mobility-related constraints (Manton, Stallard, & Corder, 1995). On the other
hand, hobbies and interests that do not require physical mobility should be more salient for older individuals. After retirement, they have ample time to invest in this life domain.
Based on Carstensen's Socioemotional Selectivity Theory (SST, Carstensen, 1993), it can be expected that family
represents an important area for older persons' self-definitions. According to SST, people are increasingly selective
with regard to their social partners when approaching the
end of life. Whereas social partners become less important
for gaining information with increasing age, emotion regulation becomes an increasingly important motive for social
contact in old age. Family members and relatives, usually
long-standing members of one's social convoy (Antonucci,
1990), might have particular emotional salience in old age
as the number of same-aged friends decreases because of
their health-related problems, reduced mobility, or because
they have died.
There are very few studies that include substantial numbers of individuals older than 85 years, and use open-ended
measures (rather than questionnaires or adjective checklists) to access central self-defining domains (one exception
is Johnson & Barer, 1997). The few investigations of selfdefinition that have included individuals over the age of 60
tend to confirm our expectations, based on theoretical considerations, concerning the saliency of certain themes and
life domains for this age group, but the results are somewhat inconsistent. Dittmann-Kohli (1995), for example,
found that previous life events and life history, as well as
physical integrity and functional capacity versus physical
decline and its consequences, represent essential aspects of
the self-definition of adults between the ages of 60 and 90
years compared with younger adults (15-34 years). McCrae
and Costa (1988) found age-related differences in the selfdefinition of persons between 32 and 84 years of age: With
increasing age, themes such as age, health, life events, life
situations, hobbies/interests, and attitudes were mentioned
more frequently, whereas the frequency of the domains
family roles, social relationships, neuroticism, personality
traits, and daily living routines decreased. George and
Okun (1985), however, found no significant age-related differences in their cross-sectional comparison of the content
of the self-definition of three age groups (45-54, 55-64,
and over 65 years). Likewise, Filipp and Klauer (1986), reported no age differences in a cross-sectional comparison
of five male cohorts (born between 1905 and 1945). A longitudinal analysis of the changes in self-definitions over 26
months, however, showed that themes such as social roles,
political or religious attitudes, body image, and social style
were increasingly reported to be self-defining, whereas
emotionality and autonomy were mentioned less frequently.
Functional Characteristics of the Self-Definition:
Evaluation and Multifacetness
Two aspects of the self-definition are considered to be
essential for a positive self-view, and for maintaining continuity of the self-definition: (a) the personal evaluation of
self-defining domains as positive or negative, and (b) the
multifacetness of the self-definition. In order to empirically
investigate their proposed function, both dimensions of the
self-definition (evaluation and multifacetness) were included in the present study and related to an independent
measure of emotional well-being.
Several authors have repeatedly stressed that positive or
negative evaluation of self-defining aspects is very important for a general sense of self-worth and emotional wellbeing (Epstein, 1990; Pelham & Swann, 1989; Marsh,
1987; Showers, 1992; Showers & Ryff, 1996). Research on
a global sense of self-worth has revealed few age-related
differences (cf., Bengtson, Reedy, & Gordon, 1985; George,
1981; Herzog & Rodgers, 1981). Whether and how old and
very old adults evaluate themselves in their spontaneous
self-definition is not known. This was one goal of the present study. Further, we were interested in whether these
evaluations are related to emotional well-being in advanced
age. We expected that, in late adulthood, the increasingly
negative gain-loss-balance (Heckhausen, Dixon, & P. B.
Baltes, 1989) would be reflected in a positive correlation
between age and the number of negatively evaluated selfdefining domains, and a negative correlation between age
and the number of positively evaluated domains.
Multifacetness refers to the number of self-defining domains and their richness (i.e., the number of aspects within
self-defining domains), and serves to guarantee continuity
of the self-definition (Freund, 1995). A multifaceted selfdefinition (i.e., a self-definition that is comprised of a large
number of different domains that contain many aspects)
provides alternative sources for a self-definition and offers
more possibilities to compensate for loss and impairment,
and, therefore, is likely to favor continuity (Linville, 1987;
Rosenberg & Gara, 1985; Thoits, 1983; Ogilvie, 1987). For
example, with regard to richness, if the self-definition of a
person in the domain "hobbies" contains only the aspect "I
play tennis," this self-defining domain cannot be maintained,
if playing tennis is impossible due to physical constraints. If,
however, the domain "hobbies," apart from playing tennis,
contains additional aspects such as coach, referee, advisor, or
spectator, as well as watching movies, then the self-defining
domain "hobbies" can be maintained. Not only is tennis
likely to continue as a field of interest if physical constraints
limit active participation, but there is the additional aspect of
enjoying movies. We therefore assumed that the higher the
degree of richness of a self-defining domain, that is, the more
SELF-DEFINITION IN OLD AGE
aspects it involves, the less vulnerable it should be to the
losses and health constraints of old age.
Similarly, a large number of self-defining domains is expected to provide more continuity because there is a greater
supply of alternative or substitute domains which come into
play when one is lost. For instance, if one out of an original
ten domains is lost, there are still nine left, which means that
the self-definition is hardly impaired. The same loss would
be severe if the self-definition had originally been based on
only two domains. This function of maintaining continuity
is, of course, best investigated in a longitudinal study.
Given the increased likelihood of losses and of healthrelated constraints in old age (e.g., Manton, Stallard, &
Corder, 1995), the stability of the self-definition is especially at risk for elderly people. The maintenance of the
self-definition is often posited as an important task in old
age, essential for the well-being of an individual (e.g., P. B.
Baltes & Baltes, 1990; Brandtstadter & Greve, 1994; Troll
& Skaff, 1997). Many self theorists have pointed out the
importance of maintaining one's self-definition (e.g., Allport, 1961; Epstein, 1990; Lecky, 1945; Swann, 1990). Epstein (1973), and later Swann (1990), suggested that the
main function of the self-definition is to structure and organize experience. According to this view, the self-definition
offers concepts, which guide behavior, provide orientation
for interacting with the environment, and provide a sense of
identity (Allport, 1961). Symbolic interactionists have long
stressed the importance of social partners for self-related
knowledge, as well as the important function of the selfdefinition for communicating to social partners who you are
and how you are likely to behave (e.g., Cooley, 1902;
Mead, 1925). More recently, Stryker (1987) stressed the
importance of personal commitment to social roles (identities). According to Stryker, identities function to guide expectations, and to orient behavior.
If a multifaceted self-definition (i. e., a self-definition
composed of a large number of rich domains) serves as a
source of alternative self-defining domains and aspects, it
might contribute to emotional well-being in advanced old
age, even in the face of poor health and physical constraints. Staudinger, Freund, Linden, and Maas (in press),
for example, reported that a multifaceted self-definition reduced the vulnerability for depression in connection with
physical impairment. To date, however, it has not been
shown whether multifacetness has an impact on the positive
aspects of subjective well-being.
Central Questions of the Present Study
The present study asked five central questions about the
content, evaluative valence, and multifacetness of the selfdefinition of individuals aged between 70 and 100+ years.
(1) Which domains are represented in older persons' selfdefinitions? We assumed that the content of the selfdefinition would reflect the themes salient for this age
group. We expected the following domains to be particularly salient for old (79 to 84 years) and very old (85
to 100+ years) persons: health, everyday competence,
daily living routines, interests / hobbies, family, lifebalance / life-review, and death / dying. Furthermore,
(2)
(3)
(4)
(5)
P57
we expected that the self-definition would be rooted in
past rather than in present or future events, achievements, and experiences. Professional status may also
play a role in older persons' self-definition, because
professional status carries personal as well as societal
meaning that goes beyond retirement. Finally, we hypothesized that body image, financial considerations,
and outdoor activities would be domains that infrequently characterize older persons' self-definition. In
the present study, the salience of a given life-domain
for older adults' self-definition was operationalized as
the number of participants nominating the respective
life-domain as self-defining.
Is the content of the self-definition different for old and
very old persons? We expected that more very old persons (85-100+ yrs) than old persons (70-84 yrs) would
name the domains of health, everyday competence,
daily living routines, interests / hobbies, family, life-balance / life-review, and death / dying as self-defining. In
addition, we hypothesized that very old adults would include statements about body image, financial considerations, and outdoor activities less frequently in their selfdefinition than persons aged between 70 to 84 years.
Do old and very old persons differ in the evaluation of
their self-defining domains?
Are there differences between old and very old persons
with regard to the multifacetness of their self-definitions? We expected that the number of different selfdefining domains, as well as their richness, decrease
due to the increasing constraints and losses experienced
in advanced old age.
Do individuals who produce a multifaceted self-definition or include positive evaluations also report higher
emotional well-being?
METHODS
Data are reported from the first measurement occasion
(1990-93) of the Berlin Aging Study (BASE: P. B. Baltes,
Mayer, Helmchen, & Steinhagen-Thiessen, in press; P. B.
Baltes & Smith, 1997). In this article, we only refer to those
aspects of BASE which are relevant for this study. The research design, sample selection, and procedure of BASE
are described in detail elsewhere (see P. B. Baltes et al., in
press; Lindenberger, Gilberg, Potter, Little, & P. B. Baltes,
in press; P. B. Baltes & Smith, 1997).
Sample
The present study is based on the first measurement sample of N = 516 participants in the Berlin Aging Study (for
details see P. B. Baltes & Smith, 1997). The heterogeneous
age-by-sex stratified sample, which was locally representative of the western districts of Berlin, included 258 women
and 258 men aged between 70 and 103 years. The sample
was divided into 6 age/cohort groups (each n = 86): 70-74
years (born 1915-1922), 75-79 years (born 1910-1917),
80-84 years (born 1905-1913), 85-89 years (born 19001908), 90-94 years (born 1896-1902), and 95-103 years
(born 1883-1897). Selectivity analyses (Lindenberger et
al., in press) indicated that the sample was positively selected. However, selection effects were relatively small and
P58
FREUND AND SMITH
did not affect in any major way the variance and covariance
structure of the major constructs of BASE.
The average age of the BASE sample at the first measurement occasion was 85 years. At the time of the study, 30% of
the sample were married, 55% widowed, 7% divorced, and
8% had never married. Eighty six percent of the BASE participants resided in their own homes and 14% in institutions
(e.g., homes for seniors, nursing homes, hospitals). Representative for these cohorts in Germany, 65% of the sample
had primary level education, 28% secondary level, and 8%
tertiary level. The percentage of non-German participants in
the BASE sample was very low (4.5%), matching the population distribution for over 70-year-olds. During the 12month period after study participation, 5.6% of the sample
had died, a mortality value similar to age-mortality statistics
for this age range and its stratified sample composition.
Measures
Self-definition.—The self-definition was assessed using
an oral version of the open question "Who am I?" (Bugental & Zelen, 1950; Kuhn & McPartland, 1954). This openended measure seemed to be especially appropriate for assessing the content of the self-definitions of old and very
old individuals, as there is practically no reliable empirical
knowledge about the salient life-domains of this age group
(cf. Filipp & Klauer, 1986; for a detailed theoretical discussion of using open-ended measures for assessing the
self-definition see Freund, 1995). Subjects were asked to
generate ten answers to the question "Who am I?". These
answers were taped and transcribed verbatim. Two examples of protocols are provided in Table 1.
There were two main steps to the procedure used to code
the answers to the question "Who am I?". In the first step,
"Who am I?" protocols were segmented into meaningful
idea units defined on the basis of a coding system consisting of 24 content domains (see Table 2 for a list) that had
been developed specifically for assessing self-definition of
old and very old people (Freund, 1995; Freund & Smith, in
press). A segment was defined as beginning with a new thematic content category and ending with the start of a different category (Krippendorf, 1980). Agreement between two
independent persons segmenting the protocols was high (M
= 95.1% and M = 86.8%, computed for a sample of n = 156;
agreement was calculated against "target" segmented protocols completed by a third person).
In a second step, idea units (segments) were coded as reflecting one of the 24 content categories (listed in Table 2)
Table 1. Two Examples of "Who am I?" Protocols Obtained From BASE Particpants
Self-defining statement (idea unit)
Domain
Richness
Example 1: 100 year old man
I am a male / who is 100 years old
Demographic variables
During my life I have been to many places in the world and have seen a lot of things
Life review
Now, I am living here in Berlin
Demographic variables
I've been on the pension for more than 40 years
Professional occupation
My hobby was furniture making and I made a lot of my present furniture myself
Hobbies/interests at home
I traveled with my wife every year. But not as far away as it is usual nowadays.
Hobbies/interests outdoors
I have had a lot of friends in my life of whom, unfortunately, only a few are still alive. /
But I am happy to know that I still have very close relationships to friends my age
as far as they are still there.
Social relations
For the last 20 years I've not been able to read or write anymore, and that makes life
really difficult. /I'm not able to watch TV anymore.
Everyday competence
Since my wife died not long ago I feel very lonely
Emotions
To divert myself I still have the radio.
Hobbies/interests at home
Example 2: 89 year old woman
I do not do a lot of things. /1 tidy up my apartment.
Politics interest me. /And how it [the reunification of Germany] will all work out. /
I read the newspaper.
I feel lonely.
I play cards and solitaire. /1 keep myself busy looking after my flowers.
I am unselfish.
I can hardly hear anymore.
I'm the grandmother of some dear grandchildren
I've come to believe that to give is more blessed than to receive
Financially I have enough money to live on
In a way, I think my life was worthless.
Daily living routine
Interests reflecting participation (at home)
Emotions
Interests/hobbies at home
Personality traits
Everyday competence
Family
Life knowledge
Financial considerations
Life review
Note. Self-defining statements (segmented idea units) were coded for content (24 categories), and for three dimensions (richness, evaluation, time). The
richness of a self-defining statement (i.e., number of aspects mentioned within each content category) is indicated by sections separated with slashes.
SELF-DEFINITION IN OLD AGE
Table 2. Category System for "Who am I?"
Example
Category
1. Sociodemographic variables
a) Gender
b) Name
c) Nationality
d) Hometown
e) Age
f) Marital status
g) Place of living
2. Financial / material
considerations
Well, of course, I'm a woman.
First my name: H. G.
I am German
Somehow, I stayed in Berlin.
I am old - being 82 years old, that's
old.
I have been a widow for the last 12
years.
I like my apartment a lot, especially its balcony.
I own three houses, all of them in
Berlin.
3. Personality traits
I am very self-conscious.
4. Likes and dislikes
Personally, Ido not like these
bright colors.
5. Emotions / mood
Unfortunately, I get angry easily.
6. Body image
Many people say I'm still attractive
7. Health
I have a lot of pain in my legs
when I walk
I am often forgetting things
8. Intellectual functioning
9. Profession / profession-like
activities
10. Interests / hobbies at home
11. Social participation (at home)
12. Interests / hobbies outdoors
13. Social participation (outdoors)
14. Daily living routines
15. Everyday competence
16. Life events
17. Social relations (not family)
18. Family
19. Interpersonal style
20. Life knowledge
21. Life balance / life review
22. Death and dying
23. Religious beliefs / attitudes
24. (Socio-) political beliefs /
attitudes
I am a trained car mechanic
I really love to watch sports on TV
I read a lot about foreign cultures
I used to travel a lot
I often go to the theater and still
have a theater subscription.
The first thing I do in the morning
is to make a nice breakfast
I can't even wash myself without
help
The war changed my whole life
I have a lot of friends.
I am mother, grandmother, and
aunt.
I can't hold a grudge against anybody for long
It is best to live your life as if every
day was your last
I've lived a satisfying and fulfilling
life.
I am not afraid of dying any more
I believe that there exists a good
principle guiding everything
The present political parties are
totally ineffective
by two independent raters, using a computer-based randomized procedure in order to avoid context effects. The coding
scheme consisted of 24 categories for the content of selfdefinitions (see Table 2). The agreement between the two
coders for these categories was satisfactory—again computed for a sample of n = 156 Kappa K = .85 (Brennan &
Prediger, 1981).
P59
Segments were also coded in separate rounds with respect to three dimensions: (a) Richness: How many different aspects were mentioned to elaborate a self-defining domain, (b) Time reference (past, present, or future); and (c)
Explicit, personal evaluation: A self-defining statement
may be simply descriptive (neutral), though it may also
convey an explicit positive or negative evaluation. Statements that included an explicit evaluation (e.g., "I am a
very good cook," or "It's so sad that I cannot get around in
the city anymore") were coded as positive or negative depending on the words used. All other statements were
coded as neutral (e.g., "I am interested in politics."). These
neutral statements may have implied desirable or less desirable characteristics, but, for the present purpose, we decided only to consider evaluations explicitly made by the
older person. This coding strategy is more objective than
one which would rely on the value judgments of coders as
to whether a statement was positive or negative. For the
three dimensions, agreement between two coders was satis- •
factory (computed for a sample of n = 156, K = .82 for richness, K = .84 for time reference, and K = .85 for personal
evaluation).
The idea-units (segments) in the transcribed protocols
were coded and assigned to one of the 24 content categories
as well as rated along the three dimensions according to the
explicit statements contained in the self-definitions, avoiding, as far as possible, any interpretation on the part of the
raters. The coding-system was adapted from previous coding schemes in the literature. Specifically, our coding system adapted and expanded the systems developed by Cross
and Markus (1991), Epstein (1973), Filipp and Klauer
(1986), George and Okun (1985), Gordon (1968), and
Kuhn and McPartland (1954) (for a more detailed discussion of how the coding system was derived, see Freund,
1995).
Measures of functional health constraints.—Physical and
sensory constraints were operationalized as a unit-weighted
standardized composite of indicators for mobility (self-reported maximal walking distance [7-point scale], motor
balance as assessed by the Romberg Test [keeping balance
while standing with closed eyes and arms outstretched],
and the number of steps needed to do a full 360 degree circle), near and distance visual acuity (using standardized
reading tables, measured in Snellen Decimals), audition
(assessed with a Bosch ST-20-1 audiometer measuring pure
tones with frequencies ranging from 0.25 to 8.00), and selfreported independence with respect to everyday competence (ADL, Katz, Down, Cash, & Grotz, 1970). Details regarding the various assessment procedures are described
in Steinhagen-Thiessen and Borchelt (in press; see also
Marsiske, Klumb, & M. M. Baltes, 1997; Lindenberger &
Baltes, 1997).
Emotional well-being.—Positive and negative emotions
were assessed by means of a checklist, containing 10 positive and 10 negative "emotional" adjectives (Watson, Clark,
& Tellegen: PANAS, 1988). Subjects were asked to indicate on a 5-point scale how often they had experienced
each emotion in the course of the last year (for detailed in-
P60
FREUND AND SMITH
formation, see Smith, Fleeson, Settersten, & Kunzmann, in
press). Mean values were computed for the 10 positive and
10 negative items (Cronbach's Alpha for positive emotions:
a = .78, for negative emotions: a = .81). In accordance
with Watson, Clark, and Tellegen (1988), positive and negative emotions were not correlated in the Berlin Aging
Study (r = .05).
Subjective well-being was assessed using a 15-item German translation of the Philadelphia Center Morale Scale
(PCGMS, Lawton, 1975; Liang & Bollen, 1983). Participants indicated on a five-point Likert-scale how much they
agreed with each item (1 = does not apply to me at all, 5 =
applies very well to me; Cronbach's Alpha a = .83).
Procedure
All testing was carried out in the subject's place of residence (e.g., private home, nursing home, hospital) by
trained research assistants. The "Who am I?" was collected
as the first item in a 90-minute interview session about self
and personality (see Smith & Baltes, in press, for details
about this session). The PANAS was included later in the
same session. The interview took place in individual, face
to face sessions. All responses were spoken by the participants and recorded by the interviewer (as well as taperecorded). Health-measures were obtained in a different
session by physicians.
RESULTS
All participants (TV = 516) were able to produce some response to the question "Who am I?". Furthermore, 86.7%
of the protocols contained at least 10 self-defining statements that could be coded according to the 24 content categories. Only 4.2% of the sample generated fewer than five
statements that could be coded as falling into one of the 24
categories. Interrater agreement, computed for a subsample
of n = 156, was Kappa K = .85 for content categories, K =
.82 for richness, K = .85 for personal evaluation, and K =
.84 for time reference (Kappa computed according to Brennan & Prediger, 1981). The rating procedure proved to be
sufficiently reliable. In sum, the "Who am I?" proved to be
a feasible method for assessing spontaneous self-definitions
in a sample including the oldest olds.
Self-Defining Domains
Figure 1 shows for the 24 content domains in the coding
system (Table 2), the percentage of participants who reported
a given domain as being self-defining at least once. The frequency distribution indicates that the self-definition of old
and very old persons comprised a broad spectrum of themes
and life-domains. As Figure 1 shows, even in advanced old
age an activity-oriented lifestyle was predominant. The majority of the sample considered hobbies and interests which
can be pursued at home (general interests: 61%; interests re-
Interests/hobbies at home
Daily living routine
Health
Family
Interests: social participation at home
Personality traits
Life review
Interpersonal style
Demographic variables
Interests/hobbies outdoors
Religious/existential attitudes
Professional occupation
Social relations (not family)
Everyday competence
Life knowledge
Cognition
Emotions
Interests: social participation outdoors
Sociopolitical attitudes
Financial considerations
Likes and dislikes
Thoughts about own death
Body image
Life events
10
30
40
50
60
70
% Participants
Figure 1. Distribution of self-defining domains (24 content categories) reported by participants (N = 516) in the Berlin Aging Study. Illustrated is the
percentage of participants who named a domain as self-defining at least once.
SELF-DEFINITION IN OLD AGE
fleeting social participation: 47%) as well as daily living routine (48%) as self-defining. Also as predicted, family proved
to be one of the most salient domains for older persons' selfdefinition (rank 4 of the distribution). We explored the possibility as to whether naming family or other social partners
depended on marital status. Marital status (married: n = 154
vs single/divorced/widowed: n = 362) did not affect the frequency with which participants mentioned family as self-describing. Marital status, however, did affect the frequency
with which participants mentioned other social partners
(friends, acquaintances) in their self-definition. Participants
who were not married (single, divorced, or widowed) were
significantly more likely to name friends/acquaintances than
married participants (x2 = 17.8,/? < .001).
Contrary to expectations, thoughts about death and dying
played a rather insignificant role, and constituted only a
P61
very small part of the self-definition (5%). At the same
time, the self-definition of older persons reflected the expected inwards orientation (life review: 42%) and preoccupation with one's own person (personal characteristics:
45%). Likewise, health played a central role for the selfdefinition of old and very old individuals (rank 3 of the frequency distribution). Everyday competence, however, was
reported as being self-defining by only one third of the
sample (30%). Old and very old persons also defined themselves as more active than expected: 41% of the sample
mentioned hobbies and interests away from home.
Age-Related Differences in the Content
of the Self-Definition
Figure 2 shows separate profiles of the content of the
self-definition for the age groups 70 to 84 years (old) and
Interests/hobbies at home
Family
Interests/hobbies outdoors
Interpersonal style
Interests: social participation at home
Personality traits
Health
Daily living routine
Life review
Religious/existential attitudes
Professional occupation
Social relations (not family)
Demographic variables
Life knowledge
Emotions
Cognition
Everyday competence
Interests: social participation outdoors
Likes and dislikes
Financial considerations
Sociopolitical attitudes
Body image
Life events
Thoughts about own death
Old (70 - 84 years, n = 258)
Very old (85 - 103 years, n = 258)
30
40
% Participants
Figure 2. Content of the self-definition of old (70-84 years) compared with very old (85-103 years) participants of BASE. Illustrated is the percentage
of participants who named a domain as self-defining at least once.
P62
FREUND AND SMITH
85 to 103 years (very old). Chi-square tests were used to investigate age-related differences with respect to the frequency of reported self-defining domains. Due to the large
number of computed frequency comparisons, the alpha
level was a priori set to a maximum of 1%. Compared to
old persons, very old individuals defined themselves less
frequently in the domains family/relatives (x2 = 7.97, p <
.01), interests and hobbies outdoors (general interests: x2 =
21.12, p < .001), interests reflecting social participation
(X2 = 3.96, p =.05), and interpersonal style (x2 = 9.25, p <
.01), and more frequently in the domains daily living routine (x2 = 8.96, p < .01) and sociodemographic variables (x2
= 23.42, p < .001). Contrary to our expectations, there was
no empirical evidence for age differences in the domains
hobbies and interests at home and life review, and only
marginally significant age differences for the domains
health (x2 = 4.11. p < .05) and thoughts about death and
dying (x2 = 2.98, p = .08).
The number of self-defining statements referring to the
present, the past, or the future, was analyzed in order to test
the hypothesis that the past plays an important role for the
self-definition in advanced old age. The majority of the
self-defining statements referred to the present (M = 8.7 of
10 statements), only a very small amount to the past (M =
.57), and the future was hardly mentioned (M = .14). This
suggests that the self-definition in advanced old age, in general, focuses on the present. There was, however, a significant difference between the two age groups with regard to
references to the past: Individuals aged 85 to 103 years referred more often to the past (M = .69) than persons aged
70 to 84 years (M = .45; r[514] = -2.40, p < .05).
Evaluative Valence of the Self-Definition
The spontaneous self-definition of old and very old individuals was mainly composed of descriptive, evaluatively
neutral statements. Out of the ten reported self-defining
items, an average of M = 6.4 contained no explicit evaluation. Only M = 1.3 statements contained an explicitly negative evaluation, and M = 2.2 an explicitly positive one.
A comparison of mean values revealed an age difference
with respect to the evaluation of self-defining domains:
Very old individuals mentioned significantly more negatively evaluated domains (/[514] = -3.84; p < .001) and significantly less positively evaluated (f[514] = 2.2, p < .05)
and purely descriptive ones (t[5l4] = 2.24, p < .05). Although even the oldest group (85+ years) reported more
positive than negative evaluations, the proportion of positively to negatively evaluated self-defining domains was
more unfavorable in advanced old age. These differences
are illustrated in Figure 3.
The evaluation of the self-definition was significantly related to emotional well-being: The number of positively
evaluated self-definitions correlated with positive emotions
(r = .16, p < .01), whereas the number of negative evaluations correlated with both positive (r = -.18, p < .01) and
negative emotions (r = .28, p < .01). These correlations remained significant even after statistical controls for chronological age and physical/sensory constraints (partial correlation for positive evaluation and positive emotions: r = .15,
p < .001; for negative evaluation and [a] positive emotions:
r = -.11, p < .01, [b] negative emotions: r = .29, p < .001).
Positive (r = . 12, p < .05) but, particularly, negative evaluation of the self-definition (r = -.42, p < .01) were significantly related to the PCGMS score of subjective well-being.
Again, the correlations remained significant after controlling for chronological age and the unit-weighted composite
measure of physical/sensory constraints (partial correlation
for positive evaluation and well-being: r = .12, p < .05; for
negative evaluation and well-being: r = -.42, p < .01).
Multifacetness
On the average, the self-definitions of old and very old
individuals included M = 6.8 (SD = 1.4) different domains.
These domains contained a mean of M = 2.3 aspects (richness). Number of domains and richness showed a low but
statistically significant correlation (r = .10, p < .05). While
the sheer number of self-defining domains was not age-related (r = -.05, n.s.), there was a small negative age-correlation for richness (r = -.10, p < .05). This relationship can
be explained by physical and sensory constraints, as the
negative age-correlation disappeared after a statistical control for physical and sensory constraints (partial correlation
r = .01, n.s.). However, both aspects of multifacetness revealed low but significant negative correlations with the
unit-weighted composite of health-related constraints (for
number of domains: r = -.09, p = < .05; for richness: r =
-.16, p < .01), which remained significant even when age
was controlled for (for number of domains: partial correlation r = -.0S,p < .05; for richness: r = -A3,p < .01).
Both the number of self-defining domains (r = .11, p <
.05) and richness (r = .15, p < .01) showed a small but statistically significant correlation with positive emotions, but
no relationship with negative emotions or subjective wellbeing. The hypothesis that multifacetness buffers the negative effect of physical and sensory constraints on emotional
well-being (r = -.30, p < .01) was not supported. Hierarchical regression analysis did not reveal any evidence for an
7n
6- •
CO
^
C
^~
i4
•
Old (70 - 84 years, n = 258)
Very old (85 -103 years, n = 258)
1"
2 3-
I 210-
_•
Positive
••
•
Ml
•
mm—i
Negative
Explicit Evaluat
Neutral
ion
Figure 3. Mean number of self-defining statements that were explicitly
evaluated positive, negative, or neutral by old (70-84 years) and very old
(85-103 years) individuals in BASE.
SELF-DEFINITION IN OLD AGE
interaction between the unit-weighted composite of healthrelated constraints and the two aspects of multifacetness
predicting positive emotions (entering the interaction term
as a last step in the equation for testing a moderating effect:
overall R2 = .03, p < .01; A/?2 = .001, p = .55, p = 5.13).
Furthermore, regression analyses revealed that the lack of
any relationship between the unit-weighted composite of
health-related constraints and negative emotions (r = .08,
NS) was not due to a moderator function of multifacetness
(again, entering the interaction term as a last step in the
equation for testing a moderating effect: overall R2 = .13, p
= .04; AR2 = .0002, p = .75, (3 = 2.77).
DISCUSSION
The question, how old and very old persons define themselves, was examined in the context of the Berlin Aging
Study with a heterogeneous sample of old (70 to 84 years)
and very old (85 to 103 years) adults. Before discussing the
important findings, we want to set out several limitations to
their interpretations. One of the limitations of the present
study is due to its cross-sectional design. Age-related differences only indirectly address questions concerning the
maintenance of continuity of the self-definition and age-related changes. Cohort effects cannot be excluded. Age and
cohort effects could only be separated in a longitudinal
study employing a cohort sequential design (P. B. Baltes,
Reese, & Nesselroade, 1988). In addition, it would be interesting to compare the results of this study with younger age
groups (young or middle-aged adults), in order to find out
whether findings regarding the content and structural characteristics of the self-definition are prototypical for advanced old age.
The specific coding system which served as the present
basis for the evaluation of the free responses to the question
"Who am I?", is, of course, but one of many possibilities
for the study of the self-definition. By segmenting the interview protocols into meaningful idea units and by categorizing these units into 24 content domains, it is likely that
interesting and relevant information is neglected: for example, the cross-linkage and the complexity of the selfdefinition (Linville, 1987), the hierarchical order of the
self-defining domains (Ogilvie, 1987; Rosenberg & Gara,
1985), the centrality of self-defining domains (Kling, Ryff,
& Essex, 1997), or the interrelatedness of self-concepts
with other people (Niedenthal & Beike, 1997). Also, it
should be noted that the question "Who am I?" might pull
for present-related, descriptive statements, rather than for
past or future-oriented and evaluative statements. Further
research addressing alternative methods for assessing various aspects and dimensions of the self-definition is needed
(Freund & Smith, in press).
Despite these general limitations, the present study is
unique in that it surveyed the spontaneous self-descriptions
in an age by sex stratified sample of 516 individuals between the ages of 70 and 103 years. The study design allowed comparisons of responses from very old individuals
(older than 85 years) with those from persons between the
ages of 70 and 84 years. Furthermore, the sample was locally
representative of the population older than 70 in a large city.
P63
Content of the Self-Definition
Old and very old persons in the BASE sample described
a broad spectrum of life domains in their self-definition. On
the one hand, an activity-oriented lifestyle proved to be the
central theme: The majority of the sample considered hobbies and interests, social participation, and daily living routine as self-defining domains. Thoughts about death and
dying played a rather insignificant role. These findings, as
well as the strong focus on the present reflected in the selfdefinitions, are in contrast to the image drawn by the disengagement theory (Cumming & Henry, 1961; see also
Achenbaum & Bengtson, 1994), according to which adults
in advanced old age mainly ponder over their past lives,
disengage from activities and society, and, above all, are
preoccupied with preparation for death. On the other hand,
life review and statements about personality traits also
played an important role in the self-definition of old and
very old persons, indicating that they are not exclusively
activity-oriented, but are also rather self-reflecting and preoccupied with their own private lives (Erikson, 1982). As
expected, health proved to be an important aspect of the
self-definition in advanced old age.
In general, there seemed to be more similarities than differences between the old (70 to 84 years) and the very old
(85 to 103 years) with regard to the content of their selfdefinition: Only five of the 24 categories revealed statistically significant age-related differences. The group of participants older than 85 years mentioned the domains family/
relatives and interpersonal style less frequently than the
younger age group (70 to 84 years), an age-related trend
that was also found by McCrae and Costa (1988) in a
younger sample. This result speaks against our expectation,
based on Carstensen's Socioemotional Selectivity Theory
(e.g., Carstensen, 1993), that people tend to focus on emotionally most meaningful social contacts when approaching
the end of life. We had predicted that family members and
relatives, people who have been in one's social convoy
(Antonucci, 1990) for much of one's lifetime, would be
more prominent than friends, especially in the very old. Although family proved to be one of the most salient life domains for self-definition in old and very old age, age-related differences showed that very old persons (85 to 100+
years) less frequently defined themselves using this domain
when compared with persons aged 70 to 84 years. One possible explanation of this finding is that extremely close social partners, such as one's husband or wife, might be
highly integrated into one's self in very old age. This integration of close family members into one's self might take
place to such a degree that they are perceived as part of
oneself, rather than external social partners who deserve
mentioning. Interestingly, there was no difference between
the group of participants who were still married and the
group of participants who were widowed, divorced, or single, with respect to the frequency of naming the life domains family as self-defining. Widowed, divorced, or single
older adults, however, did mention other social partners not
belonging to their family (i.e., friends, acquaintances) more
frequently than married persons. Not being married in old
age might make relationships with other social partners
more important.
P64
FREUND AND SMITH
As predicted, outdoors activities, usually requiring good
health and physical mobility, were less often mentioned by
the very old participants, whereas the domain health was
mentioned more frequently. These results indicate that
health is a central theme in the self-definition in advanced
old age. This interpretation is supported by the fact that
both the number of self-defining domains and their richness
were negatively correlated with physical and sensory constraints, whereas chronological age did not correlate with
these two aspects of the self-definition when statistically
controlled for the unit-weighted composite of physical and
sensory capacity.
These descriptive analyses of the content of the self-definition offer insight into the possibly age-specific topics and
challenges that older adults face. Interestingly, when old
and very old individuals define themselves, they turn both
"inward" (personal characteristics, life review) and "outward" (hobbies, social participation). This finding supports
an integration of the activity and disengagement theory (cf.
Achenbaum & Bengtson, 1994; Maddox, 1994; Thomae,
1987).
Self-Evaluation and Multifacetness
Finally, we would like to resume the discussion of the
function of the evaluation and multifacetness of the selfdefinition. The hypothesis that multifacetness of the selfdefinition promotes emotional well-being in advanced old
age was supported. These findings agree with the central argument of the activity theory, namely, that a large number of
various roles is essential for the well-being in old age (Atchley, 1982; Lemon, Bengtson, & Peterson, 1971; Maddox,
1994; Rosow, 1973; Thoits, 1987). However, there was no
evidence for a moderator function of multifacetness against
the negative impact of physical and sensory constraints.
Neither the number of self-defining domains nor their richness buffered the negative effect of physical constraints on
emotional well-being. This result supplements findings of
the Berlin Aging Study, which showed that although multifacetness reduces the vulnerability for depression in connection with physical illness (number of diagnoses) in advanced old age, it does not contribute to the maintenance of
emotional well-being in the face of physical impairment
(Staudinger, Freund, Linden, & Maas, in press). Thus, the
multifacetness of the self-definition seems to guard against
depression, but does not have a positive impact on emotional well-being in the presence of physical constraints.
One possible interpretation of this finding is that multifacetness might lead to a compartmentalization of the selfdefinition that threatens one's overall sense of identity,
rather than foster the maintenance of the self-definition in
the face of loss (Donahue, Robins, Roberts, & John, 1993;
see also Labouvie-Vief, Chiodo, Goguen, Diehl, & Orwoll,
1995). Future research will have to address the question of
the optimal margin of multifacetness of the self-definition
in different age-groups for helping to maintain one's selfdefinition and, at the same time, for not leading to a lack of
an overarching sense of a coherent self.
Another question of the present study concerned the relationship between the evaluation of the self-definition and
emotional well-being. In general, the spontaneous self-defi-
nition of old and very old adults contained relatively few
explicit evaluations—more than half of the self-defining
statements were purely descriptive. This finding, however,
does not necessarily suggest that the self-definitions consisted mainly of cold cognitive statements, as our study
considered only the explicit evaluations of the participants.
Furthermore, the question "Who am I?" might elicit descriptive rather than evaluative answers. A self-defining
statement such as "I am retired," may be linked with very
positive emotions for some individuals, but strongly negative ones for others, without these emotions being mentioned explicitly in answer to the question "Who am I?".
Although participants in the present study made relatively
few evaluations in their self-definitions, the evaluative
statements that were made spontaneously did reveal some
distinctive patterns.
Both old and very old adults proportionately expressed
more positive than negative evaluations. Furthermore, as
was expected, the persons older than 85 years evaluated
fewer self-defining domains as positive, and more as negative, compared with the group of persons aged 70 to 84
years. These findings contradict the prevailing stereotype
that negative events are most likely to occur in old, and especially advanced old age, leading to an increasingly negative ratio of gains and losses (Heckhausen, Dixon, & P. B.
Baltes, 1989). On the basis of research on social comparison processes (Higgins, 1990; Suls & Wills, 1991), one
might argue that old and very old persons use negative agerelated expectations as a standard of comparison for evaluating their own life situation. Such downward comparisons
might indeed result in positive evaluations, even though,
compared with younger age groups or their own past, the
situation might have changed for the worse (Heckhausen &
Krueger, 1993).
Further analyses relating positive and negative evaluation to indicators of subjective well-being provided evidence for the importance of this aspect of the self-definition
in old age. We found that the positive evaluation of the selfdefinition was related to positive emotional well-being. In
contrast, negative evaluation correlated with both positive
and negative aspects of well-being. These findings stress
the importance of a detailed analysis of the evaluative aspect of the self-definition, in order to specify the relationship between the different self-defining facets and measures
of well-being, similar to Showers' (1992) concept of "compartmentalization" of positively and negatively evaluated
self-defining domains (cf. Showers & Ryff, 1996; Woolfolk, Novalany, Gara, Allen, & Polino, 1995). To date, research on self-esteem in advanced old age has not revealed
large differences in comparison with younger age groups
(for an overview, see Bengtson, Reedy, & Gordon, 1985).
However, most of these studies refer to young old adults
and positively selected samples. Furthermore, they use
mostly global, domain-overarching measures for assessing
self-esteem, which do not permit separate analyses of positive and negative domain-specific evaluations of the selfdefinition and might conceal age-related differences in the
structure of self-esteem.
Taken together, the results of our study show that the
self-definition provides a window on the salient themes of a
SELF-DEFINITION IN OLD AGE
given age group. The self-definition also has an important
function for emotional well-being in advanced old age.
Positive evaluation, as well as richness and number of selfdefining domains, was associated with the experience of
positive emotions. Considering the numerous conditions
and processes which may influence the well-being of an individual (P. B. Baltes & M. M. Baltes, 1990; M. M. Baltes
& Carstensen, 1996; Brandtstadter & Greve, 1994; Carstensen & Freund, 1994; George, 1981; Smith & Baltes, 1996),
the magnitude of the effects, which at first sight might not
look very impressive, point to the importance of self-related processes in old age.
ACKNOWLEDGMENTS
The research reported was part of a dissertation by the first author
(Freund, 1995). A German version containing parts of this article was published in the Zeitschrift fiir Sozialpsychologie.
The study was conducted within the context of the Berlin Aging Study
(BASE). The study is directed by a Steering Committee consisting of P. B.
Baltes (psychology), H. Helmchen (psychiatry), K. U. Mayer (sociology),
and E. Steinhagen-Thiessen (internal medicine and geriatrics). BASE is a
project of the Committee on Aging and Societal Development of the
Berlin-Brandenburg Academy of Sciences in collaboration with institutes
and research centers affiliated with the Free University of Berlin, Humboldt University of Berlin and the Max Planck Institute for Human Development where the study is coordinated. The study is financially supported
by the Federal Ministry of Family Affairs, Senior Citizens, Women, and
Youth (314-1722-102/9, 314-1722-102/9a). Current memebers of the
BASE Psychology Unit include P. B. Baltes (co-director), A. M. Freund,
U. Kunzmann, U. Lindenberger, H. Maier, J. Smith (co-director), and U.
M. Staudinger. We thank our colleagues in the Psychology Unit for their
many contributions. In addition, we also want to thank A. Barnes, J. Boultwood, U. Kunzmann, and A. Overmann who assisted us in the content
analyses, and P. B. Baltes, W. Greve and U. M. Staudinger for helpful
comments on an earlier draft of this article.
Address correspondence to Dr. Freund at Max Planck Institute for
Human Development, Lentzeallee 94, 14195 Berlin, Germany. E-mail:
freund @ mpib-berlin.mpg.de
REFERENCES
Achenbaum, A. A., & Bengtson, V. L. (1994). Re-engaging the disengagement theory of aging: On the history and assessment of theory development in gerontology. The Gemntologist, 34, 756-763.
Allport, G. W. (1961). Patterns of growth in personality. New York: Holt,
Rinehart & Winston.
Antonucci, T. C. (1990). Social supports and social relationships. In R. H.
Binstock & L. K. George (Eds.), Handbook of aging and the social
sciences (3rd ed., pp. 205-226). San Diego, CA: Academic Press.
Atchley, R. C. (1982). The aging self. Psychological Theory, Research, &
Practice, 19, 338-396.
Atchley, R. C. (1991). The influence of aging or frailty on perceptions and
expressions of the self: Theoretical and methodological issues. In J. E.
Birren, J. E. Lubben, J. C. Rowe, & D. E. Deutchman (Eds.), The
concept and measurement of quality of life in the frail elderly (pp.
207-225). San Diego, CA: Academic Press.
Baltes, M. M., & Carstensen, L. L. (1996). The process of successful
aging. Aging and Society, 16, 397-422.
Baltes, M. M., Maas, I., Wilms, H.-U., Borchelt, M., & Little, T. D. (in
press). Everyday competence in old and very old age: Theoretical
considerations and empirical findings. In P. B. Baltes & K. U. Mayer
(Eds.), The Berlin Aging Study: Aging from 70 to 100. New York:
Cambridge University Press.
Baltes, M. M., & Silverberg, S. B. (1994). The dynamics between dependency and autonomy: Illustrations across the life span. In D. L. Featherman, R. M. Lerner, & M. Perlmutter (Eds.), Life span development
and behavior (pp. 41-90). Hillsdale, NJ: Erlbaum.
Baltes, P. B. (1987). Theoretical propositions of life span developmental
P65
psychology: On the dynamics between growth and decline. Developmental Psychology, 23, 611-623.
Baltes, P. B. (1997). On the incomplete architecture of human ontogeny.
Selection, optimization, and compensation as foundation of developmental theory. American Psychologist, 52, 366-380.
Baltes, P. B., & Baltes, M. M. (1990). Psychological perspectives on successful aging: The model of selective optimization with compensation.
In P. B. Baltes & M. M. Baltes (Eds.), Successful aging: Perspectives
from the behavioral sciences (pp. 1-34). Cambridge: Cambridge University Press.
Baltes, P. B., Mayer, K. U., Helmchen, H., & Steinhagen-Thiessen, E. (in
press). The Berlin Aging Study (BASE): Sample, design, and overview
of measures. In P. B. Baltes & K. U. Mayer (Eds.), The Berlin Aging
Study: Aging from 70 to 100. New York: Cambridge University Press.
Baltes, P. B. Reese, H. W., & Nesselroade, J. R. (1988). Life span developmental psychology: Introduction to research methods. Hillsdale, NJ:
Erlbaum.
Baltes, P. B., & Smith, J. (1997). A systemic-holistic view of psychological functioning in very old age: Introduction to a collection of articles
from the Berlin Aging Study. Psychology and Aging, 12, 395-409.
Bengtson, V. L., Reedy, M. N., & Gordon, C. (1985). Aging and self-conceptions: Personality processes and social contexts. In J. E. Birren &
K. W. Schaie (Eds.), Handbook of the psychology of aging (pp.
544-593). New York: Van Nostrand Reinhold.
Brandtstadter, J. (1985). Entwicklungsprobleme des Jugendalters als Probleme des Aufbaus von Handlungsorientierungen [Developmental
problems in adolescence as problems in building action orientation].
In D. Liepman & A. Stiksrud (Eds.), Entwicklungsaufgaben und Bewaltigungsprobleme der Adoleszenz [Developmental tasks and problems of coping in adolescence] (pp. 5-12). Gottingen: Hogrefe.
Brandtstadter, J., & Greve, W. (1994). The aging self: Stabilizing and protective processes. Developmental Review, 14, 52-80.
Brennan, R. L., & Prediger, D. J. (1981). Coefficient kappa: Some uses,
misuses, and alternatives. Educational and Psychological Measurement, 41, 687-699.
Breytspraak, L. M. (1994). The development of self in later life. Boston,
MA: Little, Brown & Co.
Bugental, J. F. T., & Zelen, S. L. (1950). Investigations into the self-concept—the WAY technique. Journal of Personality, 18, 483-498.
Cantor, N., & Kihlstrom, J. F. (1989). Social intelligence and cognitive
assessments of personality. In R. S. Wyer & T. K. Srull (Eds.), Advances in social cognition (Vol. 2, pp. 1-59). Hillsdale, NJ: Lawrence
Erlbaum.
Carstensen, L. L. (1993). Motivation for social contact across the life
span: A theory of socioemotional selectivity. In J. Jacobs (Ed.). Nebraska symposium on motivation: 1992. Developmental perspectives
on motivation (Vol. 40., pp. 209-254). Lincoln, NE: University of Nebraska Press.
Carstensen, L. L., & Freund, A.M. (1994). The resilience of the self. Developmental Review, 14, 81-92.
Cooley, G. H. (1902). Human nature and the social order. New York:
Scribner's.
Cross, S., & Markus, H. (1991). Possible selves across the life span.
Human Development, 34, 230-255.
Cumming, E., & Henry, W. E. (1961). Growing old. New York: Basic
Books.
Dittmann Kohli, F. (1995). Das persdnliche Sinnsystem: Ein Vergleich
zwischen friihem und spa'tem Erwachsenenalter [The personal meaning system: A comparison of early and late adulthood]. Gottingen:
Hogrefe.
Donahue, E. M, Robins, R. W., Roberts, B. W., & John, O. P. (1993). The
devided self: Concurrent and longitudinal effects of psychological adjustment and social roles on self-concept differentiation. Journal of
Personality and Social Psychology, 64, 834-846.
Epstein, S. (1973). The self-concept revisited. Or a theory of a theory.
American Psychologist, 28, 404-416.
Epstein, S. (1990). Cognitive-experiential self-theory. In A. Pervin (Eds.),
Handbook of personality: Theory and research (pp. 165-192). New
York: Guilford Press.
Erikson, E. H. (1959). Identity and the life cycle. New York: International
Universities.
Erikson, E. H. (1982). The life cycle completed. NY: Norton.
Filipp, S.-H. (1980). Entwicklung von Selbstkonzepten [Development of
P66
FREUND AND SMITH
self-concepts]. Zeitschrift fur Entwicklungspsychologie und Pa'dagogische Psychologic 12, 105-125.
Filipp, S.-H., & Klauer, T. (1986). Conceptions of self over the life span:
Reflections on the dialectics of change. In M. M. Baltes & P. B.
Baltes (Eds.), The psychology of control and aging (pp. 167-205).
Hillsdale, NJ: Erlbaum.
Freund, A. M. (1995). Wer bin ich? Die Selbstdefinition alter Menschen
[Who am I? The self-definition of older persons]. Berlin: Sigma.
Freund, A. M., & Baltes, P. B. (in press). The orchestration of selection,
optimization, and compensation: An action-theoretical conceptualization of a theory of developmental regulation. In W. J. Perrig & A.
Grob (Eds.), Control of human behaviour, mental processes, and consciousness. Mahwah, NJ: Erlbaum
Freund, A. M., & Smith, J. (in press). Temporal stability of older persons'
spontaneous self-definition. Experimental Aging Research.
George, L. K. (1981). Subjective well-being: Conceptual and methodological issues. In C. Eisdorfer (Ed.), Annual review of gerontology and
geriatrics (Vol. 2, pp. 267-282). New York: Springer.
George, L. K., & Okun, M. A. (1985). Self-concept content. In E. W.
Busse, J. L. Maddox, J. B. Nowlin, & J. C. Siegler (Eds.), Normal
aging HI: Reports from the Duke Longitudinal Studies 1975-1984
(pp. 267-282). Durham, NC: Duke University Press.
Gordon, C. (1968). Self-conceptions: Configuration of content. In C. Gordon & K. J. Gergen (Eds.), The self in social interaction: Classic and
contemporary perspectives. New York: Wiley.
Greenwald, G. A., & Pratkanis, A. R. (1984). The self. In R. S. Wyer & T.
K. Srull (Eds.), Handbook of social cognition (Vol. 3, pp. 129-178).
Hillsdale, NJ: Erlbaum.
Harter, S. (1983). Developmental perspectives on the self-system. In E. M.
Hetherington (Ed.), Handbook of child psychology: Social and personality development (4 ed., Vol. 4, pp. 275-385). New York: Wiley.
Havighurst, R. J. (1963). Successful aging. In R. H. Williams, C. Tibbitts,
& W. Donahue (Eds.), The process of aging: Social and psychological
perspectives (Vol. 1, pp. 299-320). New York: Atherton Press.
Heckhausen, J., Dixon, R. A., & Baltes, P. B. (1989). Gains and losses in
development throughout adulthood as perceived by different adult age
groups. Developmental Psychology, 25, 109-121.
Heckhausen, J., & Krueger, J. (1993). Developmental expectations for the
self and most other people: Age grading in three functions of social
comparison. Developmental Psychology, 29, 539-548.
Herzog, A. R., & Rodgers, W. L. (1981). Age and satisfaction: Data from
several large surveys. Research on Aging, 3, 142-165.
Higgins, E. T. (1990). Personality, social psychology, and person-situation
relations: Standards and knowledge activation as a common language.
In L. A. Pervin (Ed.), Handbook of personality: Theory and research
(pp. 301-338). New York: Guilford Press.
Hooker, K. (1992). Possible selves and perceived health in older adults
and college students. Journal of Gerontology: Psychological Sciences, 47, P85-P95.
James, W. (1890). Principles of psychology (Vol. 1). New York: Holt.
Johnson, C. L., & Barer, B. M. (1997). Life beyond 85 years: The aura of
survivorship. New York: Springer.
Katz, S. T. D., Downs, H. R., Cash, R. C , & Grotz, R. C. (1970). Progress
in development of the index ADL. The Gerontologist, 10, 20-30.
Kelly, G. A. (1955). The psychology of personal constructs (Vol. 1). New
York: Norton.
Kihlstrom, J. F., & Cantor, N. (1984). Mental representations of the self. In
L. Berkowitz (Ed.), Advances in experimental social psychology (Vol.
17, pp. 1-47). Orlando: Academic Press.
Kling, K. C , Ryff, C. D., & Essex, M. J. (1997). Adaptive changes in the
self-concept during a life transition. Personality and Social Psychological Bulletin, 23, 981-990.
Krippendorf, K. (1980). Content analysis. An introduction to its methodology. Beverly Hills, CA: Sage.
Kuhn, M. H., & McPartland, T. S. (1954). An empirical investigation of
self-attitudes. American Sociological Review, 19, 68-76.
Labouvie-Vief, G., Chiodo, L. M., Goguen, L. A., Diehl, M., & Orwoll, L.
(1995). Representations of self across the life span. Psychology and
Aging, 10, 404-415.
Lawton, M. P. (1975). The Philadelphia Center Morale Scale: A revision.
Journal of Gerontology, 30, 85-89
Lecky, P. (1945). Self-consistency: A theory of personality. New York: Island Press.
L'Ecuyer, R. (1990). Le developement du concept de soi de 0 a 100 ans:
Cent ans apres William James. Revue quebecoise de psychologie,
Special edition on the self-concept.
Lemon, B. W, Bengtson, V. L., & Peterson, J. A. (1971). An exploration
of the activity theory of aging: Activity types and life satisfactions
among in-movers to a retirement community. Journal of Gerontology,
27,511-523.
Liang, J. & Bollen, K.A. (1983). The structure of the Philadelphia Geriatric Center Morale Scale: A Reinterpretation. Journals of Gerontology, 38, 151-189.
Lindenberger, U., & Baltes, P. B. (1997). Intellectual functioning in old
and very old age: Cross-sectional results from the Berlin Aging Study.
Psychology and Aging, 12, 410-432.
Lindenberger, U., Gilberg, R., Little, T. D., Nuthmann, R., Potter, U., &
Baltes, P. B. (in press). Sample selectivity and generalizability of the
results of the Berlin Aging Study. In P. B. Baltes & K. U. Mayer
(Eds.), The Berlin Aging Study: Aging from 70 to 100. New York:
Cambridge University Press.
Linville, P. W. (1987). Self complexity as a cognitive buffer against stress
related illness and depression. Journal of Personality and Social Psychology, 52, 663-676.
Maddox, G. L. (1994). Lives through the years revisited. The Gerontologist, 6, 764-767.
Manton, K. G., Stallard, E., & Corder, L. (1995). Changes in morbidity
and chronic disability in the U. S. elderly population: Evidence from
the 1982, 1984, and 1989 National Long Term Care Surveys. Journal
of Gerontology: Social Sciences, 50B, S194-S204.
Markus, H. (1983). Self-knowledge: An expanded view. Journal of Personality, 51, 543-565.
Marsh, H. W. (1987). The hierarchical structure of self-concept and the application of hierarchical confirmatory factor analysis. Journal of Educational Measurement, 24, 17-19.
Marsiske, M., Klumb, P. & Baltes, M. M. (1997) Everyday activity patterns and sensory functioning in old age. Psychology and Aging, 12,
444-^57.
McAdams, D. P. (1996). Personality, modernity, and the storied self: A
contemporary framework for studying persons. Psychological Inquiry, 7,295-321.
McCrae, R. R., & Costa, P. T, Jr. (1988). Age, personality and the spontaneous self-concept. Journal of Gerontology: Social Sciences, 43B,
S177-S185.
McGuire, W. J. (1984). Search for the self: Going beyond self-esteem and
the reactive self. In R. A. Zucker, J. Aronoff, & A. I. Rabin (Eds.).
Personality and the prediction of behavior (pp. 73-120). New York:
Academic Press.
Mead, G. H. (1925). The genesis of self and social control. Journal of
Ethics, 35, 251-273.
Neugarten, B. L. (1964). Personality in middle and late life. New York:
Atherton Press.
Niedenthal, P. M., & Beike, D. R. (1997). Interrelated and isolated selfconcepts. Personality and Social Psychology Review, 1, 106-128.
Ogilvie, D. M. (1987). Life satisfaction and identity structure in late, middle aged men and women. Psychology and Aging, 2, 217-224.
Peck, R. (1968). Psychological developments in the second half of life. In
B. L. Neugarten (Ed.), Middle age and aging: A reader in social psychology (pp. 88-92). Chicago: University of Chicago Press.
Pelham, B. W, & Swann, W. B. (1989). From self-conceptions to selfworth: On the sources and structure of global self-esteem. Journal of
Personality and Social Psychology, 57, 672-680.
Rosenberg, S., & Gara, M. A. (1985). The multiplicity of personal identity.
In P. Shaver (Ed.), Review of personality and social psychology (Vol.
6, pp. 87-113). Beverly Hills, CA: Sage.
Rosow, I. (1973). The social context of the aging self. The Gerontologist,
13, 82-87.
Showers, C. J. (1992). Compartmentalization of positive and negative selfknowledge: Keeping bad apples out of the bunch. Journal of Personality and Social Psychology, 62, 1036-1049.
Showers, C. J., & Ryff, C. D. (1996). Self-differentiation and well-being in
a life transition. Personality and Social Psychological Bulletin, 22,
448-460.
Smith, J., & Baltes, P. B. (i996). Altern aus psychologischer Perspektive:
Trends und Profile im hohen Alter [Trends and profiles of psychological functioning in very old age]. In K. U. Mayer & P. B. Baltes (Eds.),
SELF-DEFINITION IN OLD AGE
Die Berliner Altersstudie [The Berlin Aging Study] (pp. 251-275).
Berlin: Akademie Verlag.
Smith, J., & Baltes, P. B. (in press). Trends and profiles of psychological
functioning in very old age. In P. B. Baltes & K. U. Mayer (Eds.), The
Berlin Aging Study: Aging from 70 to 100. New York: Cambridge
University Press.
Smith, J., Fleeson, W., Geiselmann, B., Settersten, R. A., Jr., & Kunzmann, U. (in press). Sources of well-being in very old age. In P. B.
Baltes & K. U. Mayer (Eds.), The Berlin Aging Study: Aging from 70
to 100. New York: Cambridge University Press.
Staudinger, U. M., Freund, A., Linden, M., & Maas, I. (in press). Self, personality, and life regulation: Facets of psychological resilience in old
age. In P. B. Baltes & K. U. Mayer (Eds.), The Berlin Aging Study:
Aging from 70 to 100. New York: Cambridge University Press.
Steinhagen-Thiessen, E., & Borchelt, M. (in press). Morbidity, medication,
and functional limitations in very old age. In P. B. Baltes & K. U.
Mayer (Eds.), The Berlin Aging Study: Aging from 70 to 100. New
York: Cambridge University Press.
Suls, J., & Wills, T. A. (1991) (Eds.), Social comparison: Contemporary
theory and research. Hillsdale, NJ: Erlbaum.
Stryker, S. (1987). The vitalization of symbolic interctionism. Social Psychology Quaterly, 50, 83-94.
Swann, W. B., Jr. (1990). To be adored or to be known: The interplay of
self-enhancement and self-verification. In R. M. Sorrentino & E. T.
Higgins (Eds.), Motivation and cognition (Vol. 2, pp. 408^148). Hillsdale, NJ: Lawrence Erlbaum.
Thoits, P. A. (1983). Multiple identities and psychological well-being. A
reformulation and test of the social isolation hypothesis. American
Sociological Review, 48, 174-187.
Thoits, P. A. (1987). Negotiating roles. In F. J. Crosby (Ed.) Spouse, parent, worker: In gender and multiple roles. New Haven: Yale University Press.
Thomae, H. (1979). The concept of development and life-span developmental psychology. In P. B. Baltes & O. G. Brim, Jr. (Eds.), Life-span
development and behavior (Vol. 2, pp. 282-312). New York: Academic Press.
Thomae, H. (1987). Patterns of psychological aging—Findings of the
Bonn Longitudinal Study of Aging. In U. Lehr & H. Thomae (Eds.),
Formen seelischen Alterns. Ergebnisse der Bonner Gerontologischen
Ldngsschnittstudie (BOLSA) [Patterns of psychological aging. Results
of the Bonn Longitudinal Study of Aging (BOLSA)] (pp. 279-286).
Stuttgart: Enke.
Troll, L. E., & Skaff, M. M. (1997). Perceived continuity of self in very
old age. Psychology and Aging, 12, 162-169.
Waterman, A. S., & Archer, S. L. (1990). A life span perspective on identity formation: Developments in form, function, and process. In P. B.
Baltes, D. L. Featherman, & R. M. Lerner (Eds.), Life-span development and behavior (Vol. 10, p. 29-57). Hillsdale, NJ: Lawrence
Erlbaum.
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS
scales. Journal of Personality and Social Psychology, 54, 1063-1070.
Whitbourne, S. K. (1985). The aging body: Physiological changes and
psychological consequences. New York: Springer.
Wicklund, R. A., & Gollwitzer, P. M. (1982). Symbolic self completion.
Hillsdale, NJ: Lawrence Erlbaum.
P67
Woolfolk, R. L., Novalany, J., Gara, M. A. Allen, L. A., & Polino (1995).
Self-complexity, self-evaluation, and depression: An examination of
form and content within the self-schema. Journal of Personality and
Social Psychology, 68, 1108-1120.
Received February 3, 1998
Accepted September 14, 1998
Appendices
Appendix 1. The verbatim instruction was the following: "In
this task you are asked to describe yourself. Imagine you sit down
in a quiet hour and think about the following question: Who am I?
This means for example: What do you do? What do you feel?
What do you represent? What interests you? Who are you? Think
about how you would answer this question. Just say in ten short
statements whatever comes into your mind, and I will write it
down for you. How would you describe yourself? I am . . . "
Appendix 2. In BASE, N = 109 individuals were diagnosed as
having suspected dementia (see also Smith & Baltes, 1996). The
diagnoses were differentiated at three levels: Level 1 (n = 37) indicated participants with symptomatic memory difficulties, and
Levels 2 and 3 (n = 33 and n = 39, respectively) indicated cases
with moderate to severe symptomatology (DSM-III-R). In all of
these individuals, language functioning was still intact. They were
able to respond to questions and to produce sentences. For research ethical reasons, individuals with severe dementia that involved the loss of language were excluded from participation in
the study.
Appendix 3. Although of less theoretical interest in the present
context, we also looked at the relationship between the number of
purely descriptive statements and positive and negative affect.
Whereas the number of purely descriptive statements was not significantly related to positive affect (r = .07), it was negatively correlated with negative affect (r = -.12, p < .01). This correlation remained after controlling for health (partial r = -.12, p < .01). A
speculative interpretation of this finding is that pure descriptions of
oneself are subjectively perceived as the outcome of having successfully avoided a negative state (or trajectory). Avoiding a negative outcome might lead to the absence of negative affect rather
than contributing to positive affect (cf., Freund & Baltes, in press).
Appendix 4. Being diagnosed as having suspected dementia (n
= 109; see Appendix 2) was associated with lower multifacetness
(for number of domains f[514] = 4.06, p < .001; for richness
t[5\4] = 5.39, p <.001). Excluding participants with suspected dementia from the analyses, however, did not change the findings.
Again, we did not find evidence supporting the hypothesis that
multifacetness of the self-definition buffers negative effects of
health-related constraints.